diff --git a/runs/single_llava/context/L01_s13/adapter/README.md b/runs/single_llava/context/L01_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L01_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L01_s13/adapter/adapter_config.json b/runs/single_llava/context/L01_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..6aa9217b9ec284899cbc9b9ab7431638ac70d20d
--- /dev/null
+++ b/runs/single_llava/context/L01_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.1.mlp.down_proj",
+    "model.language_model.layers.1.self_attn.v_proj",
+    "model.language_model.layers.1.self_attn.k_proj",
+    "model.language_model.layers.1.self_attn.o_proj",
+    "model.language_model.layers.1.mlp.gate_proj",
+    "model.language_model.layers.1.mlp.up_proj",
+    "model.language_model.layers.1.self_attn.q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L01_s13/config.json b/runs/single_llava/context/L01_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5e20f561b37ef3b9bd59e39645b3a003317de08
--- /dev/null
+++ b/runs/single_llava/context/L01_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 1,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L01_s13/steps.jsonl b/runs/single_llava/context/L01_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..8a51a549a50c4940e5ccfaa0550c860a698e0242
--- /dev/null
+++ b/runs/single_llava/context/L01_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.7376080751419067, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 18.21, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.9404605031013489, "loss_scale": 4.0, "retries": 0, "seconds": 2.2, "max_memory_gb": 18.22, "ce": 2.679309105849825, "kl": 0.010342742316424847, "context": 2.6844804864376783}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 2.125945806503296, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 18.22, "ce": 5.379236817359924, "kl": 0.1790473797263985, "context": 5.4687604904174805}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 1.406678557395935, "loss_scale": 4.0, "retries": 0, "seconds": 3.3, "max_memory_gb": 18.22, "ce": 3.815349311567843, "kl": 0.05171321358284331, "context": 3.8412059266120195}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.17895527184009552, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 18.22, "ce": 0.08575703017413616, "kl": 0.0368805134203285, "context": 0.10419728606939316}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.6453537344932556, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 18.22, "ce": 1.4614265412092209, "kl": 0.18266987416427583, "context": 1.552761483937502}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 1.6638957262039185, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 18.22, "ce": 3.4159883996471763, "kl": 0.10645121728884988, "context": 3.4692139946855605}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 1.9232434034347534, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 18.23, "ce": 3.663819432258606, "kl": 0.15977236069738865, "context": 3.743705563247204}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 1.5256216526031494, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 18.23, "ce": 2.813711278140545, "kl": 0.16244306764565408, "context": 2.894932806491852}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.8285219669342041, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 18.23, "ce": 1.6207847744226456, "kl": 0.059720152348745614, "context": 1.6506448090076447}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 1.8672510385513306, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 18.23, "ce": 2.593834443949163, "kl": 0.00929369334699004, "context": 2.598481308668852}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 1.150572419166565, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 18.23, "ce": 1.5164186605252326, "kl": 0.017532389392727055, "context": 1.5251848145853728}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 1.9375978708267212, "loss_scale": 4.0, "retries": 0, "seconds": 8.3, "max_memory_gb": 18.23, "ce": 2.633444096427411, "kl": 0.0073406050651101395, "context": 2.637114468961954}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 2.9325249195098877, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 18.23, "ce": 3.833781197667122, "kl": 0.11697630712296814, "context": 3.8922694101929665}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 2.161702871322632, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 18.23, "ce": 2.5398754951311275, "kl": 0.2698106986063067, "context": 2.6747808616491966}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 1.3229111433029175, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 18.23, "ce": 1.345980865182355, "kl": 0.008969108075689292, "context": 1.3504654557909817}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 1.4373747110366821, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 18.23, "ce": 1.5622919807210565, "kl": 0.07808439276413992, "context": 1.6013341583311558}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 3.1208200454711914, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 18.24, "ce": 3.508824534714222, "kl": 0.09419492981396616, "context": 3.5559218898415565}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 4.202141761779785, "loss_scale": 4.0, "retries": 0, "seconds": 11.5, "max_memory_gb": 18.24, "ce": 5.239079713821411, "kl": 0.3287815147486981, "context": 5.403470516204834}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.05616855248808861, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 18.24, "ce": 0.012885322561487556, "kl": 0.0020806705724680796, "context": 0.013925657724030316}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 2.0101873874664307, "loss_scale": 4.0, "retries": 0, "seconds": 12.6, "max_memory_gb": 18.24, "ce": 2.7249249818269163, "kl": 0.03076278052321868, "context": 2.740306413732469}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 3.152761459350586, "loss_scale": 4.0, "retries": 0, "seconds": 13.1, "max_memory_gb": 18.24, "ce": 2.8672371078282595, "kl": 0.052422083681449294, "context": 2.8934481646865606}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 3.090667724609375, "loss_scale": 4.0, "retries": 0, "seconds": 13.7, "max_memory_gb": 18.24, "ce": 2.5836324305273592, "kl": 0.5143351724400418, "context": 2.8407999931368977}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 5.7888102531433105, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 18.24, "ce": 4.87368643283844, "kl": 0.013864439446479082, "context": 4.880618631839752}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 5.435366153717041, "loss_scale": 4.0, "retries": 0, "seconds": 14.8, "max_memory_gb": 18.24, "ce": 4.102435946464539, "kl": 0.03623349324334413, "context": 4.120552562177181}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 7.357712268829346, "loss_scale": 4.0, "retries": 0, "seconds": 15.3, "max_memory_gb": 18.24, "ce": 5.04179984331131, "kl": 0.06775383185595274, "context": 5.07567685842514}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 4.429919242858887, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 18.24, "ce": 4.068496577441692, "kl": 0.08707543747732416, "context": 4.112034227699041}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 2.835806369781494, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 18.24, "ce": 2.7365202605724335, "kl": 0.2276319544762373, "context": 2.8503362089395523}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 3.6288249492645264, "loss_scale": 4.0, "retries": 0, "seconds": 16.9, "max_memory_gb": 18.24, "ce": 2.5310678449459374, "kl": 0.5195266348600853, "context": 2.7908312242943794}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 4.840408802032471, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 18.24, "ce": 2.605150311253965, "kl": 0.038290229509584606, "context": 2.624295435845852}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 4.741778373718262, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 18.24, "ce": 2.717921420931816, "kl": 0.02980279119219631, "context": 2.7328227683901787}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 7.1298112869262695, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 18.24, "ce": 5.1161816120147705, "kl": 0.015060827077832073, "context": 5.123711943626404}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 2.2590880393981934, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 18.24, "ce": 2.398184174671769, "kl": 0.19385266047902405, "context": 2.4951104819774628}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 5.015777111053467, "loss_scale": 4.0, "retries": 0, "seconds": 19.6, "max_memory_gb": 18.24, "ce": 2.502482812386006, "kl": 0.615818471647799, "context": 2.810392056591809}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 6.438923358917236, "loss_scale": 4.0, "retries": 0, "seconds": 20.2, "max_memory_gb": 18.24, "ce": 2.5213087359443307, "kl": 0.02885297422471922, "context": 2.5357351256534457}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 4.732455730438232, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 18.24, "ce": 2.4310567900538445, "kl": 0.007069282582961023, "context": 2.43459140881896}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 3.8157331943511963, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 18.24, "ce": 2.3629493452608585, "kl": 0.3457048118580133, "context": 2.5358018092811108}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 9.884786605834961, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 18.24, "ce": 4.407026767730713, "kl": 0.484049159073038, "context": 4.649051308631897}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 4.188363552093506, "loss_scale": 4.0, "retries": 0, "seconds": 22.3, "max_memory_gb": 18.24, "ce": 1.2989514688961208, "kl": 0.03097886056639254, "context": 1.3144409586675465}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 6.512091159820557, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 18.24, "ce": 2.5391197614371777, "kl": 0.11344576114788651, "context": 2.5958425998687744}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 2.8237619400024414, "loss_scale": 4.0, "retries": 0, "seconds": 23.4, "max_memory_gb": 18.24, "ce": 1.385099794715643, "kl": 0.17048641690053046, "context": 1.4703429713845253}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 6.935862064361572, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 18.24, "ce": 2.1633203476667404, "kl": 0.0070581468171440065, "context": 2.166849416680634}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 8.392904281616211, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 18.24, "ce": 2.163207652978599, "kl": 0.010115203549503349, "context": 2.1682652039453387}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 4.1327948570251465, "loss_scale": 4.0, "retries": 0, "seconds": 25.0, "max_memory_gb": 18.24, "ce": 1.259248135611415, "kl": 0.25925476814154536, "context": 1.3888755203224719}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 4.6808953285217285, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 18.24, "ce": 1.2641260288655758, "kl": 0.05534238356631249, "context": 1.291797213256359}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 7.29893159866333, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 18.24, "ce": 2.059255935251713, "kl": 0.07898017702973448, "context": 2.0987460240721703}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 8.377323150634766, "loss_scale": 4.0, "retries": 0, "seconds": 26.6, "max_memory_gb": 18.24, "ce": 1.7550556482747197, "kl": 0.03334357867424842, "context": 1.7717273712623864}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 14.719626426696777, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 18.24, "ce": 3.272421967238188, "kl": 0.013981290030642413, "context": 3.2794126234948635}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 18.086837768554688, "loss_scale": 4.0, "retries": 0, "seconds": 27.7, "max_memory_gb": 18.24, "ce": 2.9311758782714605, "kl": 0.07201784569770098, "context": 2.9671848230063915}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 14.323931694030762, "loss_scale": 4.0, "retries": 0, "seconds": 28.2, "max_memory_gb": 18.24, "ce": 3.183785547502339, "kl": 0.05240462432266213, "context": 3.2099878881126642}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 5.886399745941162, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 18.24, "ce": 1.1808416629210114, "kl": 0.11847176519222558, "context": 1.240077517926693}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 12.004474639892578, "loss_scale": 4.0, "retries": 0, "seconds": 29.3, "max_memory_gb": 18.24, "ce": 1.857993207871914, "kl": 0.12052774504991248, "context": 1.9182570427656174}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 19.61351203918457, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 18.24, "ce": 3.044460155069828, "kl": 0.21307231858372688, "context": 3.1509964168071747}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 22.93589210510254, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 18.24, "ce": 3.001146674156189, "kl": 0.03939057217212394, "context": 3.020841956138611}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 12.346293449401855, "loss_scale": 4.0, "retries": 0, "seconds": 30.9, "max_memory_gb": 18.24, "ce": 1.8300859830342233, "kl": 0.021023116540163755, "context": 1.8405975373461843}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 34.28090286254883, "loss_scale": 4.0, "retries": 0, "seconds": 31.4, "max_memory_gb": 18.24, "ce": 3.2736990451812744, "kl": 0.3468517635483295, "context": 3.44712495803833}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 8.0927734375, "loss_scale": 4.0, "retries": 0, "seconds": 32.0, "max_memory_gb": 18.24, "ce": 0.8157362379133701, "kl": 0.11296675889752805, "context": 0.8722196193411946}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 11.637081146240234, "loss_scale": 4.0, "retries": 0, "seconds": 32.5, "max_memory_gb": 18.24, "ce": 1.63437969237566, "kl": 0.005414601746451808, "context": 1.637087007984519}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 6.526420593261719, "loss_scale": 4.0, "retries": 0, "seconds": 33.1, "max_memory_gb": 18.24, "ce": 1.1372733525931835, "kl": 0.08285811101086438, "context": 1.178702400997281}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 14.558892250061035, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 18.24, "ce": 1.596313988789916, "kl": 0.014872873813146725, "context": 1.6037504617124796}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 22.33942413330078, "loss_scale": 4.0, "retries": 0, "seconds": 34.1, "max_memory_gb": 18.24, "ce": 1.5892736464738846, "kl": 0.09560365392826498, "context": 1.6370754390954971}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 9.46692943572998, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 18.24, "ce": 0.8395881596952677, "kl": 0.03974704205757007, "context": 0.859461679123342}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 15.739212036132812, "loss_scale": 4.0, "retries": 0, "seconds": 35.2, "max_memory_gb": 18.24, "ce": 1.8912204951047897, "kl": 0.08203302024048753, "context": 1.932236977852881}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 21.802553176879883, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 18.27, "ce": 1.8586441781371832, "kl": 0.045379314105957747, "context": 1.8813337795436382}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 19.258071899414062, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 18.27, "ce": 1.820362776517868, "kl": 0.11479807877913117, "context": 1.8777618110179901}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 4.907248020172119, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 18.27, "ce": 0.5906326770782471, "kl": 0.061760017852066085, "context": 0.6215126966126263}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 11.237154960632324, "loss_scale": 4.0, "retries": 0, "seconds": 37.4, "max_memory_gb": 18.27, "ce": 1.0416990220546722, "kl": 0.04413418215699494, "context": 1.0637661069631577}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 6.785576820373535, "loss_scale": 4.0, "retries": 0, "seconds": 37.9, "max_memory_gb": 18.27, "ce": 0.6413977444171906, "kl": 0.07025387673638761, "context": 0.6765247127041221}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 9.198704719543457, "loss_scale": 4.0, "retries": 0, "seconds": 38.5, "max_memory_gb": 18.27, "ce": 0.7608813308179379, "kl": 0.03253761597443372, "context": 0.7771501373499632}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 7.006900787353516, "loss_scale": 4.0, "retries": 0, "seconds": 39.0, "max_memory_gb": 18.27, "ce": 0.5863928198814392, "kl": 0.006654970347881317, "context": 0.589720313437283}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 10.722969055175781, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 18.27, "ce": 0.9311029314994812, "kl": 0.028550022281706333, "context": 0.9453779458999634}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 13.321512222290039, "loss_scale": 4.0, "retries": 0, "seconds": 40.1, "max_memory_gb": 18.27, "ce": 1.001702442765236, "kl": 0.1271388721652329, "context": 1.0652718991041183}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.559457540512085, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 18.27, "ce": 0.2991416547447443, "kl": 0.03991512022912502, "context": 0.3190992083400488}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 3.0521187782287598, "loss_scale": 4.0, "retries": 0, "seconds": 41.2, "max_memory_gb": 18.27, "ce": 0.4722152575850487, "kl": 0.02508535166271031, "context": 0.48475794307887554}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 2.122023582458496, "loss_scale": 4.0, "retries": 0, "seconds": 41.7, "max_memory_gb": 18.27, "ce": 0.8192267194390297, "kl": 0.19596783816814423, "context": 0.91721061617136}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 2.4752719402313232, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 18.27, "ce": 0.43397206813097, "kl": 0.21550441719591618, "context": 0.541724257171154}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 4.829584121704102, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 18.27, "ce": 0.8660785555839539, "kl": 0.04853285942226648, "context": 0.8903449773788452}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 3.106229782104492, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 18.27, "ce": 0.3800245653837919, "kl": 0.05524195544421673, "context": 0.40764554031193256}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 2.8915350437164307, "loss_scale": 4.0, "retries": 0, "seconds": 43.9, "max_memory_gb": 18.27, "ce": 0.33642882108688354, "kl": 0.09282084577716887, "context": 0.3828392457216978}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 3.7754316329956055, "loss_scale": 4.0, "retries": 0, "seconds": 44.4, "max_memory_gb": 18.27, "ce": 0.3540552780032158, "kl": 0.024641262833029032, "context": 0.3663759045302868}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 3.8144278526306152, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 18.27, "ce": 0.6233523115515709, "kl": 0.10610251920297742, "context": 0.6764035746455193}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 2.1237945556640625, "loss_scale": 4.0, "retries": 0, "seconds": 45.5, "max_memory_gb": 18.27, "ce": 0.3841622397303581, "kl": 0.07658684812486172, "context": 0.42245566844940186}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 2.8666231632232666, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 18.27, "ce": 0.46068884059786797, "kl": 0.12567527405917645, "context": 0.5235264822840691}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.7726753950119019, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 18.27, "ce": 0.2816368490457535, "kl": 0.04637070163153112, "context": 0.30482220370322466}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.9571926593780518, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 18.27, "ce": 0.14508799370378256, "kl": 0.04098024401173461, "context": 0.16557811107486486}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 3.2575180530548096, "loss_scale": 4.0, "retries": 0, "seconds": 47.6, "max_memory_gb": 18.27, "ce": 0.4529199078679085, "kl": 0.15002442721743137, "context": 0.5279321223497391}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 2.2805981636047363, "loss_scale": 4.0, "retries": 0, "seconds": 48.2, "max_memory_gb": 18.27, "ce": 0.44596934132277966, "kl": 0.06090609263628721, "context": 0.47642238810658455}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.6336581707000732, "loss_scale": 4.0, "retries": 0, "seconds": 48.7, "max_memory_gb": 18.27, "ce": 0.5194545490667224, "kl": 0.034429684543283656, "context": 0.5366693884134293}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.969314455986023, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 18.27, "ce": 0.14042979618534446, "kl": 0.00837124150712043, "context": 0.1446154131554067}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.5414490699768066, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 18.27, "ce": 0.41820533853024244, "kl": 0.03697933442890644, "context": 0.4366950076073408}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.307892084121704, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 18.27, "ce": 0.3789929747581482, "kl": 0.08912669611163437, "context": 0.4235563240945339}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.8115626573562622, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 18.27, "ce": 0.21188370138406754, "kl": 0.10883526550605893, "context": 0.26630133111029863}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 3.919248580932617, "loss_scale": 4.0, "retries": 0, "seconds": 51.4, "max_memory_gb": 18.27, "ce": 0.34589238930493593, "kl": 0.06337381163029931, "context": 0.3775792964734137}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 3.915025472640991, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 18.27, "ce": 0.3358202502131462, "kl": 0.14912031264975667, "context": 0.41038040816783905}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 2.5162248611450195, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 18.27, "ce": 0.46011318173259497, "kl": 0.0669543812982738, "context": 0.493590384721756}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 4.828794479370117, "loss_scale": 4.0, "retries": 0, "seconds": 53.0, "max_memory_gb": 18.27, "ce": 0.4551592916250229, "kl": 0.026720493158791214, "context": 0.46851953864097595}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 1.2568479776382446, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 18.27, "ce": 0.287545807659626, "kl": 0.03324792138300836, "context": 0.30416977033019066}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 2.300018310546875, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 18.27, "ce": 0.19047624245285988, "kl": 0.05071806076739449, "context": 0.2158352779224515}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 1.1289167404174805, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 18.27, "ce": 0.23005019314587116, "kl": 0.07537996012251824, "context": 0.2677401709370315}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 4.1182403564453125, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 18.27, "ce": 0.40486055612564087, "kl": 0.0708211527671665, "context": 0.44027112796902657}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 2.7486493587493896, "loss_scale": 4.0, "retries": 0, "seconds": 55.8, "max_memory_gb": 18.27, "ce": 0.3487996682524681, "kl": 0.3549562320113182, "context": 0.5262777954339981}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 4.711090087890625, "loss_scale": 4.0, "retries": 0, "seconds": 56.3, "max_memory_gb": 18.27, "ce": 0.5360232628881931, "kl": 0.1366055477410555, "context": 0.6043260172009468}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 1.5288912057876587, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 18.27, "ce": 0.15514220937620848, "kl": 0.04534657648764551, "context": 0.17781550018116832}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.038412570953369, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 18.27, "ce": 0.17608238174580038, "kl": 0.059340201172744855, "context": 0.20575248217210174}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 1.9887276887893677, "loss_scale": 4.0, "retries": 0, "seconds": 57.9, "max_memory_gb": 18.27, "ce": 0.29016633704304695, "kl": 0.18454635771922767, "context": 0.3824395090341568}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 1.9676318168640137, "loss_scale": 4.0, "retries": 0, "seconds": 58.5, "max_memory_gb": 18.27, "ce": 0.2527678832411766, "kl": 0.07872015796601772, "context": 0.29212796688079834}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.90774005651474, "loss_scale": 4.0, "retries": 0, "seconds": 59.0, "max_memory_gb": 18.27, "ce": 0.17766749113798141, "kl": 0.029372765333391726, "context": 0.19235387537628412}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.9433226585388184, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 18.27, "ce": 0.19942587381228805, "kl": 0.02466690569417551, "context": 0.2117593209259212}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 2.472810745239258, "loss_scale": 4.0, "retries": 0, "seconds": 60.1, "max_memory_gb": 18.27, "ce": 0.36278029158711433, "kl": 0.22791613650042564, "context": 0.4767383597791195}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 1.7056604623794556, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 18.27, "ce": 0.29824861511588097, "kl": 0.050531402695924044, "context": 0.32351431995630264}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 1.9673413038253784, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 18.27, "ce": 0.2347913400735706, "kl": 0.08602395202615298, "context": 0.2778033190406859}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 1.039703130722046, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 18.27, "ce": 0.1523988824337721, "kl": 0.01307673059636727, "context": 0.15893724747002125}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 1.1176447868347168, "loss_scale": 4.0, "retries": 0, "seconds": 62.2, "max_memory_gb": 18.27, "ce": 0.2059992514550686, "kl": 0.05739903077483177, "context": 0.23469876684248447}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 4.449512958526611, "loss_scale": 4.0, "retries": 0, "seconds": 62.8, "max_memory_gb": 18.27, "ce": 0.46578963845968246, "kl": 0.6899207262322307, "context": 0.8107500001788139}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.8840828537940979, "loss_scale": 4.0, "retries": 0, "seconds": 63.3, "max_memory_gb": 18.27, "ce": 0.26511120051145554, "kl": 0.18866315390914679, "context": 0.3594427816569805}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.3428514897823334, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 18.27, "ce": 0.12050719605758786, "kl": 0.01003299378498923, "context": 0.12552369572222233}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 3.7336435317993164, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 18.27, "ce": 0.3533056159503758, "kl": 0.029464719438692555, "context": 0.36803797306492925}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.5913958549499512, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 18.27, "ce": 0.17287345393560827, "kl": 0.03470964473672211, "context": 0.1902282778173685}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.9375137090682983, "loss_scale": 4.0, "retries": 0, "seconds": 65.5, "max_memory_gb": 18.27, "ce": 0.1390240271575749, "kl": 0.09065284684766084, "context": 0.1843504449352622}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 1.7702114582061768, "loss_scale": 4.0, "retries": 0, "seconds": 66.0, "max_memory_gb": 18.27, "ce": 0.3028053166344762, "kl": 0.08531555999070406, "context": 0.34546308824792504}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.461395263671875, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 18.27, "ce": 0.3155679020565003, "kl": 0.093285115348408, "context": 0.3622104513924569}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 1.0451695919036865, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 18.27, "ce": 0.1707947626709938, "kl": 0.0188785200152779, "context": 0.18023402290418744}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.900638222694397, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 18.27, "ce": 0.43279113015159965, "kl": 0.05281848809681833, "context": 0.4592003710567951}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 1.8051966428756714, "loss_scale": 4.0, "retries": 0, "seconds": 68.2, "max_memory_gb": 18.27, "ce": 0.3073861990123987, "kl": 0.04446363146416843, "context": 0.3296180162578821}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.7369602918624878, "loss_scale": 4.0, "retries": 0, "seconds": 68.7, "max_memory_gb": 18.27, "ce": 0.15775956166908145, "kl": 0.019333904841914773, "context": 0.16742650978267193}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 3.102926254272461, "loss_scale": 4.0, "retries": 0, "seconds": 69.2, "max_memory_gb": 18.27, "ce": 0.3820231668651104, "kl": 0.11471415939740837, "context": 0.4393802396953106}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.7089022994041443, "loss_scale": 4.0, "retries": 0, "seconds": 69.8, "max_memory_gb": 18.27, "ce": 0.06904018553905189, "kl": 0.004010320193629013, "context": 0.07104534422978759}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.193552017211914, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 18.27, "ce": 0.16640092711895704, "kl": 0.3295730888421531, "context": 0.33118748269043863}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 1.6725573539733887, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 18.27, "ce": 0.12042090110480785, "kl": 0.06387551012448967, "context": 0.1523586604744196}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 3.6897220611572266, "loss_scale": 4.0, "retries": 0, "seconds": 71.4, "max_memory_gb": 18.27, "ce": 0.3122415952384472, "kl": 0.37339873472228646, "context": 0.4989409688860178}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 1.7915693521499634, "loss_scale": 4.0, "retries": 0, "seconds": 71.9, "max_memory_gb": 18.27, "ce": 0.2803626134991646, "kl": 0.08119261637330055, "context": 0.32095892354846}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.6140691041946411, "loss_scale": 4.0, "retries": 0, "seconds": 72.5, "max_memory_gb": 18.27, "ce": 0.3062090203166008, "kl": 0.20018799387617037, "context": 0.40630301064811647}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 1.7286195755004883, "loss_scale": 4.0, "retries": 0, "seconds": 73.0, "max_memory_gb": 18.27, "ce": 0.22983487881720066, "kl": 0.14547022432088852, "context": 0.30256999284029007}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 1.2875639200210571, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 18.27, "ce": 0.1866238540969789, "kl": 0.01924882797175087, "context": 0.19624826894141734}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.045419692993164, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 18.27, "ce": 0.17283624596893787, "kl": 0.017797673237510026, "context": 0.18173508159816265}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 1.337147831916809, "loss_scale": 4.0, "retries": 0, "seconds": 74.6, "max_memory_gb": 18.27, "ce": 0.16459790989756584, "kl": 0.053446508012712, "context": 0.19132116436958313}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 1.1483254432678223, "loss_scale": 4.0, "retries": 0, "seconds": 75.1, "max_memory_gb": 18.27, "ce": 0.14809995936229825, "kl": 0.01915157272014767, "context": 0.1576757482253015}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 1.1148678064346313, "loss_scale": 4.0, "retries": 0, "seconds": 75.7, "max_memory_gb": 18.27, "ce": 0.10046674776822329, "kl": 0.07503685774281621, "context": 0.13798517733812332}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 6.405246734619141, "loss_scale": 4.0, "retries": 0, "seconds": 76.2, "max_memory_gb": 18.27, "ce": 0.24825645424425602, "kl": 0.1534025789005682, "context": 0.3249577395617962}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 1.1227439641952515, "loss_scale": 4.0, "retries": 0, "seconds": 76.8, "max_memory_gb": 18.27, "ce": 0.0819203145802021, "kl": 0.01695935265161097, "context": 0.09039999078959227}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.8575472831726074, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 18.27, "ce": 0.15709852147847414, "kl": 0.014898001565597951, "context": 0.1645475197583437}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.5003564357757568, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 18.27, "ce": 0.1321628550067544, "kl": 0.011639608477707952, "context": 0.13798265880905092}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.5016141533851624, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 18.27, "ce": 0.13479705224744976, "kl": 0.06092411620193161, "context": 0.16525911330245435}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 3.162001609802246, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 18.27, "ce": 0.20576429553329945, "kl": 0.16065931972116232, "context": 0.28609396144747734}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.33720842003822327, "loss_scale": 4.0, "retries": 0, "seconds": 79.5, "max_memory_gb": 18.27, "ce": 0.02571856533177197, "kl": 0.004012803090517991, "context": 0.027724966639652848}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 1.0279982089996338, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 18.27, "ce": 0.12240104237571359, "kl": 0.009813668089918792, "context": 0.1273078778758645}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.6098344326019287, "loss_scale": 4.0, "retries": 0, "seconds": 80.6, "max_memory_gb": 18.27, "ce": 0.21356428996659815, "kl": 0.44742421360570006, "context": 0.43727639492135495}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.6415915489196777, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 18.27, "ce": 0.022249258123338223, "kl": 0.05670885252766311, "context": 0.0506036845035851}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.8364565372467041, "loss_scale": 4.0, "retries": 0, "seconds": 81.6, "max_memory_gb": 18.27, "ce": 0.10364579386077821, "kl": 0.013066001047263853, "context": 0.11017879354767501}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 1.1390048265457153, "loss_scale": 4.0, "retries": 0, "seconds": 82.2, "max_memory_gb": 18.27, "ce": 0.18076832685619593, "kl": 0.033791077381465584, "context": 0.1976638687774539}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 1.6297955513000488, "loss_scale": 4.0, "retries": 0, "seconds": 82.7, "max_memory_gb": 18.27, "ce": 0.14432598720304668, "kl": 0.016880076727829874, "context": 0.1527660230640322}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 1.4583747386932373, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 18.27, "ce": 0.07238281844183803, "kl": 0.5345161808072589, "context": 0.3396409167908132}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 1.2348253726959229, "loss_scale": 4.0, "retries": 0, "seconds": 83.8, "max_memory_gb": 18.27, "ce": 0.08574983198195696, "kl": 0.3123267558403313, "context": 0.24191320315003395}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.171360969543457, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 18.27, "ce": 0.4516424057073891, "kl": 0.32365139116882347, "context": 0.6134681017138064}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 1.9591999053955078, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 18.27, "ce": 0.5308135161176324, "kl": 0.16240199308958836, "context": 0.6120145388413221}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.1911489963531494, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 18.27, "ce": 0.13841171795502305, "kl": 0.13588391672783473, "context": 0.20635367988143116}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.8642705082893372, "loss_scale": 4.0, "retries": 0, "seconds": 86.0, "max_memory_gb": 18.27, "ce": 0.14021307742223144, "kl": 0.049614881951129064, "context": 0.1650205166079104}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.2167679071426392, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 18.27, "ce": 0.25131495855748653, "kl": 0.27552137477323413, "context": 0.3890756443142891}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.7176409959793091, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 18.27, "ce": 0.061639465391635895, "kl": 0.07178685371764004, "context": 0.09753288887441158}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.8998632431030273, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 18.27, "ce": 0.10435651609441265, "kl": 0.06946066091768444, "context": 0.13908684300258756}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.7595582604408264, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 18.27, "ce": 0.22274068463593721, "kl": 0.04286016183323227, "context": 0.24417076539248228}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.4672553837299347, "loss_scale": 4.0, "retries": 0, "seconds": 88.7, "max_memory_gb": 18.27, "ce": 0.09200278576463461, "kl": 0.024913743138313293, "context": 0.10445965779945254}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 1.0424498319625854, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 18.27, "ce": 0.0677952989935875, "kl": 0.08369417945868918, "context": 0.1096423912094906}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 1.520188570022583, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 18.27, "ce": 0.2368917614221573, "kl": 0.023299435852095485, "context": 0.2485414817929268}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.7768880128860474, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 18.27, "ce": 0.12894302746281028, "kl": 0.05878130011842586, "context": 0.1583336777985096}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 1.2824703454971313, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 18.27, "ce": 0.15613108221441507, "kl": 0.05026351008564234, "context": 0.18126283772289753}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 1.068462610244751, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 18.27, "ce": 0.15427649673074484, "kl": 0.013430724386125803, "context": 0.16099185263738036}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.6846129298210144, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 18.27, "ce": 0.06816614721901715, "kl": 0.11047347821295261, "context": 0.1234028865583241}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 2.85980224609375, "loss_scale": 4.0, "retries": 0, "seconds": 92.5, "max_memory_gb": 18.27, "ce": 0.21528846956789494, "kl": 1.1068750666454434, "context": 0.7687259837985039}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.18330156803131104, "loss_scale": 4.0, "retries": 0, "seconds": 93.0, "max_memory_gb": 18.27, "ce": 0.018893915286753327, "kl": 0.006894621736137196, "context": 0.02234122605295852}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 1.9259201288223267, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 18.27, "ce": 0.03937287302687764, "kl": 0.20969522497034632, "context": 0.14422048453707248}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.6179949641227722, "loss_scale": 4.0, "retries": 0, "seconds": 94.1, "max_memory_gb": 18.27, "ce": 0.05527275800704956, "kl": 0.0088351956801489, "context": 0.059690356720238924}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 1.1574736833572388, "loss_scale": 4.0, "retries": 0, "seconds": 94.6, "max_memory_gb": 18.27, "ce": 0.10693247709423304, "kl": 0.07451780140399933, "context": 0.14419138012453914}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.9033143520355225, "loss_scale": 4.0, "retries": 0, "seconds": 95.1, "max_memory_gb": 18.27, "ce": 0.10260425880551338, "kl": 0.05646583653287962, "context": 0.13083717599511147}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 1.4407893419265747, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 18.27, "ce": 0.13256815820932388, "kl": 0.18706186302006245, "context": 0.22609909437596798}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.3724556565284729, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 18.27, "ce": 0.03103189868852496, "kl": 0.009182020497974008, "context": 0.035622910130769014}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.922215461730957, "loss_scale": 4.0, "retries": 0, "seconds": 96.7, "max_memory_gb": 18.27, "ce": 0.16474504838697612, "kl": 0.12700652939383872, "context": 0.22824831539765}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.4010976254940033, "loss_scale": 4.0, "retries": 0, "seconds": 97.3, "max_memory_gb": 18.27, "ce": 0.13646417181007564, "kl": 0.08837774145649746, "context": 0.18065304565243423}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.5884937047958374, "loss_scale": 4.0, "retries": 0, "seconds": 97.8, "max_memory_gb": 18.27, "ce": 0.3241334524936974, "kl": 0.0052914943371433765, "context": 0.3267792051192373}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 1.682346224784851, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 18.27, "ce": 0.29173649149015546, "kl": 0.06558422447415069, "context": 0.3245286112651229}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.8779621124267578, "loss_scale": 4.0, "retries": 0, "seconds": 98.9, "max_memory_gb": 18.27, "ce": 0.05643042642623186, "kl": 0.06005846429616213, "context": 0.08645966183394194}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 2.6357786655426025, "loss_scale": 4.0, "retries": 0, "seconds": 99.4, "max_memory_gb": 18.27, "ce": 0.24000838678330183, "kl": 0.009479500949964859, "context": 0.2447481278795749}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.11318846046924591, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 18.27, "ce": 0.007968273595906794, "kl": 0.0055583259963896126, "context": 0.010747436317615211}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.4859297275543213, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 18.27, "ce": 0.034213396487757564, "kl": 0.054314011067617685, "context": 0.061370401992462575}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.2843780517578125, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 18.27, "ce": 0.024334535642992705, "kl": 0.01918111101986142, "context": 0.03392509225523099}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.33393704891204834, "loss_scale": 4.0, "retries": 0, "seconds": 101.6, "max_memory_gb": 18.27, "ce": 0.05269862990826368, "kl": 0.00370490993373096, "context": 0.05455108662135899}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.9905128479003906, "loss_scale": 4.0, "retries": 0, "seconds": 102.1, "max_memory_gb": 18.27, "ce": 0.06051870831288397, "kl": 0.37434843451774213, "context": 0.24769293214194477}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.6020381450653076, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 18.27, "ce": 0.1316864985274151, "kl": 0.0512512976929429, "context": 0.15731214737752452}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.3853968679904938, "loss_scale": 4.0, "retries": 0, "seconds": 103.2, "max_memory_gb": 18.27, "ce": 0.03977008070796728, "kl": 0.006046614158549346, "context": 0.04279338754713535}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.3615370988845825, "loss_scale": 4.0, "retries": 0, "seconds": 103.7, "max_memory_gb": 18.27, "ce": 0.013416289817541838, "kl": 0.3824007372604683, "context": 0.2046166555956006}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 1.50946843624115, "loss_scale": 4.0, "retries": 0, "seconds": 104.2, "max_memory_gb": 18.27, "ce": 0.2712381179444492, "kl": 0.05438087380025536, "context": 0.2984285484999418}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.2959725558757782, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 18.27, "ce": 0.17545502143912017, "kl": 0.005233441606833367, "context": 0.17807174590416253}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.31803616881370544, "loss_scale": 4.0, "retries": 0, "seconds": 105.3, "max_memory_gb": 18.27, "ce": 0.029890983132645488, "kl": 0.03375163013697602, "context": 0.04676679871045053}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 2.371298313140869, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 18.27, "ce": 0.3568254257552326, "kl": 0.021497703157365322, "context": 0.36757426895201206}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.5532547235488892, "loss_scale": 4.0, "retries": 0, "seconds": 106.4, "max_memory_gb": 18.27, "ce": 0.10195718752220273, "kl": 0.02097570861224085, "context": 0.1124450396746397}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.38760271668434143, "loss_scale": 4.0, "retries": 0, "seconds": 106.9, "max_memory_gb": 18.27, "ce": 0.016901892027817667, "kl": 0.029720615275437012, "context": 0.03176219970919192}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.5063539147377014, "loss_scale": 4.0, "retries": 0, "seconds": 107.5, "max_memory_gb": 18.27, "ce": 0.08083097322378308, "kl": 0.026662284231861122, "context": 0.094162113731727}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.7779960036277771, "loss_scale": 4.0, "retries": 0, "seconds": 108.0, "max_memory_gb": 18.27, "ce": 0.1449711339082569, "kl": 0.03763363009784371, "context": 0.16378794866614044}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.7183777093887329, "loss_scale": 4.0, "retries": 0, "seconds": 108.5, "max_memory_gb": 18.27, "ce": 0.06198847712948918, "kl": 0.275765592494281, "context": 0.1998712751083076}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.7270234823226929, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 18.27, "ce": 0.22299440763890743, "kl": 0.014042770257219672, "context": 0.23001579800620675}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 1.0373765230178833, "loss_scale": 4.0, "retries": 0, "seconds": 109.6, "max_memory_gb": 18.27, "ce": 0.08102033328032121, "kl": 0.01935655897250399, "context": 0.09069861238822341}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.0902373269200325, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 18.27, "ce": 0.004847685224376619, "kl": 0.0021482417578226887, "context": 0.005921806208789349}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.8516091704368591, "loss_scale": 4.0, "retries": 0, "seconds": 110.7, "max_memory_gb": 18.27, "ce": 0.14594844984821975, "kl": 0.10886040702462196, "context": 0.2003786526620388}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.9585819244384766, "loss_scale": 4.0, "retries": 0, "seconds": 111.2, "max_memory_gb": 18.27, "ce": 0.11673334322404116, "kl": 0.020493449294008315, "context": 0.12698006868595257}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 1.0040624141693115, "loss_scale": 4.0, "retries": 0, "seconds": 111.8, "max_memory_gb": 18.27, "ce": 0.05533831715001725, "kl": 0.3698360526759643, "context": 0.2402563518844545}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.9245242476463318, "loss_scale": 4.0, "retries": 0, "seconds": 112.3, "max_memory_gb": 18.27, "ce": 0.11087389377644286, "kl": 0.10291522299667122, "context": 0.1623315048054792}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.758940577507019, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 18.27, "ce": 0.5894422540441155, "kl": 0.48781214299378917, "context": 0.8333483384922147}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.12893524765968323, "loss_scale": 4.0, "retries": 0, "seconds": 113.4, "max_memory_gb": 18.27, "ce": 0.008786197111476213, "kl": 0.00436470521526644, "context": 0.010968549991957843}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.45055091381073, "loss_scale": 4.0, "retries": 0, "seconds": 113.9, "max_memory_gb": 18.27, "ce": 0.1967141036875546, "kl": 0.021456575486809015, "context": 0.2074423902668059}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 1.1596418619155884, "loss_scale": 4.0, "retries": 0, "seconds": 114.5, "max_memory_gb": 18.27, "ce": 0.2252334791701287, "kl": 0.09687570261303335, "context": 0.2736713372869417}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.7370364665985107, "loss_scale": 4.0, "retries": 0, "seconds": 115.0, "max_memory_gb": 18.27, "ce": 0.18185274582356215, "kl": 0.0944755089294631, "context": 0.22909049573354423}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 1.1321278810501099, "loss_scale": 4.0, "retries": 0, "seconds": 115.6, "max_memory_gb": 18.27, "ce": 0.138224215246737, "kl": 0.7753488281741738, "context": 0.5258986242115498}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 1.9783804416656494, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 18.27, "ce": 0.34475608612410724, "kl": 0.06368847696285229, "context": 0.3766003284836188}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 1.033937931060791, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 18.27, "ce": 0.24667508981656283, "kl": 0.3053376052994281, "context": 0.3993438850156963}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 2.123635768890381, "loss_scale": 4.0, "retries": 0, "seconds": 117.2, "max_memory_gb": 18.27, "ce": 0.3820049201603979, "kl": 0.005706424017262179, "context": 0.38485813816078007}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.7783305644989014, "loss_scale": 4.0, "retries": 0, "seconds": 117.7, "max_memory_gb": 18.27, "ce": 0.1547446083277464, "kl": 0.07699066575150937, "context": 0.19323993823491037}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.7025606036186218, "loss_scale": 4.0, "retries": 0, "seconds": 118.3, "max_memory_gb": 18.27, "ce": 0.047049141372554004, "kl": 0.030262701357059996, "context": 0.062180492910556495}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 1.5090761184692383, "loss_scale": 4.0, "retries": 0, "seconds": 118.8, "max_memory_gb": 18.27, "ce": 0.18214742536656559, "kl": 0.4651445560157299, "context": 0.41471969801932573}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.61086505651474, "loss_scale": 4.0, "retries": 0, "seconds": 119.3, "max_memory_gb": 18.27, "ce": 0.11317693861201406, "kl": 0.033641582005657256, "context": 0.1299977283924818}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.8296230435371399, "loss_scale": 4.0, "retries": 0, "seconds": 119.9, "max_memory_gb": 18.27, "ce": 0.08279032935388386, "kl": 0.018339120244490914, "context": 0.0919598916079849}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.7756722569465637, "loss_scale": 4.0, "retries": 0, "seconds": 120.4, "max_memory_gb": 18.27, "ce": 0.08583884290419519, "kl": 0.1294235463719815, "context": 0.1505506169050932}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 4.110117435455322, "loss_scale": 4.0, "retries": 0, "seconds": 121.0, "max_memory_gb": 18.27, "ce": 0.2538722986355424, "kl": 0.5105878016183851, "context": 0.5091661997139454}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.7521055936813354, "loss_scale": 4.0, "retries": 0, "seconds": 121.5, "max_memory_gb": 18.27, "ce": 0.16234113671816885, "kl": 0.03361709014279768, "context": 0.1791496790247038}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 1.1576343774795532, "loss_scale": 4.0, "retries": 0, "seconds": 122.1, "max_memory_gb": 18.27, "ce": 0.07492788112722337, "kl": 0.13477078128198627, "context": 0.14231326756998897}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.8355916142463684, "loss_scale": 4.0, "retries": 0, "seconds": 122.6, "max_memory_gb": 18.27, "ce": 0.04336917609907687, "kl": 0.03853497072122991, "context": 0.06263666274026036}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 1.1491197347640991, "loss_scale": 4.0, "retries": 0, "seconds": 123.1, "max_memory_gb": 18.27, "ce": 0.07622514432296157, "kl": 0.017169718630611897, "context": 0.08481000224128366}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.16379064321517944, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 18.27, "ce": 0.011021020181942731, "kl": 0.00633978471159935, "context": 0.014190912479534745}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 1.782482385635376, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 18.27, "ce": 0.05739702144637704, "kl": 0.10981013905256987, "context": 0.11230209097266197}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 1.7153886556625366, "loss_scale": 4.0, "retries": 0, "seconds": 124.7, "max_memory_gb": 18.27, "ce": 0.2655076195951551, "kl": 0.17659910465590656, "context": 0.3538071643561125}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 1.148267149925232, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 18.27, "ce": 0.26370100560598075, "kl": 0.13006253517232835, "context": 0.3287322735413909}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 1.0218411684036255, "loss_scale": 4.0, "retries": 0, "seconds": 125.8, "max_memory_gb": 18.27, "ce": 0.25337498518638313, "kl": 0.0204470624885289, "context": 0.2635985226370394}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.2754001021385193, "loss_scale": 4.0, "retries": 0, "seconds": 126.3, "max_memory_gb": 18.27, "ce": 0.015788127318955958, "kl": 0.005739200074458495, "context": 0.018657726934179664}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.3391897678375244, "loss_scale": 4.0, "retries": 0, "seconds": 126.9, "max_memory_gb": 18.27, "ce": 0.12537333089858294, "kl": 0.13003906073026883, "context": 0.1903928592801094}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 1.263451099395752, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 18.27, "ce": 0.13309177896007895, "kl": 0.024946817429736257, "context": 0.14556519116740674}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 1.9165769815444946, "loss_scale": 4.0, "retries": 0, "seconds": 127.9, "max_memory_gb": 18.27, "ce": 0.34326622943626717, "kl": 0.037537899712333456, "context": 0.3620351916179061}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.8128596544265747, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 18.27, "ce": 0.07856150949373841, "kl": 0.03022470151336165, "context": 0.09367385937366635}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 1.1211398839950562, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 18.27, "ce": 0.1780388051411137, "kl": 0.06136838733209515, "context": 0.20872299419716}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.9573328495025635, "loss_scale": 4.0, "retries": 0, "seconds": 129.5, "max_memory_gb": 18.27, "ce": 0.03702476085163653, "kl": 0.13931799461352057, "context": 0.10668375756358728}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 1.409354567527771, "loss_scale": 4.0, "retries": 0, "seconds": 130.1, "max_memory_gb": 18.27, "ce": 0.15625964861828834, "kl": 0.028855846816441044, "context": 0.17068757535889745}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 1.1381977796554565, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 18.27, "ce": 0.09257214888930321, "kl": 0.03461581002920866, "context": 0.10988004971295595}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.7209874987602234, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 18.27, "ce": 0.15634799282997847, "kl": 0.017716048285365105, "context": 0.16520601511001587}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.9053629040718079, "loss_scale": 4.0, "retries": 0, "seconds": 131.7, "max_memory_gb": 18.27, "ce": 0.11836331791710109, "kl": 0.04876759948092513, "context": 0.14274711848702282}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 1.4664490222930908, "loss_scale": 4.0, "retries": 0, "seconds": 132.2, "max_memory_gb": 18.27, "ce": 0.0219593346118927, "kl": 0.16091331403004006, "context": 0.10241599520668387}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 1.9643288850784302, "loss_scale": 4.0, "retries": 0, "seconds": 132.8, "max_memory_gb": 18.27, "ce": 0.2867316579213366, "kl": 0.06375385566207115, "context": 0.3186085855704732}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 1.690477967262268, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 18.27, "ce": 0.40556356275919825, "kl": 0.01492776139639318, "context": 0.4130274448543787}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.4116933345794678, "loss_scale": 4.0, "retries": 0, "seconds": 133.8, "max_memory_gb": 18.27, "ce": 0.018502743390854448, "kl": 0.026049742335999326, "context": 0.03152761363890022}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.8218260407447815, "loss_scale": 4.0, "retries": 0, "seconds": 134.4, "max_memory_gb": 18.27, "ce": 0.07495753315743059, "kl": 0.038242558599449694, "context": 0.09407881158404052}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.7377862930297852, "loss_scale": 4.0, "retries": 0, "seconds": 134.9, "max_memory_gb": 18.27, "ce": 0.06782642658799887, "kl": 0.038689121953211725, "context": 0.0871709892526269}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.7146872878074646, "loss_scale": 4.0, "retries": 0, "seconds": 135.5, "max_memory_gb": 18.27, "ce": 0.06991536181885749, "kl": 0.03357025804143632, "context": 0.08670049184001982}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.5905564427375793, "loss_scale": 4.0, "retries": 0, "seconds": 136.0, "max_memory_gb": 18.27, "ce": 0.03714515594765544, "kl": 0.010892949609115021, "context": 0.04259163048118353}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.7237983345985413, "loss_scale": 4.0, "retries": 0, "seconds": 136.5, "max_memory_gb": 18.27, "ce": 0.12105125607922673, "kl": 0.019768840014876332, "context": 0.13093567651230842}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.8389970660209656, "loss_scale": 4.0, "retries": 0, "seconds": 137.1, "max_memory_gb": 18.27, "ce": 0.027609051088802516, "kl": 0.14835335276438855, "context": 0.10178573010489345}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 1.177032709121704, "loss_scale": 4.0, "retries": 0, "seconds": 137.6, "max_memory_gb": 18.27, "ce": 0.12372546608094126, "kl": 0.007000991216045804, "context": 0.1272259628167376}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 1.4015558958053589, "loss_scale": 4.0, "retries": 0, "seconds": 138.2, "max_memory_gb": 18.27, "ce": 0.13477999507449567, "kl": 0.283881991985254, "context": 0.2767209932208061}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 1.3107872009277344, "loss_scale": 4.0, "retries": 0, "seconds": 138.7, "max_memory_gb": 18.27, "ce": 0.12268446572124958, "kl": 0.2823839355260134, "context": 0.26387643069028854}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.8099509477615356, "loss_scale": 4.0, "retries": 0, "seconds": 139.2, "max_memory_gb": 18.27, "ce": 0.06657696806360036, "kl": 0.0299313853665808, "context": 0.08154266385827214}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.8935452699661255, "loss_scale": 4.0, "retries": 0, "seconds": 139.8, "max_memory_gb": 18.27, "ce": 0.06490737493732013, "kl": 0.015286608308088034, "context": 0.07255068018275779}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.8395239114761353, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 18.27, "ce": 0.047132110921666026, "kl": 0.13161121850134805, "context": 0.11293772317003459}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.40651875734329224, "loss_scale": 4.0, "retries": 0, "seconds": 140.9, "max_memory_gb": 18.27, "ce": 0.010087470058351755, "kl": 0.034320934893912636, "context": 0.02724793751258403}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 1.035726547241211, "loss_scale": 4.0, "retries": 0, "seconds": 141.4, "max_memory_gb": 18.27, "ce": 0.06258880393579602, "kl": 0.13025433564325795, "context": 0.12771597504615784}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 1.056388020515442, "loss_scale": 4.0, "retries": 0, "seconds": 141.9, "max_memory_gb": 18.27, "ce": 0.13104803184978664, "kl": 0.08780789095908403, "context": 0.17495197989046574}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 1.6960959434509277, "loss_scale": 4.0, "retries": 0, "seconds": 142.5, "max_memory_gb": 18.27, "ce": 0.18807061878032982, "kl": 0.1324387268105056, "context": 0.25428998190909624}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 1.762035846710205, "loss_scale": 4.0, "retries": 0, "seconds": 143.0, "max_memory_gb": 18.27, "ce": 0.14387820288538933, "kl": 0.057814940926618874, "context": 0.17278567608445883}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 1.9857313632965088, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 18.27, "ce": 0.13652389700291678, "kl": 0.07131201076845173, "context": 0.172179906105157}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.5747493505477905, "loss_scale": 4.0, "retries": 0, "seconds": 144.1, "max_memory_gb": 18.27, "ce": 0.0729562419001013, "kl": 0.09330397500889376, "context": 0.11960822762921453}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 2.918571710586548, "loss_scale": 4.0, "retries": 0, "seconds": 144.6, "max_memory_gb": 18.27, "ce": 0.4076607036986388, "kl": 0.16196142547414638, "context": 0.48864142887759954}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.8680586218833923, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 18.27, "ce": 0.15256826754193753, "kl": 0.004782984149642289, "context": 0.1549597557168454}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.8949910402297974, "loss_scale": 4.0, "retries": 0, "seconds": 145.7, "max_memory_gb": 18.27, "ce": 0.23211256274953485, "kl": 0.08359353878768161, "context": 0.2739093331620097}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.51489782333374, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 18.27, "ce": 0.37188689410686493, "kl": 0.7171276465596748, "context": 0.7304507121443748}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 1.5929090976715088, "loss_scale": 4.0, "retries": 0, "seconds": 146.8, "max_memory_gb": 18.27, "ce": 0.1781716364203021, "kl": 0.004029223040561192, "context": 0.18018624582327902}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.3329264223575592, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 18.27, "ce": 0.02586688648443669, "kl": 0.04865361112024402, "context": 0.050193691975437105}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 1.6199181079864502, "loss_scale": 4.0, "retries": 0, "seconds": 147.9, "max_memory_gb": 18.27, "ce": 0.012568939127959311, "kl": 0.44997088858508505, "context": 0.23755438008811325}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.922319769859314, "loss_scale": 4.0, "retries": 0, "seconds": 148.4, "max_memory_gb": 18.27, "ce": 0.015536855265963823, "kl": 0.16536758722213563, "context": 0.09822064620675519}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.5918371081352234, "loss_scale": 4.0, "retries": 0, "seconds": 148.9, "max_memory_gb": 18.27, "ce": 0.1028530653566122, "kl": 0.04146765265613794, "context": 0.12358689168468118}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 2.5933570861816406, "loss_scale": 4.0, "retries": 0, "seconds": 149.5, "max_memory_gb": 18.27, "ce": 0.5518586413236335, "kl": 0.4032446686178446, "context": 0.7534809772623703}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.20347358286380768, "loss_scale": 4.0, "retries": 0, "seconds": 150.0, "max_memory_gb": 18.27, "ce": 0.0074073984869755805, "kl": 0.010099071716467733, "context": 0.01245693422970362}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.6555309295654297, "loss_scale": 4.0, "retries": 0, "seconds": 150.6, "max_memory_gb": 18.27, "ce": 0.008829446451272815, "kl": 0.13955109403468668, "context": 0.07860499480739236}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 2.049729108810425, "loss_scale": 4.0, "retries": 0, "seconds": 151.1, "max_memory_gb": 18.27, "ce": 0.15773942531086504, "kl": 0.2499726284877397, "context": 0.28272574025322683}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 1.4921152591705322, "loss_scale": 4.0, "retries": 0, "seconds": 151.6, "max_memory_gb": 18.27, "ce": 0.06861267681233585, "kl": 0.12038337958438206, "context": 0.12880436750128865}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.35846877098083496, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 18.27, "ce": 0.020330147817730904, "kl": 0.012103136177756824, "context": 0.026381715899333358}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 4.178472518920898, "loss_scale": 4.0, "retries": 0, "seconds": 152.7, "max_memory_gb": 18.27, "ce": 0.19682622468098998, "kl": 0.09921902843052521, "context": 0.24643574468791485}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.837547242641449, "loss_scale": 4.0, "retries": 0, "seconds": 153.3, "max_memory_gb": 18.27, "ce": 0.07691209514450748, "kl": 0.008990033961708832, "context": 0.08140711003215984}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.8109620809555054, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 18.27, "ce": 0.05659930733963847, "kl": 0.18572628188849194, "context": 0.14946244843304157}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 1.5309948921203613, "loss_scale": 4.0, "retries": 0, "seconds": 154.3, "max_memory_gb": 18.27, "ce": 0.07073504576692358, "kl": 0.011538099293829873, "context": 0.07650409633060917}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.7564228177070618, "loss_scale": 4.0, "retries": 0, "seconds": 154.9, "max_memory_gb": 18.27, "ce": 0.1101901747751981, "kl": 0.01704729930497706, "context": 0.11871382058598101}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.34408038854599, "loss_scale": 4.0, "retries": 0, "seconds": 155.4, "max_memory_gb": 18.27, "ce": 0.01174092828296125, "kl": 0.014629423727456015, "context": 0.01905563974287361}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 1.456271767616272, "loss_scale": 4.0, "retries": 0, "seconds": 156.0, "max_memory_gb": 18.27, "ce": 0.09605926205404103, "kl": 0.03974268911406398, "context": 0.1159306070767343}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.17779268324375153, "loss_scale": 4.0, "retries": 0, "seconds": 156.5, "max_memory_gb": 18.27, "ce": 0.006822607945650816, "kl": 0.007834791096684057, "context": 0.010740003315731883}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.45929062366485596, "loss_scale": 4.0, "retries": 0, "seconds": 157.0, "max_memory_gb": 18.27, "ce": 0.03437817271333188, "kl": 0.07385176615207456, "context": 0.07130405365023762}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 1.4296718835830688, "loss_scale": 4.0, "retries": 0, "seconds": 157.6, "max_memory_gb": 18.27, "ce": 0.11277430359041318, "kl": 0.05729859482107713, "context": 0.14142360669211484}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.10012968629598618, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 18.27, "ce": 0.00478737335652113, "kl": 0.0022298104963738297, "context": 0.005902278644498438}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.777550458908081, "loss_scale": 4.0, "retries": 0, "seconds": 158.6, "max_memory_gb": 18.27, "ce": 0.05267486197408289, "kl": 0.7516861818730831, "context": 0.4285179548896849}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 1.2180935144424438, "loss_scale": 4.0, "retries": 0, "seconds": 159.2, "max_memory_gb": 18.27, "ce": 0.05026331765111536, "kl": 0.15612086243345402, "context": 0.12832374987192452}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 1.246163249015808, "loss_scale": 4.0, "retries": 0, "seconds": 159.7, "max_memory_gb": 18.27, "ce": 0.05173407681286335, "kl": 0.016922767157666385, "context": 0.06019546091556549}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 2.5857701301574707, "loss_scale": 4.0, "retries": 0, "seconds": 160.2, "max_memory_gb": 18.27, "ce": 0.1355925714597106, "kl": 0.09879823612573091, "context": 0.18499169335700572}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 3.3893887996673584, "loss_scale": 4.0, "retries": 0, "seconds": 160.8, "max_memory_gb": 18.27, "ce": 0.25757170165888965, "kl": 0.45304423331981525, "context": 0.48409381695091724}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.8973795175552368, "loss_scale": 4.0, "retries": 0, "seconds": 161.3, "max_memory_gb": 18.27, "ce": 0.10482552426401526, "kl": 0.011206295777810737, "context": 0.1104286735644564}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.6095316410064697, "loss_scale": 4.0, "retries": 0, "seconds": 161.8, "max_memory_gb": 18.27, "ce": 0.4643040479859337, "kl": 0.1666064270248171, "context": 0.5476072723977268}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.6307345628738403, "loss_scale": 4.0, "retries": 0, "seconds": 162.4, "max_memory_gb": 18.27, "ce": 0.01162438135361299, "kl": 0.13582801958546042, "context": 0.07953838922549039}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 1.3895097970962524, "loss_scale": 4.0, "retries": 0, "seconds": 162.9, "max_memory_gb": 18.27, "ce": 0.2727028255467303, "kl": 0.052126167891401565, "context": 0.2987659090431407}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 1.9439579248428345, "loss_scale": 4.0, "retries": 0, "seconds": 163.5, "max_memory_gb": 18.27, "ce": 0.12460789893520996, "kl": 0.13066104386962252, "context": 0.1899384242715314}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.3547251224517822, "loss_scale": 4.0, "retries": 0, "seconds": 164.0, "max_memory_gb": 18.27, "ce": 0.02170975413173437, "kl": 0.009290390076785116, "context": 0.02635494922287762}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.6411139965057373, "loss_scale": 4.0, "retries": 0, "seconds": 164.5, "max_memory_gb": 18.27, "ce": 0.028756975225405768, "kl": 0.04070900212536799, "context": 0.04911147718667053}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.36742761731147766, "loss_scale": 4.0, "retries": 0, "seconds": 165.1, "max_memory_gb": 18.27, "ce": 0.016376241692341864, "kl": 0.0046199257267289795, "context": 0.018686204683035612}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 4.5837082862854, "loss_scale": 4.0, "retries": 0, "seconds": 165.6, "max_memory_gb": 18.27, "ce": 0.12293884460814297, "kl": 0.18614506252924912, "context": 0.2160113683785312}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 1.9186280965805054, "loss_scale": 4.0, "retries": 0, "seconds": 166.1, "max_memory_gb": 18.27, "ce": 0.008420733152888715, "kl": 0.1742390120634809, "context": 0.09554024017415941}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 1.945684790611267, "loss_scale": 4.0, "retries": 0, "seconds": 166.7, "max_memory_gb": 18.27, "ce": 0.12606451264582574, "kl": 0.013734121617744677, "context": 0.13293157215230167}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 1.458121418952942, "loss_scale": 4.0, "retries": 0, "seconds": 167.2, "max_memory_gb": 18.27, "ce": 0.023277400119695812, "kl": 0.6634760596934939, "context": 0.35501542658312246}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 2.235992670059204, "loss_scale": 4.0, "retries": 0, "seconds": 167.7, "max_memory_gb": 18.27, "ce": 0.22358522305148654, "kl": 0.02403216145467013, "context": 0.23560130459372886}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.6494370698928833, "loss_scale": 4.0, "retries": 0, "seconds": 168.3, "max_memory_gb": 18.27, "ce": 0.030839846760500222, "kl": 0.25489400583319366, "context": 0.15828684996813536}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 1.6347746849060059, "loss_scale": 4.0, "retries": 0, "seconds": 168.8, "max_memory_gb": 18.27, "ce": 0.15457637811778113, "kl": 0.020120608758588787, "context": 0.16463668324286118}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.35706302523612976, "loss_scale": 4.0, "retries": 0, "seconds": 169.4, "max_memory_gb": 18.27, "ce": 0.01922017417382449, "kl": 0.010014892468461767, "context": 0.024227619636803865}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 1.7258448600769043, "loss_scale": 4.0, "retries": 0, "seconds": 169.9, "max_memory_gb": 18.27, "ce": 0.14268425371847115, "kl": 0.0420134597516153, "context": 0.16369098451104946}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 1.433876395225525, "loss_scale": 4.0, "retries": 0, "seconds": 170.4, "max_memory_gb": 18.27, "ce": 0.1540855693165213, "kl": 0.0931537717115134, "context": 0.20066246250644326}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.6230528354644775, "loss_scale": 4.0, "retries": 0, "seconds": 171.0, "max_memory_gb": 18.27, "ce": 0.05236389598576352, "kl": 0.0531218396499753, "context": 0.0789248135406524}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 1.859360933303833, "loss_scale": 4.0, "retries": 0, "seconds": 171.5, "max_memory_gb": 18.27, "ce": 0.19075387122575194, "kl": 0.6635956313693896, "context": 0.5225516790524125}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 1.6854536533355713, "loss_scale": 4.0, "retries": 0, "seconds": 172.1, "max_memory_gb": 18.27, "ce": 0.14005987020209432, "kl": 0.020671596052125096, "context": 0.15039566531777382}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 1.0934675931930542, "loss_scale": 4.0, "retries": 0, "seconds": 172.6, "max_memory_gb": 18.27, "ce": 0.03318998822942376, "kl": 0.40418884235987207, "context": 0.23528440622612834}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 1.4068986177444458, "loss_scale": 4.0, "retries": 0, "seconds": 173.1, "max_memory_gb": 18.27, "ce": 0.07186945679131895, "kl": 0.10337087980587967, "context": 0.12355489656329155}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.4654649496078491, "loss_scale": 4.0, "retries": 0, "seconds": 173.7, "max_memory_gb": 18.27, "ce": 0.13884132204111665, "kl": 0.09231135572190396, "context": 0.18499699793756008}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 1.2043445110321045, "loss_scale": 4.0, "retries": 0, "seconds": 174.2, "max_memory_gb": 18.27, "ce": 0.07734824228100479, "kl": 0.070577072823653, "context": 0.1126367764081806}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.5348256826400757, "loss_scale": 4.0, "retries": 0, "seconds": 174.7, "max_memory_gb": 18.27, "ce": 0.017170653794892132, "kl": 0.04243401484563947, "context": 0.03838766156695783}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 1.4677157402038574, "loss_scale": 4.0, "retries": 0, "seconds": 175.3, "max_memory_gb": 18.27, "ce": 0.00886768673080951, "kl": 0.16682547132950276, "context": 0.09228042303584516}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.3039880096912384, "loss_scale": 4.0, "retries": 0, "seconds": 175.8, "max_memory_gb": 18.27, "ce": 0.017947708838619292, "kl": 0.010906121533480473, "context": 0.023400768637657166}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 1.2049033641815186, "loss_scale": 4.0, "retries": 0, "seconds": 176.3, "max_memory_gb": 18.27, "ce": 0.22261770529439673, "kl": 0.029093889643263537, "context": 0.2371646584942937}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.919475257396698, "loss_scale": 4.0, "retries": 0, "seconds": 176.9, "max_memory_gb": 18.27, "ce": 0.05891689594136551, "kl": 0.06606116318289423, "context": 0.09194747940637171}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 4.368420124053955, "loss_scale": 4.0, "retries": 0, "seconds": 177.4, "max_memory_gb": 18.27, "ce": 0.24795467796502635, "kl": 0.11319563469442073, "context": 0.30455249099759385}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 1.2073978185653687, "loss_scale": 4.0, "retries": 0, "seconds": 178.0, "max_memory_gb": 18.27, "ce": 0.10751434479607269, "kl": 0.15198396773485, "context": 0.18350632744841278}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 1.2576980590820312, "loss_scale": 4.0, "retries": 0, "seconds": 178.5, "max_memory_gb": 18.27, "ce": 0.10638430726248771, "kl": 0.17995919168606633, "context": 0.1963639030000195}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.8577247262001038, "loss_scale": 4.0, "retries": 0, "seconds": 179.0, "max_memory_gb": 18.27, "ce": 0.16958582791266963, "kl": 0.007535524742706912, "context": 0.17335359356366098}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 2.6347055435180664, "loss_scale": 4.0, "retries": 0, "seconds": 179.6, "max_memory_gb": 18.27, "ce": 0.2803071979433298, "kl": 0.10052595031447709, "context": 0.3305701771751046}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.92427659034729, "loss_scale": 4.0, "retries": 0, "seconds": 180.1, "max_memory_gb": 18.27, "ce": 0.2244308905210346, "kl": 0.04680804084637202, "context": 0.24783490761183202}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 1.816411018371582, "loss_scale": 4.0, "retries": 0, "seconds": 180.7, "max_memory_gb": 18.27, "ce": 0.2887513304594904, "kl": 0.0241523461882025, "context": 0.3008275097236037}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 1.6963273286819458, "loss_scale": 4.0, "retries": 0, "seconds": 181.2, "max_memory_gb": 18.27, "ce": 0.17893402522895485, "kl": 0.15889189910376444, "context": 0.2583799664862454}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 3.836392641067505, "loss_scale": 4.0, "retries": 0, "seconds": 181.7, "max_memory_gb": 18.27, "ce": 0.3555849192198366, "kl": 0.2888507314492017, "context": 0.5000102877384052}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.4021739363670349, "loss_scale": 4.0, "retries": 0, "seconds": 182.3, "max_memory_gb": 18.27, "ce": 0.011821610270999372, "kl": 0.06261023135812138, "context": 0.04312672535888851}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 1.4116337299346924, "loss_scale": 4.0, "retries": 0, "seconds": 182.8, "max_memory_gb": 18.27, "ce": 0.18008266296237707, "kl": 0.08921977016143501, "context": 0.22469254489988089}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.504645824432373, "loss_scale": 4.0, "retries": 0, "seconds": 183.4, "max_memory_gb": 18.27, "ce": 0.21519182808697224, "kl": 0.17147478251717985, "context": 0.30092922737821937}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 1.3931608200073242, "loss_scale": 4.0, "retries": 0, "seconds": 183.9, "max_memory_gb": 18.27, "ce": 0.15267176262568682, "kl": 0.010779444564832374, "context": 0.15806149097625166}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 1.6362930536270142, "loss_scale": 4.0, "retries": 0, "seconds": 184.4, "max_memory_gb": 18.27, "ce": 0.227413481567055, "kl": 0.10065960104111582, "context": 0.27774327411316335}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 1.6872657537460327, "loss_scale": 4.0, "retries": 0, "seconds": 185.0, "max_memory_gb": 18.27, "ce": 0.13777453685179353, "kl": 0.029111874289810658, "context": 0.15233047353103757}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 2.530500888824463, "loss_scale": 4.0, "retries": 0, "seconds": 185.5, "max_memory_gb": 18.27, "ce": 0.2602877639583312, "kl": 0.9366692579351366, "context": 0.7286223880946636}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.4482686221599579, "loss_scale": 4.0, "retries": 0, "seconds": 186.1, "max_memory_gb": 18.27, "ce": 0.01396346208639443, "kl": 0.0546223018609453, "context": 0.04127461276948452}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.45992109179496765, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 18.27, "ce": 0.020820312900468707, "kl": 0.025666242523584515, "context": 0.033653434133157134}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.7541167736053467, "loss_scale": 4.0, "retries": 0, "seconds": 187.1, "max_memory_gb": 18.27, "ce": 0.13928032870171592, "kl": 0.18581626116429106, "context": 0.232188455585856}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 1.739552617073059, "loss_scale": 4.0, "retries": 0, "seconds": 187.7, "max_memory_gb": 18.27, "ce": 0.19528344686841592, "kl": 0.005306920735165477, "context": 0.19793691113591194}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 1.6074309349060059, "loss_scale": 4.0, "retries": 0, "seconds": 188.2, "max_memory_gb": 18.27, "ce": 0.2033118698745966, "kl": 0.027702058316208422, "context": 0.2171628954820335}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.18577948212623596, "loss_scale": 4.0, "retries": 0, "seconds": 188.7, "max_memory_gb": 18.27, "ce": 0.009714528685435653, "kl": 0.005118993682117434, "context": 0.012274025415536016}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.23183783888816833, "loss_scale": 4.0, "retries": 0, "seconds": 189.3, "max_memory_gb": 18.27, "ce": 0.09986394009320065, "kl": 0.028594052895641653, "context": 0.11416096845641732}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 1.0465309619903564, "loss_scale": 4.0, "retries": 0, "seconds": 189.8, "max_memory_gb": 18.27, "ce": 0.14418074302375317, "kl": 0.05957173700153362, "context": 0.17396660894155502}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.5670608282089233, "loss_scale": 4.0, "retries": 0, "seconds": 190.3, "max_memory_gb": 18.27, "ce": 0.03758130641654134, "kl": 0.39845394949952606, "context": 0.23680827650241554}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 3.571406126022339, "loss_scale": 4.0, "retries": 0, "seconds": 190.9, "max_memory_gb": 18.27, "ce": 0.1761488050688058, "kl": 0.3908702414482832, "context": 0.371583916246891}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.9650371670722961, "loss_scale": 4.0, "retries": 0, "seconds": 191.4, "max_memory_gb": 18.27, "ce": 0.1408182339509949, "kl": 0.06746369964093901, "context": 0.17455008742399514}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.6129516363143921, "loss_scale": 4.0, "retries": 0, "seconds": 191.9, "max_memory_gb": 18.27, "ce": 0.17347999627236277, "kl": 0.02621165206801379, "context": 0.18658582668285817}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.7327644228935242, "loss_scale": 4.0, "retries": 0, "seconds": 192.5, "max_memory_gb": 18.27, "ce": 0.15076474389934447, "kl": 0.02655526942544384, "context": 0.16404237621463835}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.8988538980484009, "loss_scale": 4.0, "retries": 0, "seconds": 193.0, "max_memory_gb": 18.27, "ce": 0.04112294199876487, "kl": 0.037675519204640295, "context": 0.059960703481920063}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 1.07438325881958, "loss_scale": 4.0, "retries": 0, "seconds": 193.6, "max_memory_gb": 18.27, "ce": 0.058350102975964546, "kl": 0.020424537171493284, "context": 0.06856237101601437}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 1.3525983095169067, "loss_scale": 4.0, "retries": 0, "seconds": 194.1, "max_memory_gb": 18.27, "ce": 0.19987583858892322, "kl": 0.014192771515809, "context": 0.20697222743183374}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.35641178488731384, "loss_scale": 4.0, "retries": 0, "seconds": 194.6, "max_memory_gb": 18.27, "ce": 0.021503457566723228, "kl": 0.020932350234943442, "context": 0.03196963371010497}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 1.2144279479980469, "loss_scale": 4.0, "retries": 0, "seconds": 195.2, "max_memory_gb": 18.27, "ce": 0.07852859763079323, "kl": 0.014419722603634, "context": 0.085738456225954}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.6310608386993408, "loss_scale": 4.0, "retries": 0, "seconds": 195.7, "max_memory_gb": 18.27, "ce": 0.05946476093959063, "kl": 0.4766940544177487, "context": 0.2978117811726406}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 2.783799648284912, "loss_scale": 4.0, "retries": 0, "seconds": 196.2, "max_memory_gb": 18.27, "ce": 0.34166780006489716, "kl": 0.05133456305702566, "context": 0.36733508415636607}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.6659080982208252, "loss_scale": 4.0, "retries": 0, "seconds": 196.8, "max_memory_gb": 18.27, "ce": 0.050013413187116385, "kl": 0.10908100567758083, "context": 0.10455391742289066}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 2.6098434925079346, "loss_scale": 4.0, "retries": 0, "seconds": 197.3, "max_memory_gb": 18.27, "ce": 0.011148431993206032, "kl": 0.6632454172158759, "context": 0.3427711417025421}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 1.0594772100448608, "loss_scale": 4.0, "retries": 0, "seconds": 197.9, "max_memory_gb": 18.27, "ce": 0.13930927315959707, "kl": 0.02151975808374118, "context": 0.15006915456615388}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.8561943769454956, "loss_scale": 4.0, "retries": 0, "seconds": 198.4, "max_memory_gb": 18.27, "ce": 0.04559430759400129, "kl": 0.2985337667632848, "context": 0.19486118480563164}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 1.4505263566970825, "loss_scale": 4.0, "retries": 0, "seconds": 198.9, "max_memory_gb": 18.27, "ce": 0.1492713022744283, "kl": 0.11310657741523755, "context": 0.20582459110300988}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.38179346919059753, "loss_scale": 4.0, "retries": 0, "seconds": 199.4, "max_memory_gb": 18.27, "ce": 0.017125885700806975, "kl": 0.037295684077435, "context": 0.03577372874133289}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.4710003733634949, "loss_scale": 4.0, "retries": 0, "seconds": 200.0, "max_memory_gb": 18.27, "ce": 0.032500629604328424, "kl": 0.016270420357614057, "context": 0.04063583997776732}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 3.0436670780181885, "loss_scale": 4.0, "retries": 0, "seconds": 200.5, "max_memory_gb": 18.27, "ce": 0.3460631165653467, "kl": 0.17620921356137842, "context": 0.4341677110642195}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.162925362586975, "loss_scale": 4.0, "retries": 0, "seconds": 201.1, "max_memory_gb": 18.27, "ce": 0.1849655381811317, "kl": 0.02281261437201465, "context": 0.19637184910243377}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.5105766654014587, "loss_scale": 4.0, "retries": 0, "seconds": 201.6, "max_memory_gb": 18.27, "ce": 0.03247478662524372, "kl": 0.019543267466360703, "context": 0.04224642016924918}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.6534465551376343, "loss_scale": 4.0, "retries": 0, "seconds": 202.1, "max_memory_gb": 18.27, "ce": 0.07902917033061385, "kl": 0.07417727465508506, "context": 0.11611780873499811}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 1.552009105682373, "loss_scale": 4.0, "retries": 0, "seconds": 202.7, "max_memory_gb": 18.27, "ce": 0.14213708584429696, "kl": 0.13019615902521764, "context": 0.20723515807185322}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 1.0131056308746338, "loss_scale": 4.0, "retries": 0, "seconds": 203.2, "max_memory_gb": 18.27, "ce": 0.18294148554559797, "kl": 0.18898693091978203, "context": 0.27743495197501034}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 1.0952506065368652, "loss_scale": 4.0, "retries": 0, "seconds": 203.7, "max_memory_gb": 18.27, "ce": 0.05298918858170509, "kl": 0.04235864542715717, "context": 0.07416851131711155}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.9875941276550293, "loss_scale": 4.0, "retries": 0, "seconds": 204.3, "max_memory_gb": 18.27, "ce": 0.0528663310687989, "kl": 0.1623226644587703, "context": 0.13402766559738666}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 1.11777663230896, "loss_scale": 4.0, "retries": 0, "seconds": 204.8, "max_memory_gb": 18.27, "ce": 0.07798918848857284, "kl": 0.037880147458054125, "context": 0.09692926425486803}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.4480985403060913, "loss_scale": 4.0, "retries": 0, "seconds": 205.3, "max_memory_gb": 18.27, "ce": 0.06311797874514014, "kl": 0.06217573146568611, "context": 0.09420583990868181}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 3.140845537185669, "loss_scale": 4.0, "retries": 0, "seconds": 205.9, "max_memory_gb": 18.27, "ce": 0.026550367183517665, "kl": 0.12713977532985155, "context": 0.09012025874108076}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.8189522624015808, "loss_scale": 4.0, "retries": 0, "seconds": 206.4, "max_memory_gb": 18.27, "ce": 0.2382192036602646, "kl": 0.0905292205279693, "context": 0.28348380932584405}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 3.525942802429199, "loss_scale": 4.0, "retries": 0, "seconds": 206.9, "max_memory_gb": 18.27, "ce": 0.3926349488319829, "kl": 0.31866084126522765, "context": 0.5519653734518215}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.7358250617980957, "loss_scale": 4.0, "retries": 0, "seconds": 207.5, "max_memory_gb": 18.27, "ce": 0.07958218973362818, "kl": 0.07015949989727233, "context": 0.11466193990781903}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.40925493836402893, "loss_scale": 4.0, "retries": 0, "seconds": 208.0, "max_memory_gb": 18.27, "ce": 0.010104605054948479, "kl": 0.03617165226023644, "context": 0.028190431417897344}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.6630957722663879, "loss_scale": 4.0, "retries": 0, "seconds": 208.6, "max_memory_gb": 18.27, "ce": 0.15309702791273594, "kl": 0.0026220776926493272, "context": 0.15440806577680632}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.7692887783050537, "loss_scale": 4.0, "retries": 0, "seconds": 209.1, "max_memory_gb": 18.27, "ce": 0.04934957984369248, "kl": 0.006527634170197416, "context": 0.05261339817661792}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.845862865447998, "loss_scale": 4.0, "retries": 0, "seconds": 209.6, "max_memory_gb": 18.27, "ce": 0.0821424379828386, "kl": 0.01930888430797495, "context": 0.09179688201402314}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.3858861029148102, "loss_scale": 4.0, "retries": 0, "seconds": 210.2, "max_memory_gb": 18.27, "ce": 0.03268632508115843, "kl": 0.007376135486993007, "context": 0.036374392424477264}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.9377836585044861, "loss_scale": 4.0, "retries": 0, "seconds": 210.7, "max_memory_gb": 18.27, "ce": 0.04520776169374585, "kl": 0.03945762463263236, "context": 0.06493657315149903}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 1.408658742904663, "loss_scale": 4.0, "retries": 0, "seconds": 211.3, "max_memory_gb": 18.27, "ce": 0.06948424153961241, "kl": 0.05076958704739809, "context": 0.0948690352961421}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 1.610762357711792, "loss_scale": 4.0, "retries": 0, "seconds": 211.8, "max_memory_gb": 18.27, "ce": 0.19242292642593384, "kl": 0.1981559422565624, "context": 0.29150089621543884}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.1349283754825592, "loss_scale": 4.0, "retries": 0, "seconds": 212.3, "max_memory_gb": 18.27, "ce": 0.005303243393427692, "kl": 0.005463108347612433, "context": 0.008034797574509867}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 1.4331918954849243, "loss_scale": 4.0, "retries": 0, "seconds": 212.9, "max_memory_gb": 18.27, "ce": 0.025259918649680912, "kl": 0.768647852819413, "context": 0.40958383586257696}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.7486323714256287, "loss_scale": 4.0, "retries": 0, "seconds": 213.4, "max_memory_gb": 18.27, "ce": 0.0841116335359402, "kl": 0.03239877033047378, "context": 0.10031101619824767}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 1.8169350624084473, "loss_scale": 4.0, "retries": 0, "seconds": 213.9, "max_memory_gb": 18.27, "ce": 0.05609265324892476, "kl": 0.025803052500123158, "context": 0.06899417762178928}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 2.3999319076538086, "loss_scale": 4.0, "retries": 0, "seconds": 214.5, "max_memory_gb": 18.27, "ce": 0.013969383435323834, "kl": 0.34290158598378184, "context": 0.1854201753740199}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.9335669875144958, "loss_scale": 4.0, "retries": 0, "seconds": 215.0, "max_memory_gb": 18.27, "ce": 0.10434896044898778, "kl": 0.08929129358148202, "context": 0.14899460633751005}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 1.268966794013977, "loss_scale": 4.0, "retries": 0, "seconds": 215.5, "max_memory_gb": 18.27, "ce": 0.09946869988925755, "kl": 0.025785139223444276, "context": 0.11236127279698849}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 1.9955459833145142, "loss_scale": 4.0, "retries": 0, "seconds": 216.1, "max_memory_gb": 18.27, "ce": 0.18990063371893484, "kl": 0.0010401451127108885, "context": 0.19042070885188878}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 1.4523561000823975, "loss_scale": 4.0, "retries": 0, "seconds": 216.6, "max_memory_gb": 18.27, "ce": 0.044259498827159405, "kl": 0.11596857399854343, "context": 0.10224378551356494}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 2.179068088531494, "loss_scale": 4.0, "retries": 0, "seconds": 217.2, "max_memory_gb": 18.27, "ce": 0.02797933225519955, "kl": 0.5993677678052336, "context": 0.32766321720555425}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.9605038166046143, "loss_scale": 4.0, "retries": 0, "seconds": 217.7, "max_memory_gb": 18.27, "ce": 0.13448992697522044, "kl": 0.0735767656005919, "context": 0.1712783130351454}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.15592652559280396, "loss_scale": 4.0, "retries": 0, "seconds": 218.3, "max_memory_gb": 18.27, "ce": 0.0051248540403321385, "kl": 0.009187569081404945, "context": 0.009718638670165092}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 2.552426815032959, "loss_scale": 4.0, "retries": 0, "seconds": 218.8, "max_memory_gb": 18.27, "ce": 0.22074386617168784, "kl": 0.07252069786773063, "context": 0.25700421433430165}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.7944166660308838, "loss_scale": 4.0, "retries": 0, "seconds": 219.3, "max_memory_gb": 18.27, "ce": 0.0284538131672889, "kl": 0.019170976825989783, "context": 0.038039302453398705}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 1.3168060779571533, "loss_scale": 4.0, "retries": 0, "seconds": 219.9, "max_memory_gb": 18.27, "ce": 0.14350639015901834, "kl": 0.02212698629591614, "context": 0.1545698819682002}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 3.329472303390503, "loss_scale": 4.0, "retries": 0, "seconds": 220.4, "max_memory_gb": 18.27, "ce": 0.8240674007683992, "kl": 0.45943342632381245, "context": 1.0537841483019292}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 2.7453317642211914, "loss_scale": 4.0, "retries": 0, "seconds": 220.9, "max_memory_gb": 18.27, "ce": 0.23348568961955607, "kl": 0.21898784318909748, "context": 0.3429796270793304}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.8891490697860718, "loss_scale": 4.0, "retries": 0, "seconds": 221.5, "max_memory_gb": 18.27, "ce": 0.18130165041657165, "kl": 0.07433537138513202, "context": 0.21846932679181919}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.7049877643585205, "loss_scale": 4.0, "retries": 0, "seconds": 222.0, "max_memory_gb": 18.27, "ce": 0.05225120332033839, "kl": 0.06732513230235782, "context": 0.08591377249103971}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 2.5213942527770996, "loss_scale": 4.0, "retries": 0, "seconds": 222.6, "max_memory_gb": 18.27, "ce": 0.2570409240433946, "kl": 0.21546516610396793, "context": 0.3647735173581168}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.4212394058704376, "loss_scale": 4.0, "retries": 0, "seconds": 223.1, "max_memory_gb": 18.27, "ce": 0.1429332266561687, "kl": 0.078156910429243, "context": 0.1820116783492267}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 1.4555132389068604, "loss_scale": 4.0, "retries": 0, "seconds": 223.6, "max_memory_gb": 18.27, "ce": 0.05565436207689345, "kl": 0.04724883119342849, "context": 0.07927877688780427}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 1.9970272779464722, "loss_scale": 4.0, "retries": 0, "seconds": 224.2, "max_memory_gb": 18.27, "ce": 0.19663134886650369, "kl": 0.0079805497880443, "context": 0.20062162925023586}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 2.2818801403045654, "loss_scale": 4.0, "retries": 0, "seconds": 224.7, "max_memory_gb": 18.27, "ce": 0.2934393641189672, "kl": 0.10751867410726845, "context": 0.3471987016964704}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.23352181911468506, "loss_scale": 4.0, "retries": 0, "seconds": 225.2, "max_memory_gb": 18.27, "ce": 0.014151567651424557, "kl": 0.002913121554229292, "context": 0.015608127956511453}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 2.3823273181915283, "loss_scale": 4.0, "retries": 0, "seconds": 225.8, "max_memory_gb": 18.27, "ce": 0.42897991463541985, "kl": 0.10890702367760241, "context": 0.48343343287706375}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.31868329644203186, "loss_scale": 4.0, "retries": 0, "seconds": 226.3, "max_memory_gb": 18.27, "ce": 0.02759330882690847, "kl": 0.052691590804897714, "context": 0.05393910442944616}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.3522367477416992, "loss_scale": 4.0, "retries": 0, "seconds": 226.8, "max_memory_gb": 18.27, "ce": 0.06967731937766075, "kl": 0.05884665578560089, "context": 0.09910064237192273}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 3.231227397918701, "loss_scale": 4.0, "retries": 0, "seconds": 227.4, "max_memory_gb": 18.27, "ce": 0.15861578338081017, "kl": 0.9799950258457102, "context": 0.6486132969148457}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.5896820425987244, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 18.27, "ce": 0.05033252917928621, "kl": 0.02025615647289669, "context": 0.06046060560038313}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 1.0973610877990723, "loss_scale": 4.0, "retries": 0, "seconds": 228.4, "max_memory_gb": 18.27, "ce": 0.18011719816422556, "kl": 0.014983521457907045, "context": 0.18760896133608185}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 3.9519340991973877, "loss_scale": 4.0, "retries": 0, "seconds": 229.0, "max_memory_gb": 18.27, "ce": 0.3593497443944216, "kl": 0.4157719996292144, "context": 0.5672357473522425}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.9221841096878052, "loss_scale": 4.0, "retries": 0, "seconds": 229.5, "max_memory_gb": 18.27, "ce": 0.042087744281161577, "kl": 0.0247034936764976, "context": 0.054439489962533116}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 1.530174970626831, "loss_scale": 4.0, "retries": 0, "seconds": 230.1, "max_memory_gb": 18.27, "ce": 0.05349284820840694, "kl": 0.07337021228158846, "context": 0.09017795533873141}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 2.1485726833343506, "loss_scale": 4.0, "retries": 0, "seconds": 230.6, "max_memory_gb": 18.27, "ce": 0.255918946582824, "kl": 0.0931487767957151, "context": 0.3024933347478509}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 1.405120849609375, "loss_scale": 4.0, "retries": 0, "seconds": 231.2, "max_memory_gb": 18.27, "ce": 0.08188743042410351, "kl": 0.02704378245675798, "context": 0.09540932445088401}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.594349980354309, "loss_scale": 4.0, "retries": 0, "seconds": 231.7, "max_memory_gb": 18.27, "ce": 0.005852434231201187, "kl": 0.3786265400412958, "context": 0.1951657113386318}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.9103929996490479, "loss_scale": 4.0, "retries": 0, "seconds": 232.2, "max_memory_gb": 18.27, "ce": 0.009922825003741309, "kl": 0.11668220310093602, "context": 0.06826392654329538}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 1.0834490060806274, "loss_scale": 4.0, "retries": 0, "seconds": 232.8, "max_memory_gb": 18.27, "ce": 0.10112853199825622, "kl": 0.007111308223102242, "context": 0.1046841858769767}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.444434642791748, "loss_scale": 4.0, "retries": 0, "seconds": 233.3, "max_memory_gb": 18.27, "ce": 0.21770003097481094, "kl": 0.3892503579627373, "context": 0.4123252248391509}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 1.235551357269287, "loss_scale": 4.0, "retries": 0, "seconds": 233.9, "max_memory_gb": 18.27, "ce": 0.07963447569636628, "kl": 0.07687153690494597, "context": 0.11807024478912354}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 1.4146443605422974, "loss_scale": 4.0, "retries": 0, "seconds": 234.4, "max_memory_gb": 18.27, "ce": 0.08755239105084911, "kl": 0.11837758317415137, "context": 0.14674118265975267}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.11419238895177841, "loss_scale": 4.0, "retries": 0, "seconds": 234.9, "max_memory_gb": 18.27, "ce": 0.005794662516564131, "kl": 0.0001109781296690926, "context": 0.005850151646882296}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 1.223193883895874, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 18.27, "ce": 0.11370808852370828, "kl": 0.06595237442525104, "context": 0.14668427256401628}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 1.0837273597717285, "loss_scale": 4.0, "retries": 0, "seconds": 236.0, "max_memory_gb": 18.27, "ce": 0.10764893447048962, "kl": 0.02628003020072356, "context": 0.12078894954174757}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 1.946459174156189, "loss_scale": 4.0, "retries": 0, "seconds": 236.6, "max_memory_gb": 18.27, "ce": 0.1799658517993521, "kl": 0.08568802564695943, "context": 0.2228098542545922}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.4953205585479736, "loss_scale": 4.0, "retries": 0, "seconds": 237.1, "max_memory_gb": 18.27, "ce": 0.09202708513475955, "kl": 0.039140465363743715, "context": 0.11159731296356767}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.6864622831344604, "loss_scale": 4.0, "retries": 0, "seconds": 237.6, "max_memory_gb": 18.27, "ce": 0.15662128571420908, "kl": 0.23004968838358764, "context": 0.27164612896740437}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.5302454829216003, "loss_scale": 4.0, "retries": 0, "seconds": 238.2, "max_memory_gb": 18.27, "ce": 0.037652245606295764, "kl": 0.06607015283952933, "context": 0.07068732322659343}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 3.5518369674682617, "loss_scale": 4.0, "retries": 0, "seconds": 238.7, "max_memory_gb": 18.27, "ce": 0.2540561336209066, "kl": 0.26175016415436403, "context": 0.38493122314685024}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.5570194721221924, "loss_scale": 4.0, "retries": 0, "seconds": 239.2, "max_memory_gb": 18.27, "ce": 0.07527518074493855, "kl": 0.015861586449318565, "context": 0.08320597733836621}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.8617976903915405, "loss_scale": 4.0, "retries": 0, "seconds": 239.8, "max_memory_gb": 18.27, "ce": 0.0627708351239562, "kl": 0.034237526253491524, "context": 0.0798895969055593}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 1.0187938213348389, "loss_scale": 4.0, "retries": 0, "seconds": 240.3, "max_memory_gb": 18.27, "ce": 0.10803878062870353, "kl": 0.03494075016351417, "context": 0.12550915637984872}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 1.9746018648147583, "loss_scale": 4.0, "retries": 0, "seconds": 240.9, "max_memory_gb": 18.27, "ce": 0.38421712662966456, "kl": 0.01608329784721718, "context": 0.392258766485611}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 1.978222370147705, "loss_scale": 4.0, "retries": 0, "seconds": 241.4, "max_memory_gb": 18.27, "ce": 0.1065203370526433, "kl": 0.1339748611135292, "context": 0.1735077709890902}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 1.452322006225586, "loss_scale": 4.0, "retries": 0, "seconds": 241.9, "max_memory_gb": 18.27, "ce": 0.17548845172859728, "kl": 0.027427401231761905, "context": 0.18920215975958854}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 2.9740355014801025, "loss_scale": 4.0, "retries": 0, "seconds": 242.5, "max_memory_gb": 18.27, "ce": 0.1457082089036703, "kl": 0.09688599454239011, "context": 0.19415120407938957}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.7565577030181885, "loss_scale": 4.0, "retries": 0, "seconds": 243.0, "max_memory_gb": 18.27, "ce": 0.020097907370654866, "kl": 0.011094239869635203, "context": 0.025645026849815622}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.809438169002533, "loss_scale": 4.0, "retries": 0, "seconds": 243.5, "max_memory_gb": 18.27, "ce": 0.029264238124596886, "kl": 0.032104358549077006, "context": 0.04531641828361899}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.7218322157859802, "loss_scale": 4.0, "retries": 0, "seconds": 244.1, "max_memory_gb": 18.27, "ce": 0.012115060089854524, "kl": 0.06790704608283704, "context": 0.04606858451734297}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.832519292831421, "loss_scale": 4.0, "retries": 0, "seconds": 244.6, "max_memory_gb": 18.27, "ce": 0.13635810202686116, "kl": 0.05179144035355421, "context": 0.16225382220000029}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.7276267409324646, "loss_scale": 4.0, "retries": 0, "seconds": 245.2, "max_memory_gb": 18.27, "ce": 0.044637576094828546, "kl": 0.047410215192940086, "context": 0.06834268197417259}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 1.6136538982391357, "loss_scale": 4.0, "retries": 0, "seconds": 245.7, "max_memory_gb": 18.27, "ce": 0.15316259450628422, "kl": 0.010257639572955668, "context": 0.15829141470021568}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 1.9062813520431519, "loss_scale": 4.0, "retries": 0, "seconds": 246.2, "max_memory_gb": 18.27, "ce": 0.12253723264439031, "kl": 0.194191629125271, "context": 0.2196330593433231}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.48643189668655396, "loss_scale": 4.0, "retries": 0, "seconds": 246.8, "max_memory_gb": 18.27, "ce": 0.018536318850237876, "kl": 0.013162983814254403, "context": 0.025117810524534434}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.3213292956352234, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 18.27, "ce": 0.020382124290335923, "kl": 0.04600494018814061, "context": 0.04338459437713027}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 1.3258728981018066, "loss_scale": 4.0, "retries": 0, "seconds": 247.8, "max_memory_gb": 18.27, "ce": 0.13486969214864075, "kl": 0.1257204326611827, "context": 0.19772990798810497}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 3.0623369216918945, "loss_scale": 4.0, "retries": 0, "seconds": 248.4, "max_memory_gb": 18.27, "ce": 0.11426654644310474, "kl": 0.12476756802061573, "context": 0.17665033880621195}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.41837388277053833, "loss_scale": 4.0, "retries": 0, "seconds": 248.9, "max_memory_gb": 18.27, "ce": 0.017395177972503006, "kl": 0.024046711980190594, "context": 0.029418534075375646}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 1.574770450592041, "loss_scale": 4.0, "retries": 0, "seconds": 249.5, "max_memory_gb": 18.27, "ce": 0.1769520859234035, "kl": 0.13715186146873748, "context": 0.2455280120484531}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.9375960230827332, "loss_scale": 4.0, "retries": 0, "seconds": 250.0, "max_memory_gb": 18.27, "ce": 0.04025031998753548, "kl": 0.007733426358754514, "context": 0.04411703381629195}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.6155059933662415, "loss_scale": 4.0, "retries": 0, "seconds": 250.5, "max_memory_gb": 18.27, "ce": 0.025537703113514, "kl": 0.0035372741767787375, "context": 0.027306339878123254}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.3231358230113983, "loss_scale": 4.0, "retries": 0, "seconds": 251.1, "max_memory_gb": 18.27, "ce": 0.004404757404699922, "kl": 0.0019609729934018105, "context": 0.005385243974160403}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 2.8295605182647705, "loss_scale": 4.0, "retries": 0, "seconds": 251.6, "max_memory_gb": 18.27, "ce": 0.014998095226474106, "kl": 0.7461026157779997, "context": 0.3880494182521943}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 1.764207124710083, "loss_scale": 4.0, "retries": 0, "seconds": 252.2, "max_memory_gb": 18.27, "ce": 0.009552390227327123, "kl": 0.30676184605954404, "context": 0.16293331427732483}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.8460760116577148, "loss_scale": 4.0, "retries": 0, "seconds": 252.7, "max_memory_gb": 18.27, "ce": 0.037420177133753896, "kl": 0.01539891892389278, "context": 0.04511963576078415}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.38658657670021057, "loss_scale": 4.0, "retries": 0, "seconds": 253.2, "max_memory_gb": 18.27, "ce": 0.01610936052748002, "kl": 0.005753248427936342, "context": 0.018985984497703612}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 3.4823415279388428, "loss_scale": 4.0, "retries": 0, "seconds": 253.8, "max_memory_gb": 18.27, "ce": 0.1469537508673966, "kl": 0.032787829244625755, "context": 0.16334766917861998}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 2.4041097164154053, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 18.27, "ce": 0.002808033983455971, "kl": 0.42763502969137335, "context": 0.21662554182694294}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 3.120218515396118, "loss_scale": 4.0, "retries": 0, "seconds": 254.8, "max_memory_gb": 18.27, "ce": 0.15739721927093342, "kl": 0.20346942424657755, "context": 0.25913193624000996}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.42152220010757446, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 18.27, "ce": 0.009804455563426018, "kl": 0.009900350065436214, "context": 0.014754630508832633}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.8823944330215454, "loss_scale": 4.0, "retries": 0, "seconds": 255.9, "max_memory_gb": 18.27, "ce": 0.027504941826919094, "kl": 0.01275398138386663, "context": 0.03388193290447816}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 1.2359659671783447, "loss_scale": 4.0, "retries": 0, "seconds": 256.4, "max_memory_gb": 18.27, "ce": 0.08072229230310768, "kl": 0.011350815575497109, "context": 0.0863977026892826}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 2.2182652950286865, "loss_scale": 4.0, "retries": 0, "seconds": 257.0, "max_memory_gb": 18.27, "ce": 0.20322210974700283, "kl": 0.006309062220680062, "context": 0.2063766432693228}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.6215516924858093, "loss_scale": 4.0, "retries": 0, "seconds": 257.5, "max_memory_gb": 18.27, "ce": 0.026266542423400097, "kl": 0.009564426320139319, "context": 0.031048755103256553}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 2.92107892036438, "loss_scale": 4.0, "retries": 0, "seconds": 258.1, "max_memory_gb": 18.27, "ce": 0.0065914083970710635, "kl": 0.6413221476323088, "context": 0.3272524855856318}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.6943455338478088, "loss_scale": 4.0, "retries": 0, "seconds": 258.6, "max_memory_gb": 18.27, "ce": 0.02166442564339377, "kl": 0.03772823965579164, "context": 0.04052854680776363}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 2.989164113998413, "loss_scale": 4.0, "retries": 0, "seconds": 259.1, "max_memory_gb": 18.27, "ce": 0.3977512759156525, "kl": 0.3016342272312613, "context": 0.5485683881561272}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 1.063096523284912, "loss_scale": 4.0, "retries": 0, "seconds": 259.7, "max_memory_gb": 18.27, "ce": 0.18575108633376658, "kl": 0.036963262245990336, "context": 0.20423272461630404}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.700890064239502, "loss_scale": 4.0, "retries": 0, "seconds": 260.2, "max_memory_gb": 18.27, "ce": 0.018904175667557865, "kl": 0.02535451673611533, "context": 0.0315814349160064}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.7737852334976196, "loss_scale": 4.0, "retries": 0, "seconds": 260.8, "max_memory_gb": 18.27, "ce": 0.08321654313476756, "kl": 0.017248484771698713, "context": 0.09184078499674797}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 2.28182315826416, "loss_scale": 4.0, "retries": 0, "seconds": 261.3, "max_memory_gb": 18.27, "ce": 0.1382798245176673, "kl": 0.5868465802632272, "context": 0.4317031092941761}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 3.7464444637298584, "loss_scale": 4.0, "retries": 0, "seconds": 261.8, "max_memory_gb": 18.27, "ce": 0.5001480169594288, "kl": 0.14135947357863188, "context": 0.5708277644589543}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 1.9569768905639648, "loss_scale": 4.0, "retries": 0, "seconds": 262.4, "max_memory_gb": 18.27, "ce": 0.08980335295200348, "kl": 0.05569909361656755, "context": 0.11765289772301912}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 1.577449917793274, "loss_scale": 4.0, "retries": 0, "seconds": 262.9, "max_memory_gb": 18.27, "ce": 0.138900826103054, "kl": 0.030420037684962153, "context": 0.15411084587685764}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 0.9619074463844299, "loss_scale": 4.0, "retries": 0, "seconds": 263.4, "max_memory_gb": 18.27, "ce": 0.04330778814619407, "kl": 0.058748637815369875, "context": 0.07268210692564026}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.7506069540977478, "loss_scale": 4.0, "retries": 0, "seconds": 264.0, "max_memory_gb": 18.27, "ce": 0.08726502972422168, "kl": 0.018674255697987974, "context": 0.09660215571057051}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 3.143993377685547, "loss_scale": 4.0, "retries": 0, "seconds": 264.5, "max_memory_gb": 18.27, "ce": 0.3243514540954493, "kl": 0.09358976173098199, "context": 0.37114633608143777}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.5539674758911133, "loss_scale": 4.0, "retries": 0, "seconds": 265.0, "max_memory_gb": 18.27, "ce": 0.02459730206464883, "kl": 0.010955630532407667, "context": 0.030075118207605556}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 1.142110824584961, "loss_scale": 4.0, "retries": 0, "seconds": 265.6, "max_memory_gb": 18.27, "ce": 0.09508921210363042, "kl": 0.15884848361019976, "context": 0.1745134539087303}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.619614839553833, "loss_scale": 4.0, "retries": 0, "seconds": 266.1, "max_memory_gb": 18.27, "ce": 0.01909634517505765, "kl": 0.005909678991883993, "context": 0.022051184787414968}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 1.4535185098648071, "loss_scale": 4.0, "retries": 0, "seconds": 266.7, "max_memory_gb": 18.27, "ce": 0.07676580571569502, "kl": 0.014007116435095668, "context": 0.08376936591230333}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 2.7032759189605713, "loss_scale": 4.0, "retries": 0, "seconds": 267.2, "max_memory_gb": 18.27, "ce": 0.007463175366865471, "kl": 0.5008803068412817, "context": 0.2579033203655854}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.8126949071884155, "loss_scale": 4.0, "retries": 0, "seconds": 267.7, "max_memory_gb": 18.27, "ce": 0.03549641743302345, "kl": 0.009650374588090926, "context": 0.040321605280041695}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.480910062789917, "loss_scale": 4.0, "retries": 0, "seconds": 268.3, "max_memory_gb": 18.27, "ce": 0.02494908805238083, "kl": 0.02399338036775589, "context": 0.036945778294466436}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 1.5904123783111572, "loss_scale": 4.0, "retries": 0, "seconds": 268.8, "max_memory_gb": 18.27, "ce": 0.14812951898784377, "kl": 0.026786812610225752, "context": 0.1615229233284481}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.6211350560188293, "loss_scale": 4.0, "retries": 0, "seconds": 269.3, "max_memory_gb": 18.27, "ce": 0.12013190593279433, "kl": 0.04141098202671856, "context": 0.14083739439956844}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 2.154561996459961, "loss_scale": 4.0, "retries": 0, "seconds": 269.9, "max_memory_gb": 18.27, "ce": 0.16285137044178555, "kl": 0.03715742873464478, "context": 0.1814300898113288}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 3.009082317352295, "loss_scale": 4.0, "retries": 0, "seconds": 270.4, "max_memory_gb": 18.27, "ce": 0.304044113116106, "kl": 0.009471086668781936, "context": 0.3087796498439275}
diff --git a/runs/single_llava/context/L02_s13/adapter/README.md b/runs/single_llava/context/L02_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L02_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L02_s13/adapter/adapter_config.json b/runs/single_llava/context/L02_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..6cc72169a80597067cc519358141924cd8b84d00
--- /dev/null
+++ b/runs/single_llava/context/L02_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.2.self_attn.o_proj",
+    "model.language_model.layers.2.self_attn.v_proj",
+    "model.language_model.layers.2.mlp.gate_proj",
+    "model.language_model.layers.2.mlp.up_proj",
+    "model.language_model.layers.2.mlp.down_proj",
+    "model.language_model.layers.2.self_attn.q_proj",
+    "model.language_model.layers.2.self_attn.k_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L02_s13/steps.jsonl b/runs/single_llava/context/L02_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..9517cb5e7abfd996d839b26b0b866ee92466b945
--- /dev/null
+++ b/runs/single_llava/context/L02_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.17777881026268005, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 18.06, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.23760749399662018, "loss_scale": 4.0, "retries": 0, "seconds": 2.2, "max_memory_gb": 18.07, "ce": 2.6818245879840106, "kl": 0.010183475897065364, "context": 2.6869163650553674}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.5394285321235657, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 18.07, "ce": 5.391443431377411, "kl": 0.17863947732257657, "context": 5.480763256549835}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.3190280795097351, "loss_scale": 4.0, "retries": 0, "seconds": 3.3, "max_memory_gb": 18.07, "ce": 3.825445403344929, "kl": 0.051676184542884585, "context": 3.8512835251167417}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.04965711757540703, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 18.08, "ce": 0.08617512322962284, "kl": 0.036465648328885436, "context": 0.10440794751048088}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.13985419273376465, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 18.08, "ce": 1.4656987451016903, "kl": 0.18327271682210267, "context": 1.5573351383209229}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.40952274203300476, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 18.08, "ce": 3.438128577079624, "kl": 0.10538836949854158, "context": 3.49082278739661}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.40514230728149414, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 18.08, "ce": 3.6930223554372787, "kl": 0.16066167876124382, "context": 3.773353189229965}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.32083606719970703, "loss_scale": 4.0, "retries": 0, "seconds": 6.2, "max_memory_gb": 18.08, "ce": 2.8385237865149975, "kl": 0.16233081044629216, "context": 2.9196892231702805}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.2011518031358719, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 18.08, "ce": 1.6323690488934517, "kl": 0.0599564568256028, "context": 1.662347249686718}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.40300023555755615, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 18.08, "ce": 2.623912073671818, "kl": 0.009577320954122115, "context": 2.6287007397040725}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.22581468522548676, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 18.08, "ce": 1.534041598206386, "kl": 0.01742915390059352, "context": 1.542756175622344}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.3560962677001953, "loss_scale": 4.0, "retries": 0, "seconds": 8.3, "max_memory_gb": 18.08, "ce": 2.667892591562122, "kl": 0.007436472384142689, "context": 2.6716108857654035}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.5464772582054138, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 18.09, "ce": 3.888931304216385, "kl": 0.11891112814191729, "context": 3.9483868330717087}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.3955923616886139, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 18.09, "ce": 2.5819904924719594, "kl": 0.2688325824274216, "context": 2.7164067236008123}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.2622440457344055, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 18.09, "ce": 1.370463193859905, "kl": 0.009162417023617309, "context": 1.3750443968456239}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.24788956344127655, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 18.09, "ce": 1.5879605473019183, "kl": 0.07877922442276031, "context": 1.6273501827381551}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.49363622069358826, "loss_scale": 4.0, "retries": 0, "seconds": 11.1, "max_memory_gb": 18.1, "ce": 3.5690247528254986, "kl": 0.09410755638964474, "context": 3.6160786002874374}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 0.9099277257919312, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 18.1, "ce": 5.323945760726929, "kl": 0.33149446587776765, "context": 5.48969292640686}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.01404449250549078, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 18.1, "ce": 0.01301277638413012, "kl": 0.0021321246749721467, "context": 0.014078838634304702}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.49544599652290344, "loss_scale": 4.0, "retries": 0, "seconds": 12.7, "max_memory_gb": 18.1, "ce": 2.76856482680887, "kl": 0.029737802455201745, "context": 2.783433787059039}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.514959990978241, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 18.1, "ce": 2.93149352632463, "kl": 0.05363805138040334, "context": 2.9583126604557037}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.5467601418495178, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 18.1, "ce": 2.6456924779340625, "kl": 0.51143564216909, "context": 2.901410270947963}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 0.9893494248390198, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 18.1, "ce": 4.991535723209381, "kl": 0.013971218140795827, "context": 4.998521327972412}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 0.9616848826408386, "loss_scale": 4.0, "retries": 0, "seconds": 14.9, "max_memory_gb": 18.1, "ce": 4.215441912412643, "kl": 0.03608486521989107, "context": 4.233484320342541}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.282279133796692, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 18.1, "ce": 5.189104855060577, "kl": 0.06687780050560832, "context": 5.222543835639954}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 0.9848466515541077, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 18.1, "ce": 4.160544469952583, "kl": 0.08612522669136524, "context": 4.203607130795717}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.6568374633789062, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 18.1, "ce": 2.7885872423648834, "kl": 0.23283198103308678, "context": 2.9050032794475555}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 0.5995947122573853, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 18.1, "ce": 2.5983147653751075, "kl": 0.5197234733495861, "context": 2.858176516601816}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.016492247581482, "loss_scale": 4.0, "retries": 0, "seconds": 17.6, "max_memory_gb": 18.1, "ce": 2.70434127189219, "kl": 0.037390715326182544, "context": 2.723036670126021}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 0.9388712644577026, "loss_scale": 4.0, "retries": 0, "seconds": 18.1, "max_memory_gb": 18.1, "ce": 2.808800958096981, "kl": 0.02842404553666711, "context": 2.8230129331350327}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.7123547792434692, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 18.1, "ce": 5.258619070053101, "kl": 0.0158841616648715, "context": 5.266561150550842}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 0.8633567690849304, "loss_scale": 4.0, "retries": 0, "seconds": 19.2, "max_memory_gb": 18.1, "ce": 2.434476579539478, "kl": 0.18726155417971313, "context": 2.5281072948127985}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 0.961719274520874, "loss_scale": 4.0, "retries": 0, "seconds": 19.7, "max_memory_gb": 18.1, "ce": 2.596939889714122, "kl": 0.6173741812817752, "context": 2.905626946128905}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.4132541418075562, "loss_scale": 4.0, "retries": 0, "seconds": 20.3, "max_memory_gb": 18.1, "ce": 2.644627600442618, "kl": 0.029611849604407325, "context": 2.659433432854712}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 0.8759727478027344, "loss_scale": 4.0, "retries": 0, "seconds": 20.8, "max_memory_gb": 18.1, "ce": 2.523729767650366, "kl": 0.00686372013296932, "context": 2.5271616876125336}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.0758272409439087, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 18.1, "ce": 2.4349697679281235, "kl": 0.3390898110810667, "context": 2.604514751583338}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 2.0345749855041504, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 18.1, "ce": 4.589843451976776, "kl": 0.4848613328067586, "context": 4.83227413892746}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.9929487109184265, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 18.1, "ce": 1.371702435426414, "kl": 0.03283894620835781, "context": 1.3881219243630767}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.5636160373687744, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 18.1, "ce": 2.6301096826791763, "kl": 0.11652499571209773, "context": 2.6883721500635147}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.6105539202690125, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 18.1, "ce": 1.437314085662365, "kl": 0.17276785732246935, "context": 1.5236979871988297}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.444681167602539, "loss_scale": 4.0, "retries": 0, "seconds": 24.0, "max_memory_gb": 18.1, "ce": 2.2771832728758454, "kl": 0.007155418657930568, "context": 2.2807609708979726}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.1888906955718994, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 18.1, "ce": 2.2996669290587306, "kl": 0.010296883032424375, "context": 2.3048154176212847}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.1680432558059692, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 18.1, "ce": 1.323128504678607, "kl": 0.2636870302958414, "context": 1.4549719747155905}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.3273156881332397, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 18.1, "ce": 1.3242789916694164, "kl": 0.056350849103182554, "context": 1.3524543903768063}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.9884892702102661, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 18.1, "ce": 2.1765983775258064, "kl": 0.07883745321305469, "context": 2.2160170450806618}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.638451337814331, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 18.1, "ce": 1.8591674179770052, "kl": 0.03186223113152664, "context": 1.8750985588412732}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 4.4168853759765625, "loss_scale": 4.0, "retries": 0, "seconds": 27.3, "max_memory_gb": 18.1, "ce": 3.4643557481467724, "kl": 0.015445112796442118, "context": 3.4720782190561295}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 6.091448783874512, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 18.1, "ce": 3.1578616704791784, "kl": 0.07606131193460897, "context": 3.195892257615924}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 4.237273693084717, "loss_scale": 4.0, "retries": 0, "seconds": 28.4, "max_memory_gb": 18.1, "ce": 3.3633419470861554, "kl": 0.0514054637751542, "context": 3.389044662937522}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.9098769426345825, "loss_scale": 4.0, "retries": 0, "seconds": 28.9, "max_memory_gb": 18.1, "ce": 1.2440438168123364, "kl": 0.12102978397160769, "context": 1.304558763280511}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 4.3976640701293945, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 18.1, "ce": 1.9775629676878452, "kl": 0.11689825158100575, "context": 2.0360120311379433}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 6.685367584228516, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 18.1, "ce": 3.282810002565384, "kl": 0.21744185429997742, "context": 3.391530990600586}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 8.990691184997559, "loss_scale": 4.0, "retries": 0, "seconds": 30.5, "max_memory_gb": 18.1, "ce": 3.1508410200476646, "kl": 0.040056858269963413, "context": 3.1708693355321884}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 6.372836589813232, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 18.1, "ce": 1.8363250754773617, "kl": 0.026011927257059142, "context": 1.849331026431173}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 12.7736177444458, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 18.1, "ce": 3.642251431941986, "kl": 0.3472290588542819, "context": 3.8158658742904663}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 3.108693838119507, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 18.1, "ce": 0.9485235805623233, "kl": 0.11027798953000456, "context": 1.003662589006126}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 6.140505313873291, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 18.1, "ce": 1.6423096002545208, "kl": 0.004683559844124829, "context": 1.6446513701230288}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.8465993404388428, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 18.1, "ce": 1.2067248262465, "kl": 0.08228124747984111, "context": 1.247865503653884}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 5.529757022857666, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 18.1, "ce": 1.8147636009380221, "kl": 0.015264973801095039, "context": 1.8223961107432842}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 9.862388610839844, "loss_scale": 4.0, "retries": 0, "seconds": 34.3, "max_memory_gb": 18.1, "ce": 1.8687161058187485, "kl": 0.090145482448861, "context": 1.913788877427578}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 4.108964443206787, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 18.1, "ce": 0.9163682572543621, "kl": 0.03941699909046292, "context": 0.9360767342150211}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 7.671113967895508, "loss_scale": 4.0, "retries": 0, "seconds": 35.3, "max_memory_gb": 18.1, "ce": 2.1358867986127734, "kl": 0.08073613257147372, "context": 2.1762548238039017}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 10.21703815460205, "loss_scale": 4.0, "retries": 0, "seconds": 35.9, "max_memory_gb": 18.13, "ce": 2.063472431153059, "kl": 0.042602536268532276, "context": 2.084773726761341}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 10.00862979888916, "loss_scale": 4.0, "retries": 0, "seconds": 36.4, "max_memory_gb": 18.13, "ce": 2.036473512649536, "kl": 0.1134147453121841, "context": 2.093180865049362}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 2.28200101852417, "loss_scale": 4.0, "retries": 0, "seconds": 37.0, "max_memory_gb": 18.13, "ce": 0.625395396258682, "kl": 0.05851127105415799, "context": 0.6546510527841747}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 6.518097400665283, "loss_scale": 4.0, "retries": 0, "seconds": 37.5, "max_memory_gb": 18.13, "ce": 1.204924337565899, "kl": 0.0372288660146296, "context": 1.2235387861728668}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 3.1185951232910156, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 18.13, "ce": 0.675814451649785, "kl": 0.058079122798517346, "context": 0.7048540231771767}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 5.316038608551025, "loss_scale": 4.0, "retries": 0, "seconds": 38.6, "max_memory_gb": 18.13, "ce": 0.9421472400426865, "kl": 0.025670613395050168, "context": 0.9549825433641672}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 4.315011501312256, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 18.13, "ce": 0.7571632526814938, "kl": 0.005977823166176677, "context": 0.7601521704345942}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 7.84963321685791, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 18.13, "ce": 1.1850427370518446, "kl": 0.02859638538211584, "context": 1.1993409171700478}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 9.451403617858887, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 18.13, "ce": 1.3500390499830246, "kl": 0.12400217703543603, "context": 1.4120401442050934}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.7447745203971863, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 18.13, "ce": 0.2776174321770668, "kl": 0.04010792542248964, "context": 0.29767139721661806}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.9018442630767822, "loss_scale": 4.0, "retries": 0, "seconds": 41.2, "max_memory_gb": 18.13, "ce": 0.5877009332180023, "kl": 0.02373705548234284, "context": 0.5995694622397423}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.1374289989471436, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 18.13, "ce": 0.8981253318488598, "kl": 0.1931579541414976, "context": 0.9947043135762215}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 3.1710171699523926, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 18.13, "ce": 0.5059911515563726, "kl": 0.2578839361667633, "context": 0.6349331215023994}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 5.316859722137451, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 18.13, "ce": 1.0252347439527512, "kl": 0.04562052618712187, "context": 1.0480450093746185}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 2.8594958782196045, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 18.13, "ce": 0.4542821981012821, "kl": 0.05288342107087374, "context": 0.48072390630841255}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.430938720703125, "loss_scale": 4.0, "retries": 0, "seconds": 43.9, "max_memory_gb": 18.13, "ce": 0.3508194647729397, "kl": 0.08783850132022053, "context": 0.3947387244552374}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 1.8981677293777466, "loss_scale": 4.0, "retries": 0, "seconds": 44.4, "max_memory_gb": 18.13, "ce": 0.3723255842924118, "kl": 0.021080620819702744, "context": 0.38286589831113815}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 3.150226593017578, "loss_scale": 4.0, "retries": 0, "seconds": 45.0, "max_memory_gb": 18.13, "ce": 0.7049063146114349, "kl": 0.0865417355671525, "context": 0.7481771856546402}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 2.3292102813720703, "loss_scale": 4.0, "retries": 0, "seconds": 45.5, "max_memory_gb": 18.13, "ce": 0.4605388194322586, "kl": 0.07075259368866682, "context": 0.49591512233018875}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 2.052311420440674, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 18.13, "ce": 0.5242783613502979, "kl": 0.1303627621382475, "context": 0.5894597470760345}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.6706210374832153, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 18.13, "ce": 0.34270327538251877, "kl": 0.053895653458312154, "context": 0.36965109780430794}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 2.01822829246521, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 18.13, "ce": 0.1793598160147667, "kl": 0.040348822047235444, "context": 0.19953422993421555}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 3.5415432453155518, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 18.13, "ce": 0.58699019998312, "kl": 0.13143398147076368, "context": 0.6527071744203568}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 1.6167428493499756, "loss_scale": 4.0, "retries": 0, "seconds": 48.2, "max_memory_gb": 18.13, "ce": 0.4983898922801018, "kl": 0.06645204499363899, "context": 0.5316159091889858}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.1836748123168945, "loss_scale": 4.0, "retries": 0, "seconds": 48.7, "max_memory_gb": 18.13, "ce": 0.6010293252766132, "kl": 0.031946597329806536, "context": 0.6170026250183582}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 1.7223762273788452, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 18.13, "ce": 0.1809986848384142, "kl": 0.011389180785045028, "context": 0.18669327720999718}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.018613815307617, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 18.13, "ce": 0.5185696091502905, "kl": 0.039781870087608695, "context": 0.5384605377912521}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.301755428314209, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 18.13, "ce": 0.44071876257658005, "kl": 0.08712516934610903, "context": 0.48428134620189667}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.2144951820373535, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 18.13, "ce": 0.2541889613494277, "kl": 0.07825518120080233, "context": 0.2933165468275547}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 1.4933671951293945, "loss_scale": 4.0, "retries": 0, "seconds": 51.4, "max_memory_gb": 18.13, "ce": 0.38499658834189177, "kl": 0.05921716208104044, "context": 0.41460515558719635}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 2.253922700881958, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 18.13, "ce": 0.4485592842102051, "kl": 0.14968649670481682, "context": 0.5234025195240974}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 1.5670697689056396, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 18.13, "ce": 0.5410046335309744, "kl": 0.06343141477555037, "context": 0.5727203451097012}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 1.7578293085098267, "loss_scale": 4.0, "retries": 0, "seconds": 53.0, "max_memory_gb": 18.13, "ce": 0.5331621170043945, "kl": 0.025150538538582623, "context": 0.5457373932003975}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 1.0213171243667603, "loss_scale": 4.0, "retries": 0, "seconds": 53.5, "max_memory_gb": 18.13, "ce": 0.3493302911520004, "kl": 0.034281173488125205, "context": 0.3664708770811558}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 1.0974757671356201, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 18.13, "ce": 0.2505703894421458, "kl": 0.05222148432221729, "context": 0.2766811326146126}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 1.113996148109436, "loss_scale": 4.0, "retries": 0, "seconds": 54.6, "max_memory_gb": 18.13, "ce": 0.2814726117067039, "kl": 0.06611307081766427, "context": 0.3145291437394917}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 1.9044368267059326, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 18.13, "ce": 0.45762380957603455, "kl": 0.06861835066229105, "context": 0.4919329956173897}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 1.7807459831237793, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 18.13, "ce": 0.43098197132349014, "kl": 0.35091177513822913, "context": 0.6064378619194031}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 2.6448404788970947, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 18.13, "ce": 0.6240658946335316, "kl": 0.13949250616133213, "context": 0.6938121467828751}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 1.067785382270813, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 18.13, "ce": 0.18807460879907012, "kl": 0.04529790487140417, "context": 0.21072356030344963}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 0.9120210409164429, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 18.13, "ce": 0.19975011772476137, "kl": 0.05692178977187723, "context": 0.22821100847795606}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 1.2287648916244507, "loss_scale": 4.0, "retries": 0, "seconds": 57.9, "max_memory_gb": 18.13, "ce": 0.3372184671461582, "kl": 0.15865236800163984, "context": 0.4165446609258652}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 3.1443638801574707, "loss_scale": 4.0, "retries": 0, "seconds": 58.4, "max_memory_gb": 18.13, "ce": 0.34111201390624046, "kl": 0.07996222376823425, "context": 0.3810931257903576}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.5441587567329407, "loss_scale": 4.0, "retries": 0, "seconds": 58.9, "max_memory_gb": 18.13, "ce": 0.20619555190205574, "kl": 0.026667530793929473, "context": 0.21952931862324476}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.7192615866661072, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 18.13, "ce": 0.2364714378491044, "kl": 0.02355715390876867, "context": 0.248250016476959}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 2.6419363021850586, "loss_scale": 4.0, "retries": 0, "seconds": 60.0, "max_memory_gb": 18.13, "ce": 0.4018334895372391, "kl": 0.20313198724761605, "context": 0.5033994764089584}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 1.213173270225525, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 18.13, "ce": 0.36093200743198395, "kl": 0.053443113109096885, "context": 0.38765357434749603}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 1.7750784158706665, "loss_scale": 4.0, "retries": 0, "seconds": 61.1, "max_memory_gb": 18.13, "ce": 0.26749164424836636, "kl": 0.08331354655092582, "context": 0.3091484154574573}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 0.7060455679893494, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 18.13, "ce": 0.19245238602161407, "kl": 0.015538177802227437, "context": 0.2002214714884758}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 1.4959657192230225, "loss_scale": 4.0, "retries": 0, "seconds": 62.2, "max_memory_gb": 18.13, "ce": 0.23556588776409626, "kl": 0.06504106614738703, "context": 0.26808642596006393}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 2.1323893070220947, "loss_scale": 4.0, "retries": 0, "seconds": 62.7, "max_memory_gb": 18.13, "ce": 0.5058195665478706, "kl": 0.6514444705098867, "context": 0.8315418139100075}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.7215030193328857, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 18.13, "ce": 0.29484938830137253, "kl": 0.17702723667025566, "context": 0.3833630010485649}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.4000106751918793, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 18.13, "ce": 0.1286053559742868, "kl": 0.009738166409078985, "context": 0.1334744393825531}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 2.4244964122772217, "loss_scale": 4.0, "retries": 0, "seconds": 64.3, "max_memory_gb": 18.13, "ce": 0.47528366674669087, "kl": 0.035188759880838916, "context": 0.49287804786581546}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.44871482253074646, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 18.13, "ce": 0.18721524067223072, "kl": 0.04490893188631162, "context": 0.20966970548033714}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.6715685725212097, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 18.13, "ce": 0.17910747788846493, "kl": 0.08248008077498525, "context": 0.22034750785678625}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 1.1841140985488892, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 18.13, "ce": 0.38345268927514553, "kl": 0.07337019429542124, "context": 0.4201377918943763}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 0.9809251427650452, "loss_scale": 4.0, "retries": 0, "seconds": 66.5, "max_memory_gb": 18.13, "ce": 0.3907442046329379, "kl": 0.10241441562538967, "context": 0.44195142621174455}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 0.9629122614860535, "loss_scale": 4.0, "retries": 0, "seconds": 67.0, "max_memory_gb": 18.13, "ce": 0.21321470639668405, "kl": 0.021244824747554958, "context": 0.22383711207658052}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.3771684169769287, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 18.13, "ce": 0.4246148709207773, "kl": 0.04366077156737447, "context": 0.44644525181502104}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 1.3949178457260132, "loss_scale": 4.0, "retries": 0, "seconds": 68.1, "max_memory_gb": 18.13, "ce": 0.3224717089906335, "kl": 0.033918434055522084, "context": 0.33943092450499535}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.36179018020629883, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 18.13, "ce": 0.1878470228984952, "kl": 0.020016545138787478, "context": 0.19785529468208551}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 1.2708423137664795, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 18.13, "ce": 0.4208563566207886, "kl": 0.0941838970175013, "context": 0.4679483100771904}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.5082425475120544, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 18.13, "ce": 0.09743202989920974, "kl": 0.004482680902583525, "context": 0.09967337129637599}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 0.7035898566246033, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 18.13, "ce": 0.22067668056115508, "kl": 0.2936784217308741, "context": 0.3675158894620836}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 2.008073568344116, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 18.13, "ce": 0.19360584765672684, "kl": 0.06278695710352622, "context": 0.2249993234872818}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 1.3426448106765747, "loss_scale": 4.0, "retries": 0, "seconds": 71.3, "max_memory_gb": 18.13, "ce": 0.39245809614658356, "kl": 0.31012045446550474, "context": 0.5475183222442865}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 0.9286509156227112, "loss_scale": 4.0, "retries": 0, "seconds": 71.9, "max_memory_gb": 18.13, "ce": 0.3232850395143032, "kl": 0.05618588859215379, "context": 0.3513779826462269}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.3214458227157593, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 18.13, "ce": 0.36271841055713594, "kl": 0.16857802183949389, "context": 0.4470074283890426}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 1.6757547855377197, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 18.13, "ce": 0.28794805333018303, "kl": 0.1420137584209442, "context": 0.35895492881536484}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.608802318572998, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 18.13, "ce": 0.21199921052902937, "kl": 0.01741860060428735, "context": 0.22070850990712643}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.647294282913208, "loss_scale": 4.0, "retries": 0, "seconds": 74.0, "max_memory_gb": 18.13, "ce": 0.25074583664536476, "kl": 0.018425175279844552, "context": 0.2599584273993969}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 1.2029346227645874, "loss_scale": 4.0, "retries": 0, "seconds": 74.5, "max_memory_gb": 18.13, "ce": 0.20294956397265196, "kl": 0.0545431156642735, "context": 0.23022112250328064}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.6950628161430359, "loss_scale": 4.0, "retries": 0, "seconds": 75.1, "max_memory_gb": 18.13, "ce": 0.16732441540807486, "kl": 0.0240878164768219, "context": 0.17936832457780838}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.595964252948761, "loss_scale": 4.0, "retries": 0, "seconds": 75.6, "max_memory_gb": 18.13, "ce": 0.12830720841884613, "kl": 0.07418982870876789, "context": 0.1654021218419075}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 5.699001312255859, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 18.13, "ce": 0.31426560692489147, "kl": 0.11790363979525864, "context": 0.37321742810308933}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.6328405141830444, "loss_scale": 4.0, "retries": 0, "seconds": 76.7, "max_memory_gb": 18.13, "ce": 0.13814541138708591, "kl": 0.023268692893907428, "context": 0.14977975375950336}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.5542738437652588, "loss_scale": 4.0, "retries": 0, "seconds": 77.2, "max_memory_gb": 18.13, "ce": 0.17815153487026691, "kl": 0.019654227420687675, "context": 0.18797865323722363}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.570348858833313, "loss_scale": 4.0, "retries": 0, "seconds": 77.7, "max_memory_gb": 18.13, "ce": 0.17359716910868883, "kl": 0.0136770160170272, "context": 0.18043567379936576}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.2984142303466797, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 18.13, "ce": 0.1361352577805519, "kl": 0.056329823593841866, "context": 0.16430016583763063}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 1.5844228267669678, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 18.13, "ce": 0.2751513347029686, "kl": 0.14830909250304103, "context": 0.34930588491261005}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.6773377656936646, "loss_scale": 4.0, "retries": 0, "seconds": 79.4, "max_memory_gb": 18.13, "ce": 0.0499215885065496, "kl": 0.005836294069013093, "context": 0.05283973668701947}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.6898636221885681, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 18.13, "ce": 0.15250172186642885, "kl": 0.013221093802712858, "context": 0.15911227092146873}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.5985109210014343, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 18.13, "ce": 0.19830421451479197, "kl": 0.40603998815640807, "context": 0.4013242176733911}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.5135658979415894, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 18.13, "ce": 0.04017169470898807, "kl": 0.08533379842992872, "context": 0.0828385935164988}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 2.329118013381958, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 18.13, "ce": 0.14702991815283895, "kl": 0.0256959513790207, "context": 0.15987789304926991}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.941100001335144, "loss_scale": 4.0, "retries": 0, "seconds": 82.0, "max_memory_gb": 18.13, "ce": 0.24134360440075397, "kl": 0.04401479219086468, "context": 0.26335099060088396}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 7.394458293914795, "loss_scale": 4.0, "retries": 0, "seconds": 82.6, "max_memory_gb": 18.13, "ce": 0.26245351042598486, "kl": 0.023263494920684025, "context": 0.274085262324661}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 1.692649245262146, "loss_scale": 4.0, "retries": 0, "seconds": 83.1, "max_memory_gb": 18.13, "ce": 0.12093376740813255, "kl": 0.3798830399173312, "context": 0.31087528727948666}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 3.8608546257019043, "loss_scale": 4.0, "retries": 0, "seconds": 83.6, "max_memory_gb": 18.13, "ce": 0.1422819048166275, "kl": 0.2639851216226816, "context": 0.27427446376532316}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.3722169399261475, "loss_scale": 4.0, "retries": 0, "seconds": 84.2, "max_memory_gb": 18.13, "ce": 0.5234378650784492, "kl": 0.3330222205258906, "context": 0.6899489834904671}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.061215400695801, "loss_scale": 4.0, "retries": 0, "seconds": 84.7, "max_memory_gb": 18.13, "ce": 0.5993352960795164, "kl": 0.14733342442195863, "context": 0.6730020213872194}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.0761178731918335, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 18.13, "ce": 0.19295549299567938, "kl": 0.1154255003475555, "context": 0.25066823174711317}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.8247249722480774, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 18.13, "ce": 0.17750202119350433, "kl": 0.046280269423732534, "context": 0.2006421610713005}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.6607673764228821, "loss_scale": 4.0, "retries": 0, "seconds": 86.3, "max_memory_gb": 18.13, "ce": 0.3105472084134817, "kl": 0.23827250069007277, "context": 0.42968345433473587}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.93448805809021, "loss_scale": 4.0, "retries": 0, "seconds": 86.8, "max_memory_gb": 18.13, "ce": 0.0828223405405879, "kl": 0.08571341098286211, "context": 0.12567904591560364}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.9483373165130615, "loss_scale": 4.0, "retries": 0, "seconds": 87.4, "max_memory_gb": 18.13, "ce": 0.13367719948291779, "kl": 0.09380559273995459, "context": 0.18057999201118946}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.46249744296073914, "loss_scale": 4.0, "retries": 0, "seconds": 87.9, "max_memory_gb": 18.13, "ce": 0.22215939313173294, "kl": 0.053901197359664366, "context": 0.24910998204723}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.45203113555908203, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 18.13, "ce": 0.10950454510748386, "kl": 0.029489096661563963, "context": 0.12424909137189388}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.7582813501358032, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 18.13, "ce": 0.10144473519176245, "kl": 0.09237908986688126, "context": 0.14763427712023258}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 1.4162201881408691, "loss_scale": 4.0, "retries": 0, "seconds": 89.5, "max_memory_gb": 18.13, "ce": 0.2960374616086483, "kl": 0.0264686681330204, "context": 0.30927179008722305}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.8242514133453369, "loss_scale": 4.0, "retries": 0, "seconds": 90.0, "max_memory_gb": 18.13, "ce": 0.18005754612386227, "kl": 0.051177936111344025, "context": 0.2056465158239007}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.9326037764549255, "loss_scale": 4.0, "retries": 0, "seconds": 90.6, "max_memory_gb": 18.13, "ce": 0.17881887406110764, "kl": 0.041851722984574735, "context": 0.19974473491311073}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 1.3367083072662354, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 18.13, "ce": 0.22033359669148922, "kl": 0.012366959010250866, "context": 0.22651707753539085}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 1.150089144706726, "loss_scale": 4.0, "retries": 0, "seconds": 91.7, "max_memory_gb": 18.13, "ce": 0.10485744755715132, "kl": 0.13919081684434786, "context": 0.17445285711437464}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 2.0990352630615234, "loss_scale": 4.0, "retries": 0, "seconds": 92.2, "max_memory_gb": 18.13, "ce": 0.28797953203320503, "kl": 1.0932244446594268, "context": 0.8345917407423258}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.7969131469726562, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 18.13, "ce": 0.0406992407515645, "kl": 0.0061233559681568295, "context": 0.04376091714948416}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 1.2513409852981567, "loss_scale": 4.0, "retries": 0, "seconds": 93.3, "max_memory_gb": 18.13, "ce": 0.0636693782871589, "kl": 0.20421651215292513, "context": 0.16577763273380697}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 1.084174633026123, "loss_scale": 4.0, "retries": 0, "seconds": 93.8, "max_memory_gb": 18.13, "ce": 0.0685448320582509, "kl": 0.010807325772475451, "context": 0.07394849602133036}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 1.2078890800476074, "loss_scale": 4.0, "retries": 0, "seconds": 94.3, "max_memory_gb": 18.13, "ce": 0.15645646397024393, "kl": 0.07328274985775352, "context": 0.19309783354401588}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.749703586101532, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 18.13, "ce": 0.10822360962629318, "kl": 0.05988895532209426, "context": 0.13816808816045523}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.7740225791931152, "loss_scale": 4.0, "retries": 0, "seconds": 95.4, "max_memory_gb": 18.13, "ce": 0.19783943891525269, "kl": 0.1894872933626175, "context": 0.2925830762833357}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.8113016486167908, "loss_scale": 4.0, "retries": 0, "seconds": 95.9, "max_memory_gb": 18.13, "ce": 0.048233140259981155, "kl": 0.015084716491401196, "context": 0.055775499902665615}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 2.0569939613342285, "loss_scale": 4.0, "retries": 0, "seconds": 96.4, "max_memory_gb": 18.13, "ce": 0.16236340254545212, "kl": 0.12647854042006657, "context": 0.22560267010703683}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.5253708362579346, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 18.13, "ce": 0.1679189121350646, "kl": 0.08110486180521548, "context": 0.20847133826464415}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 2.0701651573181152, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 18.13, "ce": 0.3493953205179423, "kl": 0.005588223313679919, "context": 0.3521894318982959}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 2.528369903564453, "loss_scale": 4.0, "retries": 0, "seconds": 98.0, "max_memory_gb": 18.13, "ce": 0.2689069313928485, "kl": 0.06838977831648663, "context": 0.30310181993991137}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 3.70477294921875, "loss_scale": 4.0, "retries": 0, "seconds": 98.6, "max_memory_gb": 18.13, "ce": 0.10607875557616353, "kl": 0.06662909057922661, "context": 0.1393932979553938}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 0.5641171932220459, "loss_scale": 4.0, "retries": 0, "seconds": 99.1, "max_memory_gb": 18.13, "ce": 0.17249889625236392, "kl": 0.006417280463210773, "context": 0.1757075332570821}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.13769543170928955, "loss_scale": 4.0, "retries": 0, "seconds": 99.7, "max_memory_gb": 18.13, "ce": 0.01273913576733321, "kl": 0.0050440376107872, "context": 0.015261154738254845}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.7576249837875366, "loss_scale": 4.0, "retries": 0, "seconds": 100.2, "max_memory_gb": 18.13, "ce": 0.043831250397488475, "kl": 0.0754086414235644, "context": 0.08153557474724948}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.21789203584194183, "loss_scale": 4.0, "retries": 0, "seconds": 100.7, "max_memory_gb": 18.13, "ce": 0.029292506049387157, "kl": 0.01563746836473001, "context": 0.03711123997345567}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.47756287455558777, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 18.13, "ce": 0.060712730046361685, "kl": 0.005667900084517896, "context": 0.06354668037965894}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.8112800121307373, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 18.13, "ce": 0.07262454903684556, "kl": 0.34846108398051, "context": 0.24685509479604661}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.30079740285873413, "loss_scale": 4.0, "retries": 0, "seconds": 102.3, "max_memory_gb": 18.13, "ce": 0.14830729831010103, "kl": 0.05246794707454683, "context": 0.17454126820666716}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.9136371612548828, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 18.13, "ce": 0.05633737798780203, "kl": 0.012202245532535017, "context": 0.06243849918246269}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.5060997009277344, "loss_scale": 4.0, "retries": 0, "seconds": 103.4, "max_memory_gb": 18.13, "ce": 0.01957144821062684, "kl": 0.45074672903865576, "context": 0.24494481785222888}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.5859802961349487, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 18.13, "ce": 0.2560108616016805, "kl": 0.06158142886124551, "context": 0.28680157102644444}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.45229604840278625, "loss_scale": 4.0, "retries": 0, "seconds": 104.5, "max_memory_gb": 18.13, "ce": 0.19059464358724654, "kl": 0.007753967744065449, "context": 0.19447162025608122}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.4361148476600647, "loss_scale": 4.0, "retries": 0, "seconds": 105.0, "max_memory_gb": 18.13, "ce": 0.03018669248558581, "kl": 0.035675415594596416, "context": 0.04802439908962697}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 0.985405683517456, "loss_scale": 4.0, "retries": 0, "seconds": 105.5, "max_memory_gb": 18.13, "ce": 0.32882828544825315, "kl": 0.023479077499359846, "context": 0.3405678207054734}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.30882251262664795, "loss_scale": 4.0, "retries": 0, "seconds": 106.1, "max_memory_gb": 18.13, "ce": 0.10658113704994321, "kl": 0.02103559742681682, "context": 0.11709893820807338}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.37845805287361145, "loss_scale": 4.0, "retries": 0, "seconds": 106.7, "max_memory_gb": 18.13, "ce": 0.01930234581232071, "kl": 0.025684562686365098, "context": 0.032144627533853054}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.3517100214958191, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 18.13, "ce": 0.11384389933664352, "kl": 0.0262584050942678, "context": 0.12697310140356421}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.44302237033843994, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 18.13, "ce": 0.17348051443696022, "kl": 0.041086035838816315, "context": 0.19402352953329682}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.5600108504295349, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 18.13, "ce": 0.07487664162181318, "kl": 0.28153559105703607, "context": 0.21564443595707417}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.5667054653167725, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 18.13, "ce": 0.2404058314859867, "kl": 0.010056163067929447, "context": 0.24543390609323978}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.441530704498291, "loss_scale": 4.0, "retries": 0, "seconds": 109.3, "max_memory_gb": 18.13, "ce": 0.1060524438507855, "kl": 0.016822436940856278, "context": 0.11446366203017533}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.06591080129146576, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 18.13, "ce": 0.006928950548171997, "kl": 0.003522709710523486, "context": 0.008690305519849062}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.7117025256156921, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 18.13, "ce": 0.16253717709332705, "kl": 0.11056876555085182, "context": 0.21782155707478523}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.8235529065132141, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 18.13, "ce": 0.12719767226371914, "kl": 0.02104450872866437, "context": 0.1377199268899858}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.7926010489463806, "loss_scale": 4.0, "retries": 0, "seconds": 111.5, "max_memory_gb": 18.13, "ce": 0.07726818922674283, "kl": 0.34889601278700866, "context": 0.25171618699096143}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.5448507070541382, "loss_scale": 4.0, "retries": 0, "seconds": 112.0, "max_memory_gb": 18.13, "ce": 0.12834951025433838, "kl": 0.09989057852726546, "context": 0.17829480743966997}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.3369747400283813, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 18.13, "ce": 0.5792446415871382, "kl": 0.45357338650501333, "context": 0.8060313160531223}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.13964498043060303, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 18.13, "ce": 0.01131854159757495, "kl": 0.005568102991674095, "context": 0.014102593064308167}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.32071638107299805, "loss_scale": 4.0, "retries": 0, "seconds": 113.6, "max_memory_gb": 18.13, "ce": 0.2099574557505548, "kl": 0.019603528548032045, "context": 0.21975922537967563}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.6344782710075378, "loss_scale": 4.0, "retries": 0, "seconds": 114.2, "max_memory_gb": 18.13, "ce": 0.2215642958180979, "kl": 0.09942618794593727, "context": 0.27127738611306995}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.47678324580192566, "loss_scale": 4.0, "retries": 0, "seconds": 114.7, "max_memory_gb": 18.13, "ce": 0.18556080700363964, "kl": 0.10224927100352943, "context": 0.2366854385472834}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.8566153049468994, "loss_scale": 4.0, "retries": 0, "seconds": 115.2, "max_memory_gb": 18.13, "ce": 0.1624177349731326, "kl": 0.7381514759035781, "context": 0.5314934821799397}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.6754644513130188, "loss_scale": 4.0, "retries": 0, "seconds": 115.8, "max_memory_gb": 18.13, "ce": 0.2938709738664329, "kl": 0.05795792474236805, "context": 0.3228499407414347}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 0.6156978607177734, "loss_scale": 4.0, "retries": 0, "seconds": 116.3, "max_memory_gb": 18.13, "ce": 0.2687686197459698, "kl": 0.3012234445195645, "context": 0.41938034212216735}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 0.9277918934822083, "loss_scale": 4.0, "retries": 0, "seconds": 116.9, "max_memory_gb": 18.13, "ce": 0.3568566986359656, "kl": 0.005816600758407731, "context": 0.35976500692777336}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 0.858354389667511, "loss_scale": 4.0, "retries": 0, "seconds": 117.4, "max_memory_gb": 18.13, "ce": 0.2194001330062747, "kl": 0.0818382318248041, "context": 0.26031925552524626}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.35891467332839966, "loss_scale": 4.0, "retries": 0, "seconds": 118.0, "max_memory_gb": 18.13, "ce": 0.06399082473944873, "kl": 0.036905801214743406, "context": 0.08244372787885368}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.5289703011512756, "loss_scale": 4.0, "retries": 0, "seconds": 118.5, "max_memory_gb": 18.13, "ce": 0.17706957599148154, "kl": 0.406989311799407, "context": 0.3805642258375883}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.40021035075187683, "loss_scale": 4.0, "retries": 0, "seconds": 119.0, "max_memory_gb": 18.13, "ce": 0.11999806761741638, "kl": 0.03175283211749047, "context": 0.13587448000907898}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.4500848352909088, "loss_scale": 4.0, "retries": 0, "seconds": 119.6, "max_memory_gb": 18.13, "ce": 0.09710484300740063, "kl": 0.017762708361260593, "context": 0.10598619910888374}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.5481495261192322, "loss_scale": 4.0, "retries": 0, "seconds": 120.1, "max_memory_gb": 18.13, "ce": 0.0915725517552346, "kl": 0.1316800395725295, "context": 0.15741257462650537}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.2121951580047607, "loss_scale": 4.0, "retries": 0, "seconds": 120.6, "max_memory_gb": 18.13, "ce": 0.23483041767030954, "kl": 0.4046228672086727, "context": 0.437141852453351}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.49996063113212585, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 18.13, "ce": 0.2040048912167549, "kl": 0.027154168812558055, "context": 0.21758197713643312}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.63104647397995, "loss_scale": 4.0, "retries": 0, "seconds": 121.7, "max_memory_gb": 18.13, "ce": 0.0817861994728446, "kl": 0.13728588426602073, "context": 0.1504291477613151}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.4173734486103058, "loss_scale": 4.0, "retries": 0, "seconds": 122.3, "max_memory_gb": 18.13, "ce": 0.06045149266719818, "kl": 0.05501490971073508, "context": 0.08795894868671894}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.5643438100814819, "loss_scale": 4.0, "retries": 0, "seconds": 122.8, "max_memory_gb": 18.13, "ce": 0.0994962602853775, "kl": 0.01718969497596845, "context": 0.10809110524132848}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.15058104693889618, "loss_scale": 4.0, "retries": 0, "seconds": 123.3, "max_memory_gb": 18.13, "ce": 0.016416559228673577, "kl": 0.00784880924038589, "context": 0.020340963499620557}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.6605556011199951, "loss_scale": 4.0, "retries": 0, "seconds": 123.9, "max_memory_gb": 18.13, "ce": 0.07779816607944667, "kl": 0.12311439029872417, "context": 0.1393553614616394}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.8962177038192749, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 18.13, "ce": 0.2611198795493692, "kl": 0.16293777618557215, "context": 0.3425887692719698}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.8034191131591797, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 18.13, "ce": 0.28531993879005313, "kl": 0.11684677167795599, "context": 0.34374332253355533}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.42610445618629456, "loss_scale": 4.0, "retries": 0, "seconds": 125.5, "max_memory_gb": 18.13, "ce": 0.22480911458842456, "kl": 0.02258603076916188, "context": 0.2361021307297051}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.26531127095222473, "loss_scale": 4.0, "retries": 0, "seconds": 126.0, "max_memory_gb": 18.13, "ce": 0.024080035742372274, "kl": 0.006909238523803651, "context": 0.02753465506248176}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.1359611749649048, "loss_scale": 4.0, "retries": 0, "seconds": 126.5, "max_memory_gb": 18.13, "ce": 0.1377396285533905, "kl": 0.13357503719089436, "context": 0.2045271396636963}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.5371825695037842, "loss_scale": 4.0, "retries": 0, "seconds": 127.1, "max_memory_gb": 18.13, "ce": 0.1761389367748052, "kl": 0.018252923196996562, "context": 0.18526540067978203}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.6950005888938904, "loss_scale": 4.0, "retries": 0, "seconds": 127.6, "max_memory_gb": 18.13, "ce": 0.2939075252506882, "kl": 0.031507080275332555, "context": 0.30966106813866645}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.4432331919670105, "loss_scale": 4.0, "retries": 0, "seconds": 128.1, "max_memory_gb": 18.13, "ce": 0.08853041438851506, "kl": 0.030379981006262824, "context": 0.10372040746733546}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 0.6383348107337952, "loss_scale": 4.0, "retries": 0, "seconds": 128.7, "max_memory_gb": 18.13, "ce": 0.19109724462032318, "kl": 0.05745736215430952, "context": 0.21982592856511474}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.7280159592628479, "loss_scale": 4.0, "retries": 0, "seconds": 129.2, "max_memory_gb": 18.13, "ce": 0.04279005102580413, "kl": 0.11832760493234673, "context": 0.10195385327097028}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.5266713500022888, "loss_scale": 4.0, "retries": 0, "seconds": 129.7, "max_memory_gb": 18.13, "ce": 0.19503800221718848, "kl": 0.030516765196807683, "context": 0.21029638219624758}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.5781900882720947, "loss_scale": 4.0, "retries": 0, "seconds": 130.2, "max_memory_gb": 18.13, "ce": 0.10860187350772321, "kl": 0.0390822229674086, "context": 0.12814298458397388}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.5471411943435669, "loss_scale": 4.0, "retries": 0, "seconds": 130.8, "max_memory_gb": 18.13, "ce": 0.1710886349901557, "kl": 0.016016502879210748, "context": 0.17909688223153353}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.4809970259666443, "loss_scale": 4.0, "retries": 0, "seconds": 131.3, "max_memory_gb": 18.13, "ce": 0.13143495004624128, "kl": 0.03944677839172073, "context": 0.15115834446623921}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.7776051163673401, "loss_scale": 4.0, "retries": 0, "seconds": 131.8, "max_memory_gb": 18.13, "ce": 0.02897236729040742, "kl": 0.14652578669483773, "context": 0.10223525669425726}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 0.9960392117500305, "loss_scale": 4.0, "retries": 0, "seconds": 132.4, "max_memory_gb": 18.13, "ce": 0.27200488426024094, "kl": 0.06020127126976149, "context": 0.30210553138749674}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.4609457552433014, "loss_scale": 4.0, "retries": 0, "seconds": 132.9, "max_memory_gb": 18.13, "ce": 0.3748457345645875, "kl": 0.011904943210538477, "context": 0.3807982038706541}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.3140625059604645, "loss_scale": 4.0, "retries": 0, "seconds": 133.5, "max_memory_gb": 18.13, "ce": 0.026640018390025944, "kl": 0.035228663466114085, "context": 0.044254351989366114}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.5438741445541382, "loss_scale": 4.0, "retries": 0, "seconds": 134.0, "max_memory_gb": 18.13, "ce": 0.10078005702234805, "kl": 0.035910385893657804, "context": 0.11873525194823742}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.5118505954742432, "loss_scale": 4.0, "retries": 0, "seconds": 134.5, "max_memory_gb": 18.13, "ce": 0.08163572009652853, "kl": 0.03008634364232421, "context": 0.09667889354750514}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.4814721345901489, "loss_scale": 4.0, "retries": 0, "seconds": 135.1, "max_memory_gb": 18.13, "ce": 0.07912512123584747, "kl": 0.03589636131073348, "context": 0.09707330213859677}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.3975561857223511, "loss_scale": 4.0, "retries": 0, "seconds": 135.6, "max_memory_gb": 18.13, "ce": 0.05105967971030623, "kl": 0.015466272994672181, "context": 0.05879281694069505}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.4406983256340027, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 18.13, "ce": 0.13506706879707053, "kl": 0.01888157178836991, "context": 0.14450785232475027}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.5211105942726135, "loss_scale": 4.0, "retries": 0, "seconds": 136.7, "max_memory_gb": 18.13, "ce": 0.03098157001659274, "kl": 0.12094377016182989, "context": 0.09145345445722342}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.2627886235713959, "loss_scale": 4.0, "retries": 0, "seconds": 137.2, "max_memory_gb": 18.13, "ce": 0.103332111495547, "kl": 0.009549137957947096, "context": 0.1081066785263829}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.9071010947227478, "loss_scale": 4.0, "retries": 0, "seconds": 137.7, "max_memory_gb": 18.13, "ce": 0.1539606391452253, "kl": 0.29828153760172427, "context": 0.30310140550136566}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.7749814987182617, "loss_scale": 4.0, "retries": 0, "seconds": 138.3, "max_memory_gb": 18.13, "ce": 0.16217752639204264, "kl": 0.2554493471980095, "context": 0.2899022027850151}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.31149986386299133, "loss_scale": 4.0, "retries": 0, "seconds": 138.8, "max_memory_gb": 18.13, "ce": 0.07861968816723675, "kl": 0.02770103259535972, "context": 0.09247020306065679}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.5438053607940674, "loss_scale": 4.0, "retries": 0, "seconds": 139.3, "max_memory_gb": 18.13, "ce": 0.08198045520111918, "kl": 0.012611764888788457, "context": 0.08828633982921019}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.45823752880096436, "loss_scale": 4.0, "retries": 0, "seconds": 139.9, "max_memory_gb": 18.13, "ce": 0.0487546986551024, "kl": 0.10767797098378651, "context": 0.10259368520928547}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.2420637607574463, "loss_scale": 4.0, "retries": 0, "seconds": 140.4, "max_memory_gb": 18.13, "ce": 0.017390242777764797, "kl": 0.02691588582820259, "context": 0.03084818576462567}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.6445085406303406, "loss_scale": 4.0, "retries": 0, "seconds": 140.9, "max_memory_gb": 18.13, "ce": 0.0908955973573029, "kl": 0.1415486251062248, "context": 0.16166990902274847}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.6361470222473145, "loss_scale": 4.0, "retries": 0, "seconds": 141.5, "max_memory_gb": 18.13, "ce": 0.14180783834308386, "kl": 0.10620982921682298, "context": 0.19491275772452354}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.9187567234039307, "loss_scale": 4.0, "retries": 0, "seconds": 142.0, "max_memory_gb": 18.13, "ce": 0.21029246947728097, "kl": 0.12293788089300506, "context": 0.2717614097055048}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.7292226552963257, "loss_scale": 4.0, "retries": 0, "seconds": 142.5, "max_memory_gb": 18.13, "ce": 0.18557070707902312, "kl": 0.04629110952373594, "context": 0.2087162653915584}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 1.0540428161621094, "loss_scale": 4.0, "retries": 0, "seconds": 143.1, "max_memory_gb": 18.13, "ce": 0.15371998795308173, "kl": 0.0691586494067451, "context": 0.1882993052713573}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.40358272194862366, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 18.13, "ce": 0.08690175705123693, "kl": 0.10528876772150397, "context": 0.1395461419597268}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 1.8929853439331055, "loss_scale": 4.0, "retries": 0, "seconds": 144.1, "max_memory_gb": 18.13, "ce": 0.4133352508652024, "kl": 0.16656869834696408, "context": 0.4966195967281237}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.3285132050514221, "loss_scale": 4.0, "retries": 0, "seconds": 144.7, "max_memory_gb": 18.13, "ce": 0.14963796827942133, "kl": 0.008770997956162319, "context": 0.1540234696585685}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.36200904846191406, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 18.13, "ce": 0.23941218433901668, "kl": 0.06612648773443652, "context": 0.27247542841359973}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 3.019500255584717, "loss_scale": 4.0, "retries": 0, "seconds": 145.8, "max_memory_gb": 18.13, "ce": 0.3731282725930214, "kl": 0.6783356154155626, "context": 0.7122960761189461}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.8001363277435303, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 18.13, "ce": 0.1816729698330164, "kl": 0.003799180361966137, "context": 0.1835725570563227}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.23387005925178528, "loss_scale": 4.0, "retries": 0, "seconds": 146.8, "max_memory_gb": 18.13, "ce": 0.036261651082895696, "kl": 0.04930055147269741, "context": 0.06091192667372525}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 1.0591156482696533, "loss_scale": 4.0, "retries": 0, "seconds": 147.4, "max_memory_gb": 18.13, "ce": 0.020253112306818366, "kl": 0.43253869142063195, "context": 0.23652245674747974}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.6651121973991394, "loss_scale": 4.0, "retries": 0, "seconds": 147.9, "max_memory_gb": 18.13, "ce": 0.01649898913456127, "kl": 0.15197483159136027, "context": 0.0924864059779793}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.5239766240119934, "loss_scale": 4.0, "retries": 0, "seconds": 148.4, "max_memory_gb": 18.13, "ce": 0.12221666984260082, "kl": 0.041661810129880905, "context": 0.143047577701509}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 1.1898967027664185, "loss_scale": 4.0, "retries": 0, "seconds": 149.0, "max_memory_gb": 18.13, "ce": 0.5134525602916256, "kl": 0.39563367469236255, "context": 0.7112693914677948}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.2265404462814331, "loss_scale": 4.0, "retries": 0, "seconds": 149.5, "max_memory_gb": 18.13, "ce": 0.010732293885666877, "kl": 0.014558953273080988, "context": 0.01801177058950998}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.34556084871292114, "loss_scale": 4.0, "retries": 0, "seconds": 150.0, "max_memory_gb": 18.13, "ce": 0.01289805886335671, "kl": 0.1315635348437354, "context": 0.07867982634343207}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 1.0877676010131836, "loss_scale": 4.0, "retries": 0, "seconds": 150.6, "max_memory_gb": 18.13, "ce": 0.17308247281471267, "kl": 0.1994182348717004, "context": 0.27279159365571104}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 0.8853962421417236, "loss_scale": 4.0, "retries": 0, "seconds": 151.1, "max_memory_gb": 18.13, "ce": 0.0910557599272579, "kl": 0.1150204383811797, "context": 0.14856597827747464}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.2583470940589905, "loss_scale": 4.0, "retries": 0, "seconds": 151.6, "max_memory_gb": 18.13, "ce": 0.019970238558016717, "kl": 0.012673722478211857, "context": 0.02630710043013096}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.19614839553833, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 18.13, "ce": 0.2523849457502365, "kl": 0.09519576613092795, "context": 0.2999828215688467}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.349509060382843, "loss_scale": 4.0, "retries": 0, "seconds": 152.7, "max_memory_gb": 18.13, "ce": 0.09067280570161529, "kl": 0.009137467970504076, "context": 0.09524154014070518}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.46634039282798767, "loss_scale": 4.0, "retries": 0, "seconds": 153.2, "max_memory_gb": 18.13, "ce": 0.07318844040855765, "kl": 0.1501952586841071, "context": 0.1482860692194663}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 0.6463713049888611, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 18.13, "ce": 0.07067645504139364, "kl": 0.018083606912114192, "context": 0.0797182593960315}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.39848846197128296, "loss_scale": 4.0, "retries": 0, "seconds": 154.3, "max_memory_gb": 18.13, "ce": 0.11921580624766648, "kl": 0.026110363949555904, "context": 0.13227098900824785}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.1983410120010376, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 18.13, "ce": 0.014923833718057722, "kl": 0.018867596954805776, "context": 0.02435763319954276}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.7442871928215027, "loss_scale": 4.0, "retries": 0, "seconds": 155.4, "max_memory_gb": 18.13, "ce": 0.09440000052563846, "kl": 0.03067505406215787, "context": 0.10973753174766898}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.12921957671642303, "loss_scale": 4.0, "retries": 0, "seconds": 155.9, "max_memory_gb": 18.13, "ce": 0.009316325653344393, "kl": 0.009857131430180743, "context": 0.014244891470298171}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.2741507589817047, "loss_scale": 4.0, "retries": 0, "seconds": 156.4, "max_memory_gb": 18.13, "ce": 0.045335418777540326, "kl": 0.08445929861045443, "context": 0.08756506792269647}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.6554073691368103, "loss_scale": 4.0, "retries": 0, "seconds": 157.0, "max_memory_gb": 18.13, "ce": 0.13579147803829983, "kl": 0.06440483043479617, "context": 0.1679938858433161}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.06726719439029694, "loss_scale": 4.0, "retries": 0, "seconds": 157.5, "max_memory_gb": 18.13, "ce": 0.005778747407021001, "kl": 0.0028597783148143208, "context": 0.00720863655442372}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.2766884565353394, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 18.13, "ce": 0.07074425788596272, "kl": 0.7156100384891033, "context": 0.4285492654889822}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.783372163772583, "loss_scale": 4.0, "retries": 0, "seconds": 158.6, "max_memory_gb": 18.13, "ce": 0.05754755507223308, "kl": 0.15468147318460979, "context": 0.13488829252310097}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 0.914565920829773, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 18.13, "ce": 0.06795249041169882, "kl": 0.030962342279963195, "context": 0.08343365998007357}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 1.3352957963943481, "loss_scale": 4.0, "retries": 0, "seconds": 159.6, "max_memory_gb": 18.13, "ce": 0.17926661134697497, "kl": 0.11092903731332626, "context": 0.23473113705404103}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 1.3997856378555298, "loss_scale": 4.0, "retries": 0, "seconds": 160.2, "max_memory_gb": 18.13, "ce": 0.2516310317441821, "kl": 0.4112283686408773, "context": 0.45724522322416306}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.5481031537055969, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 18.13, "ce": 0.1123112792847678, "kl": 0.01072049024514854, "context": 0.11767152301035821}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.1478464603424072, "loss_scale": 4.0, "retries": 0, "seconds": 161.2, "max_memory_gb": 18.13, "ce": 0.4573989117052406, "kl": 0.176290275354404, "context": 0.5455440494697541}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.31634485721588135, "loss_scale": 4.0, "retries": 0, "seconds": 161.8, "max_memory_gb": 18.13, "ce": 0.018727322341874242, "kl": 0.11566468709497713, "context": 0.07655966625316069}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.5719009637832642, "loss_scale": 4.0, "retries": 0, "seconds": 162.3, "max_memory_gb": 18.13, "ce": 0.2757548422086984, "kl": 0.04803247991731041, "context": 0.2997710796771571}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 1.0872371196746826, "loss_scale": 4.0, "retries": 0, "seconds": 162.8, "max_memory_gb": 18.13, "ce": 0.13557271449826658, "kl": 0.12432798632289632, "context": 0.19773670777794905}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.18450553715229034, "loss_scale": 4.0, "retries": 0, "seconds": 163.4, "max_memory_gb": 18.13, "ce": 0.02167699090205133, "kl": 0.006053465433069505, "context": 0.02470372337847948}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.4690784811973572, "loss_scale": 4.0, "retries": 0, "seconds": 163.9, "max_memory_gb": 18.13, "ce": 0.025993385585024953, "kl": 0.05216132776331506, "context": 0.0520740479696542}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.24574097990989685, "loss_scale": 4.0, "retries": 0, "seconds": 164.4, "max_memory_gb": 18.13, "ce": 0.023073100368492305, "kl": 0.0071718576800776646, "context": 0.026659028721041977}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 3.6119930744171143, "loss_scale": 4.0, "retries": 0, "seconds": 165.0, "max_memory_gb": 18.13, "ce": 0.12817610811907798, "kl": 0.1932943853607867, "context": 0.2248233009595424}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 1.249635100364685, "loss_scale": 4.0, "retries": 0, "seconds": 165.5, "max_memory_gb": 18.13, "ce": 0.010953821358270943, "kl": 0.18879628495778888, "context": 0.10535196610726416}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.6838323473930359, "loss_scale": 4.0, "retries": 0, "seconds": 166.0, "max_memory_gb": 18.13, "ce": 0.11536776600405574, "kl": 0.011870896993059432, "context": 0.12130321632139385}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 0.8931965231895447, "loss_scale": 4.0, "retries": 0, "seconds": 166.6, "max_memory_gb": 18.13, "ce": 0.0275835677748546, "kl": 0.6961840914154891, "context": 0.3756756002549082}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 0.910308837890625, "loss_scale": 4.0, "retries": 0, "seconds": 167.1, "max_memory_gb": 18.13, "ce": 0.21760360477492213, "kl": 0.013849008159013465, "context": 0.22452811093535274}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.5634714365005493, "loss_scale": 4.0, "retries": 0, "seconds": 167.6, "max_memory_gb": 18.13, "ce": 0.030057017458602786, "kl": 0.23249720933381468, "context": 0.14630562742240727}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.4647507965564728, "loss_scale": 4.0, "retries": 0, "seconds": 168.2, "max_memory_gb": 18.13, "ce": 0.12935228447895497, "kl": 0.017625038424739614, "context": 0.13816480396781117}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.32678887248039246, "loss_scale": 4.0, "retries": 0, "seconds": 168.7, "max_memory_gb": 18.13, "ce": 0.025730452151037753, "kl": 0.01230764102365356, "context": 0.03188427194254473}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.9174039959907532, "loss_scale": 4.0, "retries": 0, "seconds": 169.2, "max_memory_gb": 18.13, "ce": 0.14251825280371122, "kl": 0.03456980176270008, "context": 0.15980315738124773}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.5572632551193237, "loss_scale": 4.0, "retries": 0, "seconds": 169.8, "max_memory_gb": 18.13, "ce": 0.1458021472208202, "kl": 0.0885100236046128, "context": 0.1900571584701538}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.44850629568099976, "loss_scale": 4.0, "retries": 0, "seconds": 170.3, "max_memory_gb": 18.13, "ce": 0.06009350181557238, "kl": 0.05206902860663831, "context": 0.086128014838323}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 1.083039402961731, "loss_scale": 4.0, "retries": 0, "seconds": 170.8, "max_memory_gb": 18.13, "ce": 0.19285017810761929, "kl": 0.614505156641826, "context": 0.5001027509570122}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.7367078065872192, "loss_scale": 4.0, "retries": 0, "seconds": 171.4, "max_memory_gb": 18.13, "ce": 0.15970505960285664, "kl": 0.020112639293074608, "context": 0.16976138204336166}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.641428530216217, "loss_scale": 4.0, "retries": 0, "seconds": 171.9, "max_memory_gb": 18.13, "ce": 0.04393380228430033, "kl": 0.3596039368712809, "context": 0.22373577253893018}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 0.6575608253479004, "loss_scale": 4.0, "retries": 0, "seconds": 172.4, "max_memory_gb": 18.13, "ce": 0.07611557026393712, "kl": 0.1260551901650615, "context": 0.13914316520094872}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 0.9863933324813843, "loss_scale": 4.0, "retries": 0, "seconds": 173.0, "max_memory_gb": 18.13, "ce": 0.14890545583330095, "kl": 0.0813500089734589, "context": 0.1895804621744901}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.6970862746238708, "loss_scale": 4.0, "retries": 0, "seconds": 173.5, "max_memory_gb": 18.13, "ce": 0.08134434605017304, "kl": 0.08697016691439785, "context": 0.1248294310644269}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.29004931449890137, "loss_scale": 4.0, "retries": 0, "seconds": 174.0, "max_memory_gb": 18.13, "ce": 0.02345211908686906, "kl": 0.04367734097468201, "context": 0.04529078886844218}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.7430822253227234, "loss_scale": 4.0, "retries": 0, "seconds": 174.6, "max_memory_gb": 18.13, "ce": 0.012411714531481266, "kl": 0.1566028604283929, "context": 0.09071314474567771}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.2165803760290146, "loss_scale": 4.0, "retries": 0, "seconds": 175.1, "max_memory_gb": 18.13, "ce": 0.02171084238216281, "kl": 0.010897365915297996, "context": 0.027159525314345956}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.626341700553894, "loss_scale": 4.0, "retries": 0, "seconds": 175.6, "max_memory_gb": 18.13, "ce": 0.2223193707759492, "kl": 0.028526564132334897, "context": 0.23658266081474721}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.4268134832382202, "loss_scale": 4.0, "retries": 0, "seconds": 176.1, "max_memory_gb": 18.13, "ce": 0.058747304312419146, "kl": 0.052896585839334875, "context": 0.08519559679552913}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 2.478804588317871, "loss_scale": 4.0, "retries": 0, "seconds": 176.7, "max_memory_gb": 18.13, "ce": 0.23244936001719907, "kl": 0.08912786360451719, "context": 0.2770132990553975}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.9678987860679626, "loss_scale": 4.0, "retries": 0, "seconds": 177.2, "max_memory_gb": 18.13, "ce": 0.11772536393254995, "kl": 0.16917031434422825, "context": 0.20231051929295063}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.6930617690086365, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 18.13, "ce": 0.11437333119101822, "kl": 0.15901246391877066, "context": 0.19387956289574504}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.46455976366996765, "loss_scale": 4.0, "retries": 0, "seconds": 178.3, "max_memory_gb": 18.13, "ce": 0.15498760726768523, "kl": 0.0076245766977081075, "context": 0.1587998932809569}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.8051185011863708, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 18.13, "ce": 0.26066880766302347, "kl": 0.09995841234922409, "context": 0.31064801197499037}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.3107939958572388, "loss_scale": 4.0, "retries": 0, "seconds": 179.3, "max_memory_gb": 18.13, "ce": 0.2061341756489128, "kl": 0.05576012718665879, "context": 0.23401423287577927}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 1.0067378282546997, "loss_scale": 4.0, "retries": 0, "seconds": 179.9, "max_memory_gb": 18.13, "ce": 0.31384811783209443, "kl": 0.033256494672968984, "context": 0.3304763650521636}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 1.0415083169937134, "loss_scale": 4.0, "retries": 0, "seconds": 180.4, "max_memory_gb": 18.13, "ce": 0.17001572728622705, "kl": 0.13787504634819925, "context": 0.2389532569795847}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 1.969709038734436, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 18.13, "ce": 0.34562497958540916, "kl": 0.2879151402739808, "context": 0.48958256747573614}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.23808278143405914, "loss_scale": 4.0, "retries": 0, "seconds": 181.5, "max_memory_gb": 18.13, "ce": 0.014377410698216408, "kl": 0.06613633900997229, "context": 0.04744557902449742}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 0.6771491765975952, "loss_scale": 4.0, "retries": 0, "seconds": 182.0, "max_memory_gb": 18.13, "ce": 0.18960639089345932, "kl": 0.09546255972236395, "context": 0.23733767122030258}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.141492486000061, "loss_scale": 4.0, "retries": 0, "seconds": 182.6, "max_memory_gb": 18.13, "ce": 0.22626706631854177, "kl": 0.16120831295847893, "context": 0.30687121395021677}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.696418285369873, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 18.13, "ce": 0.1463359512272291, "kl": 0.013799579937767703, "context": 0.15323573420755565}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.9866936802864075, "loss_scale": 4.0, "retries": 0, "seconds": 183.6, "max_memory_gb": 18.13, "ce": 0.2175189801491797, "kl": 0.09011382004246116, "context": 0.2625758901704103}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 0.9746090173721313, "loss_scale": 4.0, "retries": 0, "seconds": 184.2, "max_memory_gb": 18.13, "ce": 0.16022584564052522, "kl": 0.03602508921176195, "context": 0.17823838721960783}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 1.5722509622573853, "loss_scale": 4.0, "retries": 0, "seconds": 184.7, "max_memory_gb": 18.13, "ce": 0.2538867898983881, "kl": 0.9311323587899096, "context": 0.7194529734551907}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.2989123463630676, "loss_scale": 4.0, "retries": 0, "seconds": 185.2, "max_memory_gb": 18.13, "ce": 0.016306769335642457, "kl": 0.047533344448311254, "context": 0.040073440643027425}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.784498393535614, "loss_scale": 4.0, "retries": 0, "seconds": 185.8, "max_memory_gb": 18.13, "ce": 0.022348262136802077, "kl": 0.04179720295360312, "context": 0.04324686387553811}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.4955652952194214, "loss_scale": 4.0, "retries": 0, "seconds": 186.3, "max_memory_gb": 18.13, "ce": 0.1430146197672002, "kl": 0.18921464096638374, "context": 0.23762194212758914}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.8561049699783325, "loss_scale": 4.0, "retries": 0, "seconds": 186.8, "max_memory_gb": 18.13, "ce": 0.20103250368265435, "kl": 0.006703335006022826, "context": 0.20438417582772672}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 1.0286121368408203, "loss_scale": 4.0, "retries": 0, "seconds": 187.4, "max_memory_gb": 18.13, "ce": 0.2214655540883541, "kl": 0.026755719212815166, "context": 0.23484341893345118}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.12291893362998962, "loss_scale": 4.0, "retries": 0, "seconds": 187.9, "max_memory_gb": 18.13, "ce": 0.01115021127043292, "kl": 0.004813931445823982, "context": 0.013557177153415978}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.18246489763259888, "loss_scale": 4.0, "retries": 0, "seconds": 188.4, "max_memory_gb": 18.13, "ce": 0.10582899849396199, "kl": 0.02235066643333994, "context": 0.1170043307938613}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.7355397343635559, "loss_scale": 4.0, "retries": 0, "seconds": 189.0, "max_memory_gb": 18.13, "ce": 0.15669849887490273, "kl": 0.050795241608284414, "context": 0.18209612369537354}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.1352393627166748, "loss_scale": 4.0, "retries": 0, "seconds": 189.5, "max_memory_gb": 18.13, "ce": 0.042065642890520394, "kl": 0.40797299228142947, "context": 0.24605213897302747}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.988677740097046, "loss_scale": 4.0, "retries": 0, "seconds": 190.0, "max_memory_gb": 18.13, "ce": 0.1951778964139521, "kl": 0.3950110534206033, "context": 0.392683420330286}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.8584052920341492, "loss_scale": 4.0, "retries": 0, "seconds": 190.5, "max_memory_gb": 18.13, "ce": 0.13836240954697132, "kl": 0.058259495403035544, "context": 0.16749216488096863}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.523113489151001, "loss_scale": 4.0, "retries": 0, "seconds": 191.1, "max_memory_gb": 18.13, "ce": 0.1824417457683012, "kl": 0.029878427332732826, "context": 0.19738095521461219}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.5091478228569031, "loss_scale": 4.0, "retries": 0, "seconds": 191.6, "max_memory_gb": 18.13, "ce": 0.15380693721817806, "kl": 0.024443782109301537, "context": 0.16602882172446698}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.41736170649528503, "loss_scale": 4.0, "retries": 0, "seconds": 192.1, "max_memory_gb": 18.13, "ce": 0.04019689839333296, "kl": 0.022944970784010366, "context": 0.05166938516777009}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.5663952231407166, "loss_scale": 4.0, "retries": 0, "seconds": 192.7, "max_memory_gb": 18.13, "ce": 0.06778512400342152, "kl": 0.020839903270825744, "context": 0.07820507662836462}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 1.0940155982971191, "loss_scale": 4.0, "retries": 0, "seconds": 193.2, "max_memory_gb": 18.13, "ce": 0.24396851425990462, "kl": 0.01403059670701623, "context": 0.25098380958661437}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.3296302258968353, "loss_scale": 4.0, "retries": 0, "seconds": 193.8, "max_memory_gb": 18.13, "ce": 0.028277406643610448, "kl": 0.020226855223882012, "context": 0.03839083289494738}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.7030841112136841, "loss_scale": 4.0, "retries": 0, "seconds": 194.3, "max_memory_gb": 18.13, "ce": 0.07480380852939561, "kl": 0.014321919996291399, "context": 0.0819647689932026}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.2727309465408325, "loss_scale": 4.0, "retries": 0, "seconds": 194.8, "max_memory_gb": 18.13, "ce": 0.06218502763658762, "kl": 0.4830419609643286, "context": 0.3037060017231852}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 1.8512554168701172, "loss_scale": 4.0, "retries": 0, "seconds": 195.4, "max_memory_gb": 18.13, "ce": 0.3265358933422249, "kl": 0.05449975685769459, "context": 0.35378577216761187}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.6590532064437866, "loss_scale": 4.0, "retries": 0, "seconds": 195.9, "max_memory_gb": 18.13, "ce": 0.061953263357281685, "kl": 0.12107956036925316, "context": 0.12249304121360183}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.8247246742248535, "loss_scale": 4.0, "retries": 0, "seconds": 196.5, "max_memory_gb": 18.13, "ce": 0.015873170574195683, "kl": 0.6435450366952864, "context": 0.3376456836849684}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.6658159494400024, "loss_scale": 4.0, "retries": 0, "seconds": 197.0, "max_memory_gb": 18.13, "ce": 0.14489704242441803, "kl": 0.018929134304926265, "context": 0.15436160587705672}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.6215296983718872, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 18.13, "ce": 0.058122911024838686, "kl": 0.280995409557363, "context": 0.19862060993909836}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 1.088510274887085, "loss_scale": 4.0, "retries": 0, "seconds": 198.1, "max_memory_gb": 18.13, "ce": 0.15108130383305252, "kl": 0.11007330639040447, "context": 0.20611795352306217}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.2559429705142975, "loss_scale": 4.0, "retries": 0, "seconds": 198.6, "max_memory_gb": 18.13, "ce": 0.019115111092105508, "kl": 0.033494409390186775, "context": 0.03586231672670692}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.2994956970214844, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 18.13, "ce": 0.04537457518745214, "kl": 0.01609112958976766, "context": 0.05342014110647142}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 1.9981343746185303, "loss_scale": 4.0, "retries": 0, "seconds": 199.7, "max_memory_gb": 18.13, "ce": 0.34453762881457806, "kl": 0.1655000827740878, "context": 0.427287669852376}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 0.9418389797210693, "loss_scale": 4.0, "retries": 0, "seconds": 200.3, "max_memory_gb": 18.13, "ce": 0.17549912963295355, "kl": 0.02391001670002879, "context": 0.18745414540171623}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.3761734068393707, "loss_scale": 4.0, "retries": 0, "seconds": 200.8, "max_memory_gb": 18.13, "ce": 0.04339196241926402, "kl": 0.023439516211510636, "context": 0.055111720226705074}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.3938453197479248, "loss_scale": 4.0, "retries": 0, "seconds": 201.3, "max_memory_gb": 18.13, "ce": 0.09072226635180414, "kl": 0.050713309610728174, "context": 0.11607892322354019}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.7974587082862854, "loss_scale": 4.0, "retries": 0, "seconds": 201.8, "max_memory_gb": 18.13, "ce": 0.13926017249468714, "kl": 0.12848235159799515, "context": 0.203501352109015}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.7029949426651001, "loss_scale": 4.0, "retries": 0, "seconds": 202.4, "max_memory_gb": 18.13, "ce": 0.1824398769531399, "kl": 0.15837380565790227, "context": 0.2616267802659422}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.7520042061805725, "loss_scale": 4.0, "retries": 0, "seconds": 202.9, "max_memory_gb": 18.13, "ce": 0.0689277850324288, "kl": 0.04912508801498916, "context": 0.09349033108446747}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.5511409044265747, "loss_scale": 4.0, "retries": 0, "seconds": 203.5, "max_memory_gb": 18.13, "ce": 0.05654632917139679, "kl": 0.13821568012645002, "context": 0.12565416970755905}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.7172291278839111, "loss_scale": 4.0, "retries": 0, "seconds": 204.0, "max_memory_gb": 18.13, "ce": 0.08790628053247929, "kl": 0.04091383336344734, "context": 0.10836319904774427}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.32877451181411743, "loss_scale": 4.0, "retries": 0, "seconds": 204.5, "max_memory_gb": 18.13, "ce": 0.07635526719968766, "kl": 0.05642270518001169, "context": 0.10456662159413099}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.9728020429611206, "loss_scale": 4.0, "retries": 0, "seconds": 205.0, "max_memory_gb": 18.13, "ce": 0.03865350363776088, "kl": 0.1594270322821103, "context": 0.11836702190339565}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.5661332607269287, "loss_scale": 4.0, "retries": 0, "seconds": 205.6, "max_memory_gb": 18.13, "ce": 0.24404388456605375, "kl": 0.08360433508642018, "context": 0.2858460578136146}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 2.5758962631225586, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 18.13, "ce": 0.328239728230983, "kl": 0.35914920072536916, "context": 0.5078143214341253}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.4183306396007538, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 18.13, "ce": 0.0838389991549775, "kl": 0.06529061975015793, "context": 0.11648430925561115}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.25796017050743103, "loss_scale": 4.0, "retries": 0, "seconds": 207.2, "max_memory_gb": 18.13, "ce": 0.011619541444815695, "kl": 0.03070120506163221, "context": 0.026970144361257553}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.5240888595581055, "loss_scale": 4.0, "retries": 0, "seconds": 207.7, "max_memory_gb": 18.13, "ce": 0.16694386024028063, "kl": 0.0020170924144622404, "context": 0.167952413787134}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.5011123418807983, "loss_scale": 4.0, "retries": 0, "seconds": 208.2, "max_memory_gb": 18.13, "ce": 0.05572506913449615, "kl": 0.009181067547615385, "context": 0.06031560304109007}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.40264537930488586, "loss_scale": 4.0, "retries": 0, "seconds": 208.8, "max_memory_gb": 18.13, "ce": 0.07295194093603641, "kl": 0.014501827186904848, "context": 0.08020285831298679}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.25601741671562195, "loss_scale": 4.0, "retries": 0, "seconds": 209.3, "max_memory_gb": 18.13, "ce": 0.04365852795308456, "kl": 0.005040520030888729, "context": 0.04617878815042786}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.5448484420776367, "loss_scale": 4.0, "retries": 0, "seconds": 209.9, "max_memory_gb": 18.13, "ce": 0.04296704102307558, "kl": 0.0387480956123909, "context": 0.06234109029173851}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 0.9086353182792664, "loss_scale": 4.0, "retries": 0, "seconds": 210.4, "max_memory_gb": 18.13, "ce": 0.076562381349504, "kl": 0.051435247412882745, "context": 0.10228000394999981}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 1.0185848474502563, "loss_scale": 4.0, "retries": 0, "seconds": 210.9, "max_memory_gb": 18.13, "ce": 0.22350455075502396, "kl": 0.1750101128127426, "context": 0.31100960820913315}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.12984110414981842, "loss_scale": 4.0, "retries": 0, "seconds": 211.5, "max_memory_gb": 18.13, "ce": 0.006190678424900398, "kl": 0.007402412127703428, "context": 0.009891884663375095}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 1.3162380456924438, "loss_scale": 4.0, "retries": 0, "seconds": 212.0, "max_memory_gb": 18.13, "ce": 0.03000148420687765, "kl": 0.7014723278116435, "context": 0.3807376637123525}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.34901002049446106, "loss_scale": 4.0, "retries": 0, "seconds": 212.6, "max_memory_gb": 18.13, "ce": 0.08261623739963397, "kl": 0.024243196938186884, "context": 0.09473783546127379}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 1.1203255653381348, "loss_scale": 4.0, "retries": 0, "seconds": 213.1, "max_memory_gb": 18.13, "ce": 0.04996839736122638, "kl": 0.015393998008221388, "context": 0.057665397645905614}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 1.0347672700881958, "loss_scale": 4.0, "retries": 0, "seconds": 213.6, "max_memory_gb": 18.13, "ce": 0.014713569427840412, "kl": 0.29421801488570054, "context": 0.16182258178014308}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.5261873006820679, "loss_scale": 4.0, "retries": 0, "seconds": 214.1, "max_memory_gb": 18.13, "ce": 0.1248817655723542, "kl": 0.08337945432867855, "context": 0.16657149279490113}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.656940758228302, "loss_scale": 4.0, "retries": 0, "seconds": 214.7, "max_memory_gb": 18.13, "ce": 0.099410783033818, "kl": 0.020240091587766074, "context": 0.10953082889318466}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.9473696351051331, "loss_scale": 4.0, "retries": 0, "seconds": 215.2, "max_memory_gb": 18.13, "ce": 0.14178762737719808, "kl": 0.0016942512593232095, "context": 0.14263475380721502}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 0.9985241889953613, "loss_scale": 4.0, "retries": 0, "seconds": 215.7, "max_memory_gb": 18.13, "ce": 0.04125498025678098, "kl": 0.10623133250919636, "context": 0.0943706447724253}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.497058391571045, "loss_scale": 4.0, "retries": 0, "seconds": 216.3, "max_memory_gb": 18.13, "ce": 0.028850043192505836, "kl": 0.5331650975276716, "context": 0.2954325887840241}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.4540048837661743, "loss_scale": 4.0, "retries": 0, "seconds": 216.8, "max_memory_gb": 18.13, "ce": 0.13075652602128685, "kl": 0.06569735403172672, "context": 0.16360520478338003}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.15398067235946655, "loss_scale": 4.0, "retries": 0, "seconds": 217.4, "max_memory_gb": 18.13, "ce": 0.006129248766228557, "kl": 0.010946775815682486, "context": 0.011602636717725545}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 1.4047340154647827, "loss_scale": 4.0, "retries": 0, "seconds": 217.9, "max_memory_gb": 18.13, "ce": 0.19811904814559966, "kl": 0.06455968030786607, "context": 0.23039889126084745}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.4680037796497345, "loss_scale": 4.0, "retries": 0, "seconds": 218.4, "max_memory_gb": 18.13, "ce": 0.03430485978606157, "kl": 0.024494343902915716, "context": 0.0465520303696394}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.7044392824172974, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 18.13, "ce": 0.1500555519014597, "kl": 0.016958996769972146, "context": 0.1585350544191897}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 2.7578020095825195, "loss_scale": 4.0, "retries": 0, "seconds": 219.5, "max_memory_gb": 18.13, "ce": 0.8118859196547419, "kl": 0.4272914974426385, "context": 1.0255316458642483}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 1.5371367931365967, "loss_scale": 4.0, "retries": 0, "seconds": 220.0, "max_memory_gb": 18.13, "ce": 0.23331059468910098, "kl": 0.2552175459713908, "context": 0.3609193686861545}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.843351423740387, "loss_scale": 4.0, "retries": 0, "seconds": 220.6, "max_memory_gb": 18.13, "ce": 0.19542041805107147, "kl": 0.08166999024979305, "context": 0.23625540666398592}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.7110081911087036, "loss_scale": 4.0, "retries": 0, "seconds": 221.1, "max_memory_gb": 18.13, "ce": 0.05768731876742095, "kl": 0.05135850756778382, "context": 0.08336657244944945}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 2.53642201423645, "loss_scale": 4.0, "retries": 0, "seconds": 221.7, "max_memory_gb": 18.13, "ce": 0.30527748726308346, "kl": 0.23973394148924854, "context": 0.42514444305561483}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.320726603269577, "loss_scale": 4.0, "retries": 0, "seconds": 222.2, "max_memory_gb": 18.13, "ce": 0.14223268977366388, "kl": 0.06557608177536167, "context": 0.17502072686329484}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 1.020475149154663, "loss_scale": 4.0, "retries": 0, "seconds": 222.7, "max_memory_gb": 18.13, "ce": 0.059411578346043825, "kl": 0.0442377784056589, "context": 0.08153046574443579}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 1.8277900218963623, "loss_scale": 4.0, "retries": 0, "seconds": 223.3, "max_memory_gb": 18.13, "ce": 0.25459609407698736, "kl": 0.017108091120462632, "context": 0.2631501330761239}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 1.7696988582611084, "loss_scale": 4.0, "retries": 0, "seconds": 223.8, "max_memory_gb": 18.13, "ce": 0.28512165480060503, "kl": 0.09291108162142336, "context": 0.33157719229348004}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.21363551914691925, "loss_scale": 4.0, "retries": 0, "seconds": 224.3, "max_memory_gb": 18.13, "ce": 0.014360152941662818, "kl": 0.0031671897231717594, "context": 0.015943747421260923}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 1.4966506958007812, "loss_scale": 4.0, "retries": 0, "seconds": 224.8, "max_memory_gb": 18.13, "ce": 0.4132779957726598, "kl": 0.09851468808483332, "context": 0.4625353291630745}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.2563181519508362, "loss_scale": 4.0, "retries": 0, "seconds": 225.4, "max_memory_gb": 18.13, "ce": 0.03006453881971538, "kl": 0.05165369913447648, "context": 0.05589138832874596}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.34430840611457825, "loss_scale": 4.0, "retries": 0, "seconds": 225.9, "max_memory_gb": 18.13, "ce": 0.07267211272846907, "kl": 0.060125026844616514, "context": 0.10273462545592338}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 2.3994550704956055, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 18.13, "ce": 0.1365675643319264, "kl": 0.8516666874056682, "context": 0.562400905881077}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.410554975271225, "loss_scale": 4.0, "retries": 0, "seconds": 227.0, "max_memory_gb": 18.13, "ce": 0.039253302034921944, "kl": 0.009523352942778729, "context": 0.044014979677740484}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 0.8189825415611267, "loss_scale": 4.0, "retries": 0, "seconds": 227.5, "max_memory_gb": 18.13, "ce": 0.19934995070798323, "kl": 0.016672798483341467, "context": 0.20768635204876773}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 2.0244481563568115, "loss_scale": 4.0, "retries": 0, "seconds": 228.0, "max_memory_gb": 18.13, "ce": 0.32592791924253106, "kl": 0.3710335048381239, "context": 0.511444691568613}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.8388458490371704, "loss_scale": 4.0, "retries": 0, "seconds": 228.6, "max_memory_gb": 18.13, "ce": 0.05408090149285272, "kl": 0.027256285393377766, "context": 0.06770904664881527}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.9298650026321411, "loss_scale": 4.0, "retries": 0, "seconds": 229.1, "max_memory_gb": 18.13, "ce": 0.05748373130336404, "kl": 0.07431807418470271, "context": 0.09464276931248605}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 1.3816431760787964, "loss_scale": 4.0, "retries": 0, "seconds": 229.7, "max_memory_gb": 18.13, "ce": 0.20260197669267654, "kl": 0.06479931948706508, "context": 0.23500163108110428}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 1.075618863105774, "loss_scale": 4.0, "retries": 0, "seconds": 230.2, "max_memory_gb": 18.13, "ce": 0.08471120051399339, "kl": 0.025711018058700574, "context": 0.09756671327340882}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.6981760263442993, "loss_scale": 4.0, "retries": 0, "seconds": 230.7, "max_memory_gb": 18.13, "ce": 0.007715717074461281, "kl": 0.32735884216526756, "context": 0.17139514058362693}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.6519555449485779, "loss_scale": 4.0, "retries": 0, "seconds": 231.3, "max_memory_gb": 18.13, "ce": 0.012419502600096166, "kl": 0.08656748381690704, "context": 0.05570324556902051}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.8284320831298828, "loss_scale": 4.0, "retries": 0, "seconds": 231.8, "max_memory_gb": 18.13, "ce": 0.11674529872834682, "kl": 0.013214260925451526, "context": 0.12335243279812858}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 3.6782267093658447, "loss_scale": 4.0, "retries": 0, "seconds": 232.4, "max_memory_gb": 18.13, "ce": 0.2021258996683173, "kl": 0.15418787426460767, "context": 0.2792198367533274}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.9138026833534241, "loss_scale": 4.0, "retries": 0, "seconds": 232.9, "max_memory_gb": 18.13, "ce": 0.06873382060439326, "kl": 0.06769371475093067, "context": 0.10258067864924669}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 1.4055671691894531, "loss_scale": 4.0, "retries": 0, "seconds": 233.4, "max_memory_gb": 18.13, "ce": 0.08905361720826477, "kl": 0.12412868734099902, "context": 0.15111795900156721}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.19460220634937286, "loss_scale": 4.0, "retries": 0, "seconds": 233.9, "max_memory_gb": 18.13, "ce": 0.006814036169089377, "kl": 0.00013870367365598213, "context": 0.006883388035930693}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 1.369215965270996, "loss_scale": 4.0, "retries": 0, "seconds": 234.5, "max_memory_gb": 18.13, "ce": 0.12075853138230741, "kl": 0.07615136698768765, "context": 0.15883421117905527}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 1.3328807353973389, "loss_scale": 4.0, "retries": 0, "seconds": 235.0, "max_memory_gb": 18.13, "ce": 0.11167210480198264, "kl": 0.02370445516135078, "context": 0.12352433148771524}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 1.3997846841812134, "loss_scale": 4.0, "retries": 0, "seconds": 235.6, "max_memory_gb": 18.13, "ce": 0.1662965051655192, "kl": 0.07956592579921562, "context": 0.206079471798148}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.3800504207611084, "loss_scale": 4.0, "retries": 0, "seconds": 236.1, "max_memory_gb": 18.13, "ce": 0.11226121906656772, "kl": 0.046208072802983224, "context": 0.13536525587551296}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.326578974723816, "loss_scale": 4.0, "retries": 0, "seconds": 236.6, "max_memory_gb": 18.13, "ce": 0.1578868255019188, "kl": 0.1672554916003719, "context": 0.24151456728577614}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.6909452080726624, "loss_scale": 4.0, "retries": 0, "seconds": 237.2, "max_memory_gb": 18.13, "ce": 0.03755658399313688, "kl": 0.05109686588502882, "context": 0.06310501683037728}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 3.158529758453369, "loss_scale": 4.0, "retries": 0, "seconds": 237.7, "max_memory_gb": 18.13, "ce": 0.2449694313108921, "kl": 0.31239486775302794, "context": 0.40116687264526263}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.521504819393158, "loss_scale": 4.0, "retries": 0, "seconds": 238.2, "max_memory_gb": 18.13, "ce": 0.08117597398813814, "kl": 0.017634755844483152, "context": 0.08999335468979552}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.8120843172073364, "loss_scale": 4.0, "retries": 0, "seconds": 238.8, "max_memory_gb": 18.13, "ce": 0.06421744590625167, "kl": 0.030231701937736943, "context": 0.07933329651132226}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 1.1538386344909668, "loss_scale": 4.0, "retries": 0, "seconds": 239.3, "max_memory_gb": 18.13, "ce": 0.11094819870777428, "kl": 0.04311392200179398, "context": 0.13250516122207046}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 1.7328029870986938, "loss_scale": 4.0, "retries": 0, "seconds": 239.8, "max_memory_gb": 18.13, "ce": 0.38204582009348087, "kl": 0.023890887505331193, "context": 0.3939912747882772}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 1.1470592021942139, "loss_scale": 4.0, "retries": 0, "seconds": 240.4, "max_memory_gb": 18.13, "ce": 0.07956416625529528, "kl": 0.12242523740860634, "context": 0.14077678346075118}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.8538029193878174, "loss_scale": 4.0, "retries": 0, "seconds": 240.9, "max_memory_gb": 18.13, "ce": 0.18205065850634128, "kl": 0.015501209547437611, "context": 0.1898012583842501}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 2.6545662879943848, "loss_scale": 4.0, "retries": 0, "seconds": 241.4, "max_memory_gb": 18.13, "ce": 0.16840671189129353, "kl": 0.08745247311890125, "context": 0.21213294751942158}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.515339195728302, "loss_scale": 4.0, "retries": 0, "seconds": 242.0, "max_memory_gb": 18.13, "ce": 0.018074496008921415, "kl": 0.008603769276305684, "context": 0.022376379725756124}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.635081946849823, "loss_scale": 4.0, "retries": 0, "seconds": 242.5, "max_memory_gb": 18.13, "ce": 0.029016132291872054, "kl": 0.03395968119366444, "context": 0.04599597104243003}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.8307676911354065, "loss_scale": 4.0, "retries": 0, "seconds": 243.0, "max_memory_gb": 18.13, "ce": 0.013869462243746966, "kl": 0.05530695141351316, "context": 0.04152293820516206}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.9375873804092407, "loss_scale": 4.0, "retries": 0, "seconds": 243.6, "max_memory_gb": 18.13, "ce": 0.1339153545559384, "kl": 0.06515558293904178, "context": 0.16649314598180354}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 1.6096993684768677, "loss_scale": 4.0, "retries": 0, "seconds": 244.1, "max_memory_gb": 18.13, "ce": 0.056432632787618786, "kl": 0.04400160620571114, "context": 0.07843343494459987}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 2.150031805038452, "loss_scale": 4.0, "retries": 0, "seconds": 244.6, "max_memory_gb": 18.13, "ce": 0.172699443413876, "kl": 0.008079597388132242, "context": 0.17673924018163234}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 2.2200844287872314, "loss_scale": 4.0, "retries": 0, "seconds": 245.2, "max_memory_gb": 18.13, "ce": 0.15165520494338125, "kl": 0.2642354285926558, "context": 0.2837729206075892}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.3440312445163727, "loss_scale": 4.0, "retries": 0, "seconds": 245.7, "max_memory_gb": 18.13, "ce": 0.01416297908872366, "kl": 0.012322629161644727, "context": 0.02032429410610348}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.3769032657146454, "loss_scale": 4.0, "retries": 0, "seconds": 246.2, "max_memory_gb": 18.13, "ce": 0.01828916007070802, "kl": 0.02605944862443721, "context": 0.03131888445932418}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 2.4767489433288574, "loss_scale": 4.0, "retries": 0, "seconds": 246.8, "max_memory_gb": 18.13, "ce": 0.1600519585190341, "kl": 0.14575854960276047, "context": 0.23293122817995027}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 5.859697341918945, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 18.13, "ce": 0.19009268935769796, "kl": 0.18881903862347826, "context": 0.2845022161491215}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.42730212211608887, "loss_scale": 4.0, "retries": 0, "seconds": 247.8, "max_memory_gb": 18.13, "ce": 0.01851736448588781, "kl": 0.03270148515002802, "context": 0.03486810642061755}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 2.508920192718506, "loss_scale": 4.0, "retries": 0, "seconds": 248.4, "max_memory_gb": 18.13, "ce": 0.21447969874134287, "kl": 0.17092097346903756, "context": 0.29994018271099776}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.8337326645851135, "loss_scale": 4.0, "retries": 0, "seconds": 248.9, "max_memory_gb": 18.13, "ce": 0.037579870942863636, "kl": 0.008599387083449983, "context": 0.0418795653822599}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.6565382480621338, "loss_scale": 4.0, "retries": 0, "seconds": 249.4, "max_memory_gb": 18.13, "ce": 0.023375077144009992, "kl": 0.0010754342201835243, "context": 0.023912794655188918}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.10848314315080643, "loss_scale": 4.0, "retries": 0, "seconds": 250.0, "max_memory_gb": 18.13, "ce": 0.00272446169401519, "kl": 0.0029920493652753066, "context": 0.004220486443955451}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 2.122861862182617, "loss_scale": 4.0, "retries": 0, "seconds": 250.5, "max_memory_gb": 18.13, "ce": 0.01670023093902273, "kl": 0.6796784662437858, "context": 0.35653947009996045}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 1.0136992931365967, "loss_scale": 4.0, "retries": 0, "seconds": 251.1, "max_memory_gb": 18.13, "ce": 0.010640993335982785, "kl": 0.26213223906870553, "context": 0.141707108799892}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.492336630821228, "loss_scale": 4.0, "retries": 0, "seconds": 251.6, "max_memory_gb": 18.13, "ce": 0.038714419526513666, "kl": 0.01691179961198941, "context": 0.04717031854670495}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.445721298456192, "loss_scale": 4.0, "retries": 0, "seconds": 252.1, "max_memory_gb": 18.13, "ce": 0.023217750422190875, "kl": 0.011435611188062467, "context": 0.02893555600894615}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 1.592365026473999, "loss_scale": 4.0, "retries": 0, "seconds": 252.7, "max_memory_gb": 18.13, "ce": 0.12110593417310156, "kl": 0.03786422119446797, "context": 0.14003803912783042}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 2.2318544387817383, "loss_scale": 4.0, "retries": 0, "seconds": 253.2, "max_memory_gb": 18.13, "ce": 0.002919728562119417, "kl": 0.4050485061161453, "context": 0.20544398110359907}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.8030204772949219, "loss_scale": 4.0, "retries": 0, "seconds": 253.7, "max_memory_gb": 18.13, "ce": 0.1554102890950162, "kl": 0.1928567698560073, "context": 0.2518386694137007}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.22450990974903107, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 18.13, "ce": 0.008381774125155061, "kl": 0.008616335791884921, "context": 0.012689941911958158}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.7944514751434326, "loss_scale": 4.0, "retries": 0, "seconds": 254.8, "max_memory_gb": 18.13, "ce": 0.03067887932411395, "kl": 0.012026619908283465, "context": 0.03669218870345503}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.642738938331604, "loss_scale": 4.0, "retries": 0, "seconds": 255.3, "max_memory_gb": 18.13, "ce": 0.08052715737721883, "kl": 0.006198170696734451, "context": 0.08362624284927733}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 1.1831098794937134, "loss_scale": 4.0, "retries": 0, "seconds": 255.8, "max_memory_gb": 18.13, "ce": 0.16630138018808793, "kl": 0.006548654484049621, "context": 0.16957570498925634}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.5319647789001465, "loss_scale": 4.0, "retries": 0, "seconds": 256.4, "max_memory_gb": 18.13, "ce": 0.029673506098333746, "kl": 0.007635977934114635, "context": 0.0334914947161451}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 2.3246777057647705, "loss_scale": 4.0, "retries": 0, "seconds": 256.9, "max_memory_gb": 18.13, "ce": 0.008540694980183616, "kl": 0.5931286932172952, "context": 0.3051050463691354}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.7231789231300354, "loss_scale": 4.0, "retries": 0, "seconds": 257.4, "max_memory_gb": 18.13, "ce": 0.027087367481726687, "kl": 0.040870752447631276, "context": 0.04752274361817399}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 2.24965763092041, "loss_scale": 4.0, "retries": 0, "seconds": 258.0, "max_memory_gb": 18.13, "ce": 0.3942906819283962, "kl": 0.28574467069120146, "context": 0.5371630173176527}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 1.5784399509429932, "loss_scale": 4.0, "retries": 0, "seconds": 258.5, "max_memory_gb": 18.13, "ce": 0.17497409624047577, "kl": 0.02299026062246412, "context": 0.18646922381594777}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.428879976272583, "loss_scale": 4.0, "retries": 0, "seconds": 259.0, "max_memory_gb": 18.13, "ce": 0.021539779525483027, "kl": 0.030485650662740227, "context": 0.03678260574815795}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.8007647395133972, "loss_scale": 4.0, "retries": 0, "seconds": 259.6, "max_memory_gb": 18.13, "ce": 0.10669111413881183, "kl": 0.014404808462131768, "context": 0.11389351973775774}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 1.9371771812438965, "loss_scale": 4.0, "retries": 0, "seconds": 260.1, "max_memory_gb": 18.13, "ce": 0.16202965821139514, "kl": 0.5692800562828779, "context": 0.44666967540979385}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 2.703615188598633, "loss_scale": 4.0, "retries": 0, "seconds": 260.6, "max_memory_gb": 18.13, "ce": 0.46539414674043655, "kl": 0.19809188484214246, "context": 0.5644401055760682}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 1.4569271802902222, "loss_scale": 4.0, "retries": 0, "seconds": 261.2, "max_memory_gb": 18.13, "ce": 0.0871082553640008, "kl": 0.0535448151640594, "context": 0.11388066317886114}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 1.7627936601638794, "loss_scale": 4.0, "retries": 0, "seconds": 261.7, "max_memory_gb": 18.13, "ce": 0.1478412776486948, "kl": 0.030267361435107887, "context": 0.16297495743492618}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 0.9847998023033142, "loss_scale": 4.0, "retries": 0, "seconds": 262.2, "max_memory_gb": 18.13, "ce": 0.051028352056164294, "kl": 0.061282114234927576, "context": 0.08166941272793338}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.8436630964279175, "loss_scale": 4.0, "retries": 0, "seconds": 262.7, "max_memory_gb": 18.13, "ce": 0.1072176315356046, "kl": 0.02048382951761596, "context": 0.11745954700745642}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 2.8735625743865967, "loss_scale": 4.0, "retries": 0, "seconds": 263.3, "max_memory_gb": 18.13, "ce": 0.3180700148222968, "kl": 0.06912517198361456, "context": 0.35263259458588436}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.5227972269058228, "loss_scale": 4.0, "retries": 0, "seconds": 263.8, "max_memory_gb": 18.13, "ce": 0.027707642555469647, "kl": 0.015408545674290508, "context": 0.035411914461292326}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 1.1634923219680786, "loss_scale": 4.0, "retries": 0, "seconds": 264.3, "max_memory_gb": 18.13, "ce": 0.08338304577046074, "kl": 0.17609792701841798, "context": 0.17143201394355856}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.28293102979660034, "loss_scale": 4.0, "retries": 0, "seconds": 264.9, "max_memory_gb": 18.13, "ce": 0.010501827113330364, "kl": 0.002503621013602242, "context": 0.011753637343645096}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 1.1946747303009033, "loss_scale": 4.0, "retries": 0, "seconds": 265.4, "max_memory_gb": 18.13, "ce": 0.08134778239764273, "kl": 0.022499673767015338, "context": 0.09259762009605765}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 2.3924057483673096, "loss_scale": 4.0, "retries": 0, "seconds": 265.9, "max_memory_gb": 18.13, "ce": 0.007671054219827056, "kl": 0.4979574400204001, "context": 0.25664977525593713}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.9154782295227051, "loss_scale": 4.0, "retries": 0, "seconds": 266.5, "max_memory_gb": 18.13, "ce": 0.05512366071343422, "kl": 0.016294496483169496, "context": 0.06327090971171856}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.5892864465713501, "loss_scale": 4.0, "retries": 0, "seconds": 267.0, "max_memory_gb": 18.13, "ce": 0.025186540122376755, "kl": 0.027987919631414115, "context": 0.039180500782094896}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 3.119968891143799, "loss_scale": 4.0, "retries": 0, "seconds": 267.5, "max_memory_gb": 18.13, "ce": 0.2329122223163722, "kl": 0.02836413610202726, "context": 0.2470942905638367}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.5541741847991943, "loss_scale": 4.0, "retries": 0, "seconds": 268.1, "max_memory_gb": 18.13, "ce": 0.116815960558597, "kl": 0.036704663070850074, "context": 0.13516829221043736}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 1.6632716655731201, "loss_scale": 4.0, "retries": 0, "seconds": 268.6, "max_memory_gb": 18.13, "ce": 0.14770698481879663, "kl": 0.04947088276821887, "context": 0.17244242274318822}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 2.543524742126465, "loss_scale": 4.0, "retries": 0, "seconds": 269.1, "max_memory_gb": 18.13, "ce": 0.283497712953249, "kl": 0.012850654165958986, "context": 0.28992304467828944}
diff --git a/runs/single_llava/context/L08_s13/adapter/README.md b/runs/single_llava/context/L08_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L08_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L08_s13/adapter/adapter_config.json b/runs/single_llava/context/L08_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..bae381e925c44dd2569aaf0c8a32c23c46965e7d
--- /dev/null
+++ b/runs/single_llava/context/L08_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.8.mlp.up_proj",
+    "model.language_model.layers.8.self_attn.o_proj",
+    "model.language_model.layers.8.mlp.down_proj",
+    "model.language_model.layers.8.mlp.gate_proj",
+    "model.language_model.layers.8.self_attn.q_proj",
+    "model.language_model.layers.8.self_attn.v_proj",
+    "model.language_model.layers.8.self_attn.k_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L08_s13/config.json b/runs/single_llava/context/L08_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..60f91aca79fa8606d1219dd9b69576b0dd198f84
--- /dev/null
+++ b/runs/single_llava/context/L08_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 8,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L08_s13/steps.jsonl b/runs/single_llava/context/L08_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..053d32184e7132e6b8e4ffc348de7a3e2ea6a38e
--- /dev/null
+++ b/runs/single_llava/context/L08_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.3501115143299103, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 17.25, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.32153457403182983, "loss_scale": 4.0, "retries": 0, "seconds": 2.1, "max_memory_gb": 17.26, "ce": 2.6823360128328204, "kl": 0.010380933992564678, "context": 2.687526488676667}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.742525041103363, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 17.26, "ce": 5.388859391212463, "kl": 0.17870125122135505, "context": 5.478209972381592}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.48634278774261475, "loss_scale": 4.0, "retries": 0, "seconds": 3.2, "max_memory_gb": 17.26, "ce": 3.8216089056804776, "kl": 0.05127976627409225, "context": 3.847248855512589}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.10691273212432861, "loss_scale": 4.0, "retries": 0, "seconds": 3.7, "max_memory_gb": 17.27, "ce": 0.08600145718082786, "kl": 0.03648475254885852, "context": 0.10424383170902729}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.2593233585357666, "loss_scale": 4.0, "retries": 0, "seconds": 4.3, "max_memory_gb": 17.27, "ce": 1.4640915859490633, "kl": 0.18289266794454306, "context": 1.555537961423397}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.5847917795181274, "loss_scale": 4.0, "retries": 0, "seconds": 4.8, "max_memory_gb": 17.27, "ce": 3.431172906421125, "kl": 0.10590594093810068, "context": 3.4841258144006133}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.6165064573287964, "loss_scale": 4.0, "retries": 0, "seconds": 5.3, "max_memory_gb": 17.27, "ce": 3.682084336876869, "kl": 0.15979337505996227, "context": 3.7619809955358505}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.48970186710357666, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 17.27, "ce": 2.8296582251787186, "kl": 0.16236457007471472, "context": 2.910840503871441}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.3534274995326996, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 17.27, "ce": 1.6274888068437576, "kl": 0.05976504017598927, "context": 1.6573713719844818}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.5893758535385132, "loss_scale": 4.0, "retries": 0, "seconds": 6.9, "max_memory_gb": 17.27, "ce": 2.6135123930871487, "kl": 0.009334306843811646, "context": 2.6181795466691256}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.4571392834186554, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 17.27, "ce": 1.523090383503586, "kl": 0.017309109040070325, "context": 1.53174497326836}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.5902856588363647, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 17.27, "ce": 2.65126466518268, "kl": 0.007462664332706481, "context": 2.6549959937110543}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.896729052066803, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 17.28, "ce": 3.8617706559598446, "kl": 0.1175258873263374, "context": 3.9205335080623627}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.6707322001457214, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 17.28, "ce": 2.557223074778449, "kl": 0.2681808924244251, "context": 2.691313466988504}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.46828359365463257, "loss_scale": 4.0, "retries": 0, "seconds": 9.6, "max_memory_gb": 17.28, "ce": 1.3571383537491783, "kl": 0.008815409877570346, "context": 1.3615460422588512}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.4538321793079376, "loss_scale": 4.0, "retries": 0, "seconds": 10.1, "max_memory_gb": 17.28, "ce": 1.5709773665294051, "kl": 0.07769034011289477, "context": 1.6098225330933928}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.9216547608375549, "loss_scale": 4.0, "retries": 0, "seconds": 10.6, "max_memory_gb": 17.28, "ce": 3.5253367200493813, "kl": 0.0950031375978142, "context": 3.572838183492422}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.462794303894043, "loss_scale": 4.0, "retries": 0, "seconds": 11.1, "max_memory_gb": 17.28, "ce": 5.262769818305969, "kl": 0.3288692579662893, "context": 5.427204489707947}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.026488585397601128, "loss_scale": 4.0, "retries": 0, "seconds": 11.7, "max_memory_gb": 17.28, "ce": 0.012771624024026096, "kl": 0.002086976121063344, "context": 0.013815111829899251}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.8733869194984436, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 17.28, "ce": 2.724796053953469, "kl": 0.02949150148924673, "context": 2.739541701041162}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.9190129637718201, "loss_scale": 4.0, "retries": 0, "seconds": 12.8, "max_memory_gb": 17.28, "ce": 2.88486354239285, "kl": 0.0525436804164201, "context": 2.911135444417596}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.8866432309150696, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 17.28, "ce": 2.6017609296832234, "kl": 0.5165351094765356, "context": 2.860028521157801}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.6769890785217285, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 17.28, "ce": 4.899322867393494, "kl": 0.013786440249532461, "context": 4.906216025352478}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.6441535949707031, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 17.28, "ce": 4.123236283659935, "kl": 0.03708008350804448, "context": 4.141776263713837}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.2485289573669434, "loss_scale": 4.0, "retries": 0, "seconds": 14.9, "max_memory_gb": 17.28, "ce": 5.045658588409424, "kl": 0.06825647898949683, "context": 5.079786717891693}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.783875823020935, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 17.28, "ce": 4.04110711812973, "kl": 0.088502511032857, "context": 4.085358411073685}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.239180326461792, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 17.28, "ce": 2.699384890496731, "kl": 0.22524749487638474, "context": 2.8120086044073105}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.1133934259414673, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 17.28, "ce": 2.5269465323071927, "kl": 0.5205841325514484, "context": 2.7872385892551392}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.6120398044586182, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 17.28, "ce": 2.5987060526385903, "kl": 0.0383079043822363, "context": 2.6178600676357746}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.5996850728988647, "loss_scale": 4.0, "retries": 0, "seconds": 17.6, "max_memory_gb": 17.28, "ce": 2.6939403861761093, "kl": 0.02717684127856046, "context": 2.7075287476181984}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.8298659324645996, "loss_scale": 4.0, "retries": 0, "seconds": 18.1, "max_memory_gb": 17.28, "ce": 5.050431847572327, "kl": 0.01573188661132008, "context": 5.058297753334045}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.4439830780029297, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 17.28, "ce": 2.3346222192049026, "kl": 0.2012975076213479, "context": 2.435271006077528}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.7408798933029175, "loss_scale": 4.0, "retries": 0, "seconds": 19.2, "max_memory_gb": 17.28, "ce": 2.460272994823754, "kl": 0.6181081989780068, "context": 2.7693270538002253}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 2.0163490772247314, "loss_scale": 4.0, "retries": 0, "seconds": 19.7, "max_memory_gb": 17.28, "ce": 2.499628375750035, "kl": 0.03060317796189338, "context": 2.5149300396442413}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.4943325519561768, "loss_scale": 4.0, "retries": 0, "seconds": 20.2, "max_memory_gb": 17.28, "ce": 2.3935726024210453, "kl": 0.007020051241852343, "context": 2.397082634270191}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.7675186395645142, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 17.28, "ce": 2.2781947664916515, "kl": 0.3476068885065615, "context": 2.451998256146908}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.369934558868408, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 17.28, "ce": 4.305767178535461, "kl": 0.4850118836038746, "context": 4.548273146152496}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.5870540142059326, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 17.28, "ce": 1.2479636678472161, "kl": 0.030298168770968914, "context": 1.2631128067150712}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.331643581390381, "loss_scale": 4.0, "retries": 0, "seconds": 22.3, "max_memory_gb": 17.28, "ce": 2.443806506693363, "kl": 0.11484158900566399, "context": 2.5012272000312805}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.0150007009506226, "loss_scale": 4.0, "retries": 0, "seconds": 22.8, "max_memory_gb": 17.28, "ce": 1.348920002579689, "kl": 0.17397643742151558, "context": 1.4359082356095314}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.542598247528076, "loss_scale": 4.0, "retries": 0, "seconds": 23.4, "max_memory_gb": 17.28, "ce": 2.023146112449467, "kl": 0.006931582494871691, "context": 2.026611866429448}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.871532440185547, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 17.28, "ce": 2.0709287328645587, "kl": 0.009805085122934543, "context": 2.075831243302673}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.4619114398956299, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 17.28, "ce": 1.2305533112958074, "kl": 0.25883717654505745, "context": 1.359971899073571}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.620240330696106, "loss_scale": 4.0, "retries": 0, "seconds": 25.0, "max_memory_gb": 17.28, "ce": 1.2043628543615341, "kl": 0.056055446388199925, "context": 1.2323905527591705}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.44065523147583, "loss_scale": 4.0, "retries": 0, "seconds": 25.5, "max_memory_gb": 17.28, "ce": 1.9692677706480026, "kl": 0.07853203287231736, "context": 2.008533827960491}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.850043296813965, "loss_scale": 4.0, "retries": 0, "seconds": 26.0, "max_memory_gb": 17.28, "ce": 1.6451494712382555, "kl": 0.03427122611901723, "context": 1.662285112310201}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 6.002554893493652, "loss_scale": 4.0, "retries": 0, "seconds": 26.6, "max_memory_gb": 17.28, "ce": 3.0180594846606255, "kl": 0.014649336852016859, "context": 3.0253841876983643}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 6.114142894744873, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 17.28, "ce": 2.8261447362601757, "kl": 0.08852424815995619, "context": 2.8704068921506405}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 4.934144973754883, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 17.28, "ce": 2.9140635896474123, "kl": 0.052675935468869284, "context": 2.9404015727341175}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 2.127584457397461, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 17.28, "ce": 1.1005317270755768, "kl": 0.1177013578126207, "context": 1.159382394514978}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 4.538797378540039, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 17.28, "ce": 1.6957248412072659, "kl": 0.11321533244336024, "context": 1.7523325607180595}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 7.7559614181518555, "loss_scale": 4.0, "retries": 0, "seconds": 29.2, "max_memory_gb": 17.28, "ce": 2.8584571927785873, "kl": 0.19532693293876946, "context": 2.956120677292347}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 9.53759765625, "loss_scale": 4.0, "retries": 0, "seconds": 29.7, "max_memory_gb": 17.28, "ce": 2.6963201612234116, "kl": 0.042425124207511544, "context": 2.717532694339752}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 6.090407371520996, "loss_scale": 4.0, "retries": 0, "seconds": 30.2, "max_memory_gb": 17.28, "ce": 1.6308602462522686, "kl": 0.023445551662007347, "context": 1.6425830493681133}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 13.728044509887695, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 17.28, "ce": 3.0182334780693054, "kl": 0.3529200409539044, "context": 3.1946935057640076}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 4.044471263885498, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 17.28, "ce": 0.7948597697541118, "kl": 0.1057327234884724, "context": 0.8477261317893863}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 7.449426174163818, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 17.28, "ce": 1.297148211626336, "kl": 0.005397579705459066, "context": 1.299846997950226}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.6698005199432373, "loss_scale": 4.0, "retries": 0, "seconds": 32.4, "max_memory_gb": 17.28, "ce": 1.14050729945302, "kl": 0.07976912311278284, "context": 1.180391889065504}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 7.09375524520874, "loss_scale": 4.0, "retries": 0, "seconds": 32.9, "max_memory_gb": 17.28, "ce": 1.350893273949623, "kl": 0.013296823963173665, "context": 1.357541685923934}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 11.803340911865234, "loss_scale": 4.0, "retries": 0, "seconds": 33.4, "max_memory_gb": 17.28, "ce": 1.4161477908492088, "kl": 0.08789781329687685, "context": 1.460096687078476}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 4.080582618713379, "loss_scale": 4.0, "retries": 0, "seconds": 34.0, "max_memory_gb": 17.28, "ce": 0.6655149720609188, "kl": 0.039893337758257985, "context": 0.6854616729542613}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 10.005477905273438, "loss_scale": 4.0, "retries": 0, "seconds": 34.5, "max_memory_gb": 17.28, "ce": 1.726084841415286, "kl": 0.074944774270989, "context": 1.76355723105371}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 10.824371337890625, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 17.31, "ce": 1.5528994835913181, "kl": 0.04280357854440808, "context": 1.5743013098835945}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 9.748082160949707, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 17.31, "ce": 1.6376375555992126, "kl": 0.11077502346597612, "context": 1.6930250823497772}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 2.8076956272125244, "loss_scale": 4.0, "retries": 0, "seconds": 36.1, "max_memory_gb": 17.31, "ce": 0.5713371178135276, "kl": 0.05987227882724255, "context": 0.6012732670642436}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 6.742191791534424, "loss_scale": 4.0, "retries": 0, "seconds": 36.6, "max_memory_gb": 17.31, "ce": 1.0073913782835007, "kl": 0.031834261608310044, "context": 1.0233084931969643}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 4.349191665649414, "loss_scale": 4.0, "retries": 0, "seconds": 37.2, "max_memory_gb": 17.31, "ce": 0.47386644780635834, "kl": 0.08049791841767728, "context": 0.5141154043376446}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 5.915610313415527, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 17.31, "ce": 0.6568926870822906, "kl": 0.02452388615347445, "context": 0.6691546281799674}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 4.944156169891357, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 17.31, "ce": 0.5657218396663666, "kl": 0.006182321929372847, "context": 0.5688130054622889}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 7.8542070388793945, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 17.31, "ce": 0.7388896597549319, "kl": 0.026828508591279387, "context": 0.7523039057850838}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 8.23542594909668, "loss_scale": 4.0, "retries": 0, "seconds": 39.3, "max_memory_gb": 17.31, "ce": 0.811671644449234, "kl": 0.11462378618307412, "context": 0.8689835369586945}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.251171350479126, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 17.31, "ce": 0.2813606094568968, "kl": 0.038638157304376364, "context": 0.30067968368530273}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.509859561920166, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 17.31, "ce": 0.47612260840833187, "kl": 0.025165360304526985, "context": 0.488705288618803}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.9280591011047363, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 17.31, "ce": 0.8244384601712227, "kl": 0.21889693848788738, "context": 0.9338869154453278}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 1.79644775390625, "loss_scale": 4.0, "retries": 0, "seconds": 41.5, "max_memory_gb": 17.31, "ce": 0.3678727261722088, "kl": 0.25349682196974754, "context": 0.4946211352944374}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 3.434199094772339, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 17.31, "ce": 0.6998389735817909, "kl": 0.04077795147895813, "context": 0.7202279642224312}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 3.1540727615356445, "loss_scale": 4.0, "retries": 0, "seconds": 42.5, "max_memory_gb": 17.31, "ce": 0.3900018446147442, "kl": 0.054773012176156044, "context": 0.41738835349678993}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.9898221492767334, "loss_scale": 4.0, "retries": 0, "seconds": 43.0, "max_memory_gb": 17.31, "ce": 0.28150592371821404, "kl": 0.08375801099464297, "context": 0.3233849201351404}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 2.097543954849243, "loss_scale": 4.0, "retries": 0, "seconds": 43.5, "max_memory_gb": 17.31, "ce": 0.26746197789907455, "kl": 0.01932937395758927, "context": 0.277126656845212}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 2.853348970413208, "loss_scale": 4.0, "retries": 0, "seconds": 44.0, "max_memory_gb": 17.31, "ce": 0.4967803806066513, "kl": 0.09204847272485495, "context": 0.5428046137094498}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 2.140270948410034, "loss_scale": 4.0, "retries": 0, "seconds": 44.6, "max_memory_gb": 17.31, "ce": 0.36223412305116653, "kl": 0.06423027161508799, "context": 0.39434925466775894}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 2.3591504096984863, "loss_scale": 4.0, "retries": 0, "seconds": 45.1, "max_memory_gb": 17.31, "ce": 0.3818027526140213, "kl": 0.15497652627527714, "context": 0.4592910185456276}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.3590075969696045, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 17.31, "ce": 0.253342071082443, "kl": 0.05413555935956538, "context": 0.28040984785184264}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.1619586944580078, "loss_scale": 4.0, "retries": 0, "seconds": 46.2, "max_memory_gb": 17.31, "ce": 0.10065779136493802, "kl": 0.05137639775057323, "context": 0.12634598789736629}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 1.7661291360855103, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 17.31, "ce": 0.30522673577070236, "kl": 0.19366014376282692, "context": 0.4020567983388901}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 2.2914116382598877, "loss_scale": 4.0, "retries": 0, "seconds": 47.3, "max_memory_gb": 17.31, "ce": 0.318920586258173, "kl": 0.05774483038112521, "context": 0.34779300168156624}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.442716598510742, "loss_scale": 4.0, "retries": 0, "seconds": 47.8, "max_memory_gb": 17.31, "ce": 0.3915816955268383, "kl": 0.02657187134900596, "context": 0.4048676243983209}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.6973505020141602, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 17.31, "ce": 0.13634020742028952, "kl": 0.010221322358120233, "context": 0.14145086985081434}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 1.4331114292144775, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 17.31, "ce": 0.2702927617356181, "kl": 0.03005680162459612, "context": 0.28532116673886776}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.8917362093925476, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 17.31, "ce": 0.3487848900258541, "kl": 0.09514591936022043, "context": 0.39635784178972244}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 0.6744269132614136, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 17.31, "ce": 0.15382895153015852, "kl": 0.08982566092163324, "context": 0.19874178292229772}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 1.6936548948287964, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 17.31, "ce": 0.23274757340550423, "kl": 0.0724674909724854, "context": 0.2689813217148185}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 1.3218252658843994, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 17.31, "ce": 0.21258662268519402, "kl": 0.12593467580154538, "context": 0.27555395662784576}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 1.2699075937271118, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 17.31, "ce": 0.3129979409277439, "kl": 0.06061949860304594, "context": 0.3433077000081539}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 1.5093657970428467, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 17.31, "ce": 0.26707007363438606, "kl": 0.03325556928757578, "context": 0.28369786217808723}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 0.7330729365348816, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 17.31, "ce": 0.21979250013828278, "kl": 0.032665110658854246, "context": 0.23612505570054054}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 0.8087990880012512, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 17.31, "ce": 0.10192470764741302, "kl": 0.03891857046983205, "context": 0.12138399574905634}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.8008635640144348, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 17.31, "ce": 0.22671360755339265, "kl": 0.08230331511003897, "context": 0.2678652557078749}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 0.958234429359436, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 17.31, "ce": 0.21847852878272533, "kl": 0.09856998559553176, "context": 0.267763527110219}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 1.0460113286972046, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 17.31, "ce": 0.14395414665341377, "kl": 0.3220844694878906, "context": 0.3049963749945164}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 1.5724685192108154, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 17.31, "ce": 0.2859016936272383, "kl": 0.1505555547773838, "context": 0.3611794635653496}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 0.4496290683746338, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 17.31, "ce": 0.08163913991302252, "kl": 0.04943536513019353, "context": 0.10635682148858905}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 0.44039347767829895, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 17.31, "ce": 0.09392358211334795, "kl": 0.03491676214616746, "context": 0.11138196266256273}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 0.632160484790802, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 17.31, "ce": 0.1332938652485609, "kl": 0.2110132813686505, "context": 0.23880050145089626}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 0.5654012560844421, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 17.31, "ce": 0.12509449757635593, "kl": 0.05745322071015835, "context": 0.15382110513746738}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.39299651980400085, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 17.31, "ce": 0.14887710381299257, "kl": 0.015123219971428625, "context": 0.15643871761858463}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.4832007884979248, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 17.31, "ce": 0.13799157040193677, "kl": 0.016566563252126798, "context": 0.14627485163509846}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 0.8732224702835083, "loss_scale": 4.0, "retries": 0, "seconds": 58.9, "max_memory_gb": 17.31, "ce": 0.20987021178007126, "kl": 0.3987513071624562, "context": 0.40924585796892643}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 0.6900612115859985, "loss_scale": 4.0, "retries": 0, "seconds": 59.4, "max_memory_gb": 17.31, "ce": 0.18657705932855606, "kl": 0.05096447782125324, "context": 0.2120592948049307}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 0.5274605751037598, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 17.31, "ce": 0.07576625468209386, "kl": 0.08258657835540362, "context": 0.11705954559147358}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 0.42056727409362793, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 17.31, "ce": 0.11015826277434826, "kl": 0.013008369016461074, "context": 0.11666244640946388}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 0.45002034306526184, "loss_scale": 4.0, "retries": 0, "seconds": 61.0, "max_memory_gb": 17.31, "ce": 0.13176196534186602, "kl": 0.05414530448615551, "context": 0.15883461572229862}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 1.1519849300384521, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 17.31, "ce": 0.3058849722146988, "kl": 0.7359866488841362, "context": 0.6738782785832882}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.5690091252326965, "loss_scale": 4.0, "retries": 0, "seconds": 62.1, "max_memory_gb": 17.31, "ce": 0.2073957109823823, "kl": 0.250431246124208, "context": 0.3326113307848573}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.2629123628139496, "loss_scale": 4.0, "retries": 0, "seconds": 62.6, "max_memory_gb": 17.31, "ce": 0.11856873845681548, "kl": 0.005918201801250689, "context": 0.12152783828787506}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 0.8886095881462097, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 17.31, "ce": 0.2361123412847519, "kl": 0.013206938718212768, "context": 0.2427158112404868}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.23512877523899078, "loss_scale": 4.0, "retries": 0, "seconds": 63.7, "max_memory_gb": 17.31, "ce": 0.14335224265232682, "kl": 0.03070148672850337, "context": 0.15870298724621534}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.3496552109718323, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 17.31, "ce": 0.09033380448818207, "kl": 0.07614231191109866, "context": 0.12840496376156807}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 0.8325417041778564, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 17.31, "ce": 0.21072012279182673, "kl": 0.07497465470805764, "context": 0.2482074461877346}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 0.8792006373405457, "loss_scale": 4.0, "retries": 0, "seconds": 65.3, "max_memory_gb": 17.31, "ce": 0.27943603647872806, "kl": 0.10039752822194714, "context": 0.32963479519821703}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 0.3851802349090576, "loss_scale": 4.0, "retries": 0, "seconds": 65.8, "max_memory_gb": 17.31, "ce": 0.13024078495800495, "kl": 0.021120331410202198, "context": 0.14080095174722373}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.3178349733352661, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 17.31, "ce": 0.4051503138616681, "kl": 0.05565796376322396, "context": 0.4329793034121394}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 0.6956835985183716, "loss_scale": 4.0, "retries": 0, "seconds": 66.9, "max_memory_gb": 17.31, "ce": 0.26724923215806484, "kl": 0.04402976940036751, "context": 0.2892641145735979}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.2864537835121155, "loss_scale": 4.0, "retries": 0, "seconds": 67.4, "max_memory_gb": 17.31, "ce": 0.13819484878331423, "kl": 0.01674119505332783, "context": 0.1465654424391687}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 0.6942158937454224, "loss_scale": 4.0, "retries": 0, "seconds": 67.9, "max_memory_gb": 17.31, "ce": 0.3263224884867668, "kl": 0.09149184986017644, "context": 0.3720684126019478}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.22669188678264618, "loss_scale": 4.0, "retries": 0, "seconds": 68.4, "max_memory_gb": 17.31, "ce": 0.04747417266480625, "kl": 0.004130926987272687, "context": 0.04953963658772409}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 0.8630371689796448, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 17.31, "ce": 0.11310996487736702, "kl": 0.4245481215475593, "context": 0.32538402872160077}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 0.3134177029132843, "loss_scale": 4.0, "retries": 0, "seconds": 69.5, "max_memory_gb": 17.31, "ce": 0.058648133650422096, "kl": 0.07241490262094885, "context": 0.09485558699816465}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 1.086403250694275, "loss_scale": 4.0, "retries": 0, "seconds": 70.0, "max_memory_gb": 17.31, "ce": 0.2815661393105984, "kl": 0.458545340137789, "context": 0.5108388243243098}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 0.743634819984436, "loss_scale": 4.0, "retries": 0, "seconds": 70.6, "max_memory_gb": 17.31, "ce": 0.2699838913977146, "kl": 0.0489353584125638, "context": 0.2944515645503998}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.2843722105026245, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 17.31, "ce": 0.30846239998936653, "kl": 0.2686764882237185, "context": 0.4428006294183433}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 0.7770033478736877, "loss_scale": 4.0, "retries": 0, "seconds": 71.6, "max_memory_gb": 17.31, "ce": 0.19307538494467735, "kl": 0.17152075795456767, "context": 0.27883576042950153}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.8236839175224304, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 17.31, "ce": 0.20719997119158506, "kl": 0.02354418343747966, "context": 0.21897206641733646}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 0.36101800203323364, "loss_scale": 4.0, "retries": 0, "seconds": 72.7, "max_memory_gb": 17.31, "ce": 0.11350650526583195, "kl": 0.023040972970193252, "context": 0.12502699438482523}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.6318720579147339, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 17.31, "ce": 0.17281199898570776, "kl": 0.052070675417780876, "context": 0.1988473366945982}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.3626278340816498, "loss_scale": 4.0, "retries": 0, "seconds": 73.7, "max_memory_gb": 17.31, "ce": 0.13101507257670164, "kl": 0.01665068173315376, "context": 0.13934041559696198}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.4209684133529663, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 17.31, "ce": 0.0944138653576374, "kl": 0.07279989495873451, "context": 0.1308138146996498}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 0.984993040561676, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 17.31, "ce": 0.1666637286543846, "kl": 0.17428221739828587, "context": 0.2538048289716244}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.39856547117233276, "loss_scale": 4.0, "retries": 0, "seconds": 75.4, "max_memory_gb": 17.31, "ce": 0.06432047206908464, "kl": 0.016046553413616493, "context": 0.07234374666586518}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.3644014000892639, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 17.31, "ce": 0.14672667160630226, "kl": 0.014352761209011078, "context": 0.15390304941684008}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.3355778753757477, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 17.31, "ce": 0.13627874432131648, "kl": 0.015949952823575586, "context": 0.14425372052937746}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.3008485436439514, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 17.31, "ce": 0.12085978931281716, "kl": 0.061976206921826815, "context": 0.15184789279010147}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.8460353016853333, "loss_scale": 4.0, "retries": 0, "seconds": 77.5, "max_memory_gb": 17.31, "ce": 0.16911103576421738, "kl": 0.14192323887255043, "context": 0.24007265642285347}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.1998051404953003, "loss_scale": 4.0, "retries": 0, "seconds": 78.1, "max_memory_gb": 17.31, "ce": 0.030992957297712564, "kl": 0.00389461888698861, "context": 0.03294026548974216}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.24294373393058777, "loss_scale": 4.0, "retries": 0, "seconds": 78.6, "max_memory_gb": 17.31, "ce": 0.11579207051545382, "kl": 0.00997239351272583, "context": 0.120778264477849}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.6967992186546326, "loss_scale": 4.0, "retries": 0, "seconds": 79.1, "max_memory_gb": 17.31, "ce": 0.19307366793509573, "kl": 0.48962880531325936, "context": 0.43788806442171335}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.3979666233062744, "loss_scale": 4.0, "retries": 0, "seconds": 79.7, "max_memory_gb": 17.31, "ce": 0.021870573866181076, "kl": 0.06165591016178951, "context": 0.05269852816127241}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.304395467042923, "loss_scale": 4.0, "retries": 0, "seconds": 80.2, "max_memory_gb": 17.31, "ce": 0.09718300169333816, "kl": 0.011231955530092819, "context": 0.10279897670261562}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.5129801034927368, "loss_scale": 4.0, "retries": 0, "seconds": 80.7, "max_memory_gb": 17.31, "ce": 0.17795653734356165, "kl": 0.021926115674432367, "context": 0.18891958985477686}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.4483717083930969, "loss_scale": 4.0, "retries": 0, "seconds": 81.3, "max_memory_gb": 17.31, "ce": 0.11959654185920954, "kl": 0.016764269283157773, "context": 0.12797867669723928}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 0.8290218114852905, "loss_scale": 4.0, "retries": 0, "seconds": 81.8, "max_memory_gb": 17.31, "ce": 0.06445967918261886, "kl": 0.5822445483063348, "context": 0.3555819382891059}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.5013607740402222, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 17.31, "ce": 0.06385140353813767, "kl": 0.29060838534496725, "context": 0.20915559027343988}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.3466306924819946, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 17.31, "ce": 0.42117075342684984, "kl": 0.31456920778146014, "context": 0.5784553568810225}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 1.3425027132034302, "loss_scale": 4.0, "retries": 0, "seconds": 83.4, "max_memory_gb": 17.31, "ce": 0.45161721925251186, "kl": 0.1727532872610027, "context": 0.5379938834812492}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.4861384332180023, "loss_scale": 4.0, "retries": 0, "seconds": 83.9, "max_memory_gb": 17.31, "ce": 0.14470019889995456, "kl": 0.1520438327715965, "context": 0.2207221173448488}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.32267117500305176, "loss_scale": 4.0, "retries": 0, "seconds": 84.4, "max_memory_gb": 17.31, "ce": 0.1346647720783949, "kl": 0.04461959363834467, "context": 0.15697456942871213}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.6762065887451172, "loss_scale": 4.0, "retries": 0, "seconds": 85.0, "max_memory_gb": 17.31, "ce": 0.2546574678272009, "kl": 0.32543730176985264, "context": 0.4173761121928692}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.2911030352115631, "loss_scale": 4.0, "retries": 0, "seconds": 85.5, "max_memory_gb": 17.31, "ce": 0.05614792089909315, "kl": 0.08177929336670786, "context": 0.09703756729140878}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.5670896768569946, "loss_scale": 4.0, "retries": 0, "seconds": 86.0, "max_memory_gb": 17.31, "ce": 0.10553374234586954, "kl": 0.09814910846762359, "context": 0.15460830088704824}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.4945656955242157, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 17.31, "ce": 0.2214462235569954, "kl": 0.03743432534975, "context": 0.24016338121145964}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.3623470067977905, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 17.31, "ce": 0.09997089486569166, "kl": 0.02791523074847646, "context": 0.1139285103417933}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.6275845766067505, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 17.31, "ce": 0.07295440696179867, "kl": 0.07871932900161482, "context": 0.11231407127343118}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.6638659238815308, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 17.31, "ce": 0.2513243518769741, "kl": 0.023624629946425557, "context": 0.2631366699934006}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.4532228410243988, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 17.31, "ce": 0.12222063960507512, "kl": 0.038819276245703804, "context": 0.141630282625556}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.5494240522384644, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 17.31, "ce": 0.13587534986436367, "kl": 0.040056281723082066, "context": 0.155903491191566}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.5891830325126648, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 17.31, "ce": 0.17402229504659772, "kl": 0.013954785390524194, "context": 0.1809996860101819}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.38591548800468445, "loss_scale": 4.0, "retries": 0, "seconds": 90.2, "max_memory_gb": 17.31, "ce": 0.07316302414983511, "kl": 0.093643824220635, "context": 0.11998493573628366}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.466064691543579, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 17.31, "ce": 0.24621396651491523, "kl": 1.1543025372666307, "context": 0.8233652561903}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.15395094454288483, "loss_scale": 4.0, "retries": 0, "seconds": 91.3, "max_memory_gb": 17.31, "ce": 0.018247545696794987, "kl": 0.003218179917894304, "context": 0.01985663583036512}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 1.3091715574264526, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 17.31, "ce": 0.0494005725486204, "kl": 0.25599079101812094, "context": 0.17739597707986832}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.33759036660194397, "loss_scale": 4.0, "retries": 0, "seconds": 92.4, "max_memory_gb": 17.31, "ce": 0.05595280439592898, "kl": 0.006169497122755274, "context": 0.05903755407780409}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 1.0809568166732788, "loss_scale": 4.0, "retries": 0, "seconds": 92.9, "max_memory_gb": 17.31, "ce": 0.16120622865855694, "kl": 0.07105220993980765, "context": 0.19673233479261398}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.38136711716651917, "loss_scale": 4.0, "retries": 0, "seconds": 93.4, "max_memory_gb": 17.31, "ce": 0.07728991657495499, "kl": 0.07416634389664978, "context": 0.11437309253960848}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.7583032846450806, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 17.31, "ce": 0.1468974594026804, "kl": 0.19108009757474065, "context": 0.24243750795722008}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.45920294523239136, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 17.31, "ce": 0.03621493326500058, "kl": 0.00989415540243499, "context": 0.04116201167926192}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.7462453246116638, "loss_scale": 4.0, "retries": 0, "seconds": 95.0, "max_memory_gb": 17.31, "ce": 0.16766311298124492, "kl": 0.15712523373076692, "context": 0.24622573237866163}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.2660755217075348, "loss_scale": 4.0, "retries": 0, "seconds": 95.6, "max_memory_gb": 17.31, "ce": 0.1365776197053492, "kl": 0.07873135988484137, "context": 0.17594330152496696}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.32525527477264404, "loss_scale": 4.0, "retries": 0, "seconds": 96.1, "max_memory_gb": 17.31, "ce": 0.3114231340587139, "kl": 0.004092548333574086, "context": 0.3134694101754576}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 0.9531481266021729, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 17.31, "ce": 0.2605496351607144, "kl": 0.08044393491582014, "context": 0.30077159870415926}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.42385852336883545, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 17.31, "ce": 0.0599279017187655, "kl": 0.05851178721059114, "context": 0.08918379270471632}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.1782619953155518, "loss_scale": 4.0, "retries": 0, "seconds": 97.6, "max_memory_gb": 17.31, "ce": 0.22500404436141253, "kl": 0.006420498961233534, "context": 0.2282142920885235}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.09127902239561081, "loss_scale": 4.0, "retries": 0, "seconds": 98.2, "max_memory_gb": 17.31, "ce": 0.011530480813235044, "kl": 0.0045153245628171135, "context": 0.013788143056444824}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.6272978186607361, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 17.31, "ce": 0.04698747012298554, "kl": 0.08340933213185053, "context": 0.08869213273283094}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.20818136632442474, "loss_scale": 4.0, "retries": 0, "seconds": 99.2, "max_memory_gb": 17.31, "ce": 0.03276892425492406, "kl": 0.014722215317306109, "context": 0.04013003152795136}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.29831603169441223, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 17.31, "ce": 0.05292245070450008, "kl": 0.004649852751754224, "context": 0.05524737713858485}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.7118575572967529, "loss_scale": 4.0, "retries": 0, "seconds": 100.3, "max_memory_gb": 17.31, "ce": 0.05877503287047148, "kl": 0.3987406614032807, "context": 0.25814536726102233}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.24600012600421906, "loss_scale": 4.0, "retries": 0, "seconds": 100.8, "max_memory_gb": 17.31, "ce": 0.12996303453110158, "kl": 0.04426770548161585, "context": 0.15209688863251358}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.34218576550483704, "loss_scale": 4.0, "retries": 0, "seconds": 101.4, "max_memory_gb": 17.31, "ce": 0.04064429411664605, "kl": 0.009711910475743935, "context": 0.045500250067561865}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 0.9171738028526306, "loss_scale": 4.0, "retries": 0, "seconds": 101.9, "max_memory_gb": 17.31, "ce": 0.01810719375498593, "kl": 0.47089195635635406, "context": 0.2535531776957214}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.6218774914741516, "loss_scale": 4.0, "retries": 0, "seconds": 102.4, "max_memory_gb": 17.31, "ce": 0.26954474369995296, "kl": 0.05215549026615918, "context": 0.2956224884837866}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.1965685486793518, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 17.31, "ce": 0.17718808609060943, "kl": 0.006569789205968846, "context": 0.18047297629527748}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.3801434338092804, "loss_scale": 4.0, "retries": 0, "seconds": 103.5, "max_memory_gb": 17.31, "ce": 0.03608480258844793, "kl": 0.045974723441759124, "context": 0.05907216388732195}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 1.2206788063049316, "loss_scale": 4.0, "retries": 0, "seconds": 104.0, "max_memory_gb": 17.31, "ce": 0.35850896034389734, "kl": 0.015015122713521123, "context": 0.36601652298122644}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.3429056406021118, "loss_scale": 4.0, "retries": 0, "seconds": 104.5, "max_memory_gb": 17.31, "ce": 0.11713524255901575, "kl": 0.01880000624805689, "context": 0.12653524661436677}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.20307783782482147, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 17.31, "ce": 0.02109605073928833, "kl": 0.020278318668715656, "context": 0.031235211063176394}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.3486374318599701, "loss_scale": 4.0, "retries": 0, "seconds": 105.6, "max_memory_gb": 17.31, "ce": 0.10087721073068678, "kl": 0.02514367114054039, "context": 0.11344904592260718}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.3518693149089813, "loss_scale": 4.0, "retries": 0, "seconds": 106.1, "max_memory_gb": 17.31, "ce": 0.15137714939191937, "kl": 0.038468061087769456, "context": 0.17061118548735976}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.5663413405418396, "loss_scale": 4.0, "retries": 0, "seconds": 106.7, "max_memory_gb": 17.31, "ce": 0.06130932131782174, "kl": 0.2830252815037966, "context": 0.20282195694744587}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.3986760079860687, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 17.31, "ce": 0.21747575653716922, "kl": 0.011964146004174836, "context": 0.22345782443881035}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.375204473733902, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 17.31, "ce": 0.08313428168185055, "kl": 0.016762442042818293, "context": 0.09151550522074103}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.05085955560207367, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 17.31, "ce": 0.005918028065934777, "kl": 0.003546084859408438, "context": 0.0076910704374313354}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.48549944162368774, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 17.31, "ce": 0.14997660741209984, "kl": 0.09319222439080477, "context": 0.19657272286713123}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.7101966142654419, "loss_scale": 4.0, "retries": 0, "seconds": 109.3, "max_memory_gb": 17.31, "ce": 0.10473509202711284, "kl": 0.02824576525017619, "context": 0.1188579760491848}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.6124159693717957, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 17.31, "ce": 0.05368633707985282, "kl": 0.42077706981945084, "context": 0.26407486968673766}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.5774468183517456, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 17.31, "ce": 0.0982862095697783, "kl": 0.09528099445742555, "context": 0.14592671470018104}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.3483184576034546, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 17.31, "ce": 0.5894387243315578, "kl": 0.4458990932616871, "context": 0.8123882894869894}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.15722393989562988, "loss_scale": 4.0, "retries": 0, "seconds": 111.5, "max_memory_gb": 17.31, "ce": 0.014036372303962708, "kl": 0.006083735264837742, "context": 0.017078240169212222}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.328183650970459, "loss_scale": 4.0, "retries": 0, "seconds": 112.0, "max_memory_gb": 17.31, "ce": 0.19830993935465813, "kl": 0.018065654847305268, "context": 0.20734276669099927}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.4709593653678894, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 17.31, "ce": 0.21748003433458507, "kl": 0.06291195648373105, "context": 0.24893600377254188}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.44390004873275757, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 17.31, "ce": 0.18741470121312886, "kl": 0.09930891799740493, "context": 0.23706915974617004}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.922776460647583, "loss_scale": 4.0, "retries": 0, "seconds": 113.6, "max_memory_gb": 17.31, "ce": 0.16214507445693016, "kl": 0.7130837154109031, "context": 0.5186869474127889}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.657627522945404, "loss_scale": 4.0, "retries": 0, "seconds": 114.2, "max_memory_gb": 17.31, "ce": 0.3126186514273286, "kl": 0.05181240833189804, "context": 0.3385248628910631}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 0.6676064133644104, "loss_scale": 4.0, "retries": 0, "seconds": 114.7, "max_memory_gb": 17.31, "ce": 0.24424990569241345, "kl": 0.3160908173304051, "context": 0.4022953067906201}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 1.1713052988052368, "loss_scale": 4.0, "retries": 0, "seconds": 115.2, "max_memory_gb": 17.31, "ce": 0.38992783334106207, "kl": 0.003980013854743447, "context": 0.3919178490759805}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 0.8775510787963867, "loss_scale": 4.0, "retries": 0, "seconds": 115.8, "max_memory_gb": 17.31, "ce": 0.18338247225619853, "kl": 0.09767247649142519, "context": 0.23221870372071862}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.3707541227340698, "loss_scale": 4.0, "retries": 0, "seconds": 116.3, "max_memory_gb": 17.31, "ce": 0.0539249338908121, "kl": 0.03339826303817972, "context": 0.07062406511977315}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.7479272484779358, "loss_scale": 4.0, "retries": 0, "seconds": 116.8, "max_memory_gb": 17.31, "ce": 0.18433392979204655, "kl": 0.4463620139285922, "context": 0.4075149283744395}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.45830655097961426, "loss_scale": 4.0, "retries": 0, "seconds": 117.4, "max_memory_gb": 17.31, "ce": 0.12164728995412588, "kl": 0.039231973874848336, "context": 0.14126327587291598}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.5249963998794556, "loss_scale": 4.0, "retries": 0, "seconds": 117.9, "max_memory_gb": 17.31, "ce": 0.07484377757646143, "kl": 0.02391863492084667, "context": 0.08680309518240392}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.6606734395027161, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 17.31, "ce": 0.08465573191642761, "kl": 0.12366760650184005, "context": 0.14648953638970852}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.0656707286834717, "loss_scale": 4.0, "retries": 0, "seconds": 118.9, "max_memory_gb": 17.31, "ce": 0.24141343077644706, "kl": 0.3807823080132948, "context": 0.43180457316339016}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.5284172296524048, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 17.31, "ce": 0.1973175562452525, "kl": 0.0367796040372923, "context": 0.21570735785644501}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.6260899305343628, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 17.31, "ce": 0.07516905013471842, "kl": 0.1542320899461629, "context": 0.1522850999608636}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.567842423915863, "loss_scale": 4.0, "retries": 0, "seconds": 120.5, "max_memory_gb": 17.31, "ce": 0.057706828229129314, "kl": 0.060113005340099335, "context": 0.08776332810521126}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.4762272834777832, "loss_scale": 4.0, "retries": 0, "seconds": 121.1, "max_memory_gb": 17.31, "ce": 0.08611893048509955, "kl": 0.019921063212677836, "context": 0.09607945894822478}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.11030173301696777, "loss_scale": 4.0, "retries": 0, "seconds": 121.6, "max_memory_gb": 17.31, "ce": 0.013056036899797618, "kl": 0.005483839777298272, "context": 0.015797956846654415}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.6101220846176147, "loss_scale": 4.0, "retries": 0, "seconds": 122.1, "max_memory_gb": 17.31, "ce": 0.06245389441028237, "kl": 0.1128436652943492, "context": 0.11887572705745697}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.882317841053009, "loss_scale": 4.0, "retries": 0, "seconds": 122.7, "max_memory_gb": 17.31, "ce": 0.25516056222841144, "kl": 0.15603859373368323, "context": 0.333179862704128}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.6177744269371033, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 17.31, "ce": 0.24405605223728344, "kl": 0.11256400548154488, "context": 0.3003380538430065}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.5275479555130005, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 17.31, "ce": 0.24186997511424124, "kl": 0.025365587003761902, "context": 0.2545527598122135}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.16105249524116516, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 17.31, "ce": 0.021581514738500118, "kl": 0.007469087722711265, "context": 0.02531605865806341}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 0.843372106552124, "loss_scale": 4.0, "retries": 0, "seconds": 124.8, "max_memory_gb": 17.31, "ce": 0.11875067092478275, "kl": 0.14435137226064398, "context": 0.19092635437846184}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.5075666308403015, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 17.31, "ce": 0.15065224608406425, "kl": 0.021270427765557542, "context": 0.1612874581478536}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.8198975920677185, "loss_scale": 4.0, "retries": 0, "seconds": 125.8, "max_memory_gb": 17.31, "ce": 0.3211551171261817, "kl": 0.02800367900636047, "context": 0.33515697182156146}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.4141565263271332, "loss_scale": 4.0, "retries": 0, "seconds": 126.4, "max_memory_gb": 17.31, "ce": 0.07638939353637397, "kl": 0.025914561410900205, "context": 0.08934667194262147}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 0.7192825078964233, "loss_scale": 4.0, "retries": 0, "seconds": 126.9, "max_memory_gb": 17.31, "ce": 0.17906224727630615, "kl": 0.0682364865801901, "context": 0.21318049193359911}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.618060827255249, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 17.31, "ce": 0.04089265677612275, "kl": 0.1336710979376221, "context": 0.10772820515558124}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.5367211699485779, "loss_scale": 4.0, "retries": 0, "seconds": 127.9, "max_memory_gb": 17.31, "ce": 0.16780223650857806, "kl": 0.02963124809321016, "context": 0.18261786550283432}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.6299831867218018, "loss_scale": 4.0, "retries": 0, "seconds": 128.4, "max_memory_gb": 17.31, "ce": 0.09506811411119998, "kl": 0.03376985786599107, "context": 0.11195304291322827}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.42268580198287964, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 17.31, "ce": 0.1666904529556632, "kl": 0.019919565820600837, "context": 0.1766502344980836}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.5328677296638489, "loss_scale": 4.0, "retries": 0, "seconds": 129.5, "max_memory_gb": 17.31, "ce": 0.1275370311923325, "kl": 0.05627860099775717, "context": 0.1556763283442706}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.8258840441703796, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 17.31, "ce": 0.018053512321785092, "kl": 0.14838893536943942, "context": 0.09224798274226487}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 1.0244351625442505, "loss_scale": 4.0, "retries": 0, "seconds": 130.5, "max_memory_gb": 17.31, "ce": 0.30208298831712455, "kl": 0.05544973534415476, "context": 0.3298078615916893}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.5395340919494629, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 17.31, "ce": 0.38440681947395205, "kl": 0.007441687630489469, "context": 0.3881276596803218}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.22423775494098663, "loss_scale": 4.0, "retries": 0, "seconds": 131.6, "max_memory_gb": 17.31, "ce": 0.022835711715742946, "kl": 0.02619977399081108, "context": 0.03593559854198247}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.5007067322731018, "loss_scale": 4.0, "retries": 0, "seconds": 132.1, "max_memory_gb": 17.31, "ce": 0.09211918106302619, "kl": 0.037441618216689676, "context": 0.11083999020047486}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.45221230387687683, "loss_scale": 4.0, "retries": 0, "seconds": 132.7, "max_memory_gb": 17.31, "ce": 0.06732294615358114, "kl": 0.03467731596902013, "context": 0.08466160576790571}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.5006363987922668, "loss_scale": 4.0, "retries": 0, "seconds": 133.2, "max_memory_gb": 17.31, "ce": 0.07456724671646953, "kl": 0.02622744973632507, "context": 0.08768097078427672}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.2997570037841797, "loss_scale": 4.0, "retries": 0, "seconds": 133.7, "max_memory_gb": 17.31, "ce": 0.03835258644539863, "kl": 0.011995003012998495, "context": 0.04435008857399225}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.40491774678230286, "loss_scale": 4.0, "retries": 0, "seconds": 134.3, "max_memory_gb": 17.31, "ce": 0.12532877456396818, "kl": 0.020942531526088715, "context": 0.1358000417239964}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.6332260966300964, "loss_scale": 4.0, "retries": 0, "seconds": 134.8, "max_memory_gb": 17.31, "ce": 0.037064093397930264, "kl": 0.14809921360574663, "context": 0.11111369775608182}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.3823714256286621, "loss_scale": 4.0, "retries": 0, "seconds": 135.3, "max_memory_gb": 17.31, "ce": 0.11794009548611939, "kl": 0.0071849640535219805, "context": 0.12153257604222745}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.847892701625824, "loss_scale": 4.0, "retries": 0, "seconds": 135.8, "max_memory_gb": 17.31, "ce": 0.15166640002280474, "kl": 0.2813682407140732, "context": 0.29235051944851875}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.6694350242614746, "loss_scale": 4.0, "retries": 0, "seconds": 136.4, "max_memory_gb": 17.31, "ce": 0.13113216403871775, "kl": 0.235215884167701, "context": 0.24874010309576988}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.3181670904159546, "loss_scale": 4.0, "retries": 0, "seconds": 136.9, "max_memory_gb": 17.31, "ce": 0.08449819823727012, "kl": 0.0324389099914697, "context": 0.10071765375323594}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.5867497324943542, "loss_scale": 4.0, "retries": 0, "seconds": 137.5, "max_memory_gb": 17.31, "ce": 0.09075122629292309, "kl": 0.02052477309916867, "context": 0.10101361409761012}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.452492356300354, "loss_scale": 4.0, "retries": 0, "seconds": 138.0, "max_memory_gb": 17.31, "ce": 0.04642325022723526, "kl": 0.11168937443289906, "context": 0.10226793820038438}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.20169097185134888, "loss_scale": 4.0, "retries": 0, "seconds": 138.5, "max_memory_gb": 17.31, "ce": 0.014169360278174281, "kl": 0.028313888316915836, "context": 0.028326304513029754}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.4788990616798401, "loss_scale": 4.0, "retries": 0, "seconds": 139.1, "max_memory_gb": 17.31, "ce": 0.07341025024652481, "kl": 0.12413838703650981, "context": 0.13547944370657206}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.6165534853935242, "loss_scale": 4.0, "retries": 0, "seconds": 139.6, "max_memory_gb": 17.31, "ce": 0.1428935038857162, "kl": 0.11093528661876917, "context": 0.19836115092039108}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.7802339196205139, "loss_scale": 4.0, "retries": 0, "seconds": 140.1, "max_memory_gb": 17.31, "ce": 0.17518621170893312, "kl": 0.11392252138466574, "context": 0.2321474647615105}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.6375539898872375, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 17.31, "ce": 0.1697307270951569, "kl": 0.04990725906100124, "context": 0.19468435738235712}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 1.0125186443328857, "loss_scale": 4.0, "retries": 0, "seconds": 141.2, "max_memory_gb": 17.31, "ce": 0.13214778783731163, "kl": 0.06592170752992388, "context": 0.16510864719748497}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.3274368345737457, "loss_scale": 4.0, "retries": 0, "seconds": 141.7, "max_memory_gb": 17.31, "ce": 0.07892873603850603, "kl": 0.07533793937182054, "context": 0.11659770854748785}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 1.3613032102584839, "loss_scale": 4.0, "retries": 0, "seconds": 142.2, "max_memory_gb": 17.31, "ce": 0.3889942669775337, "kl": 0.14323639750364237, "context": 0.46061246073804796}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.33364376425743103, "loss_scale": 4.0, "retries": 0, "seconds": 142.8, "max_memory_gb": 17.31, "ce": 0.14789196499623358, "kl": 0.006053333345334977, "context": 0.15091863181442022}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.3716396689414978, "loss_scale": 4.0, "retries": 0, "seconds": 143.3, "max_memory_gb": 17.31, "ce": 0.23343651508912444, "kl": 0.06515407186816446, "context": 0.2660135496407747}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 1.9943920373916626, "loss_scale": 4.0, "retries": 0, "seconds": 143.8, "max_memory_gb": 17.31, "ce": 0.3867578636854887, "kl": 0.6560513937401993, "context": 0.7147835791110992}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.7710847854614258, "loss_scale": 4.0, "retries": 0, "seconds": 144.4, "max_memory_gb": 17.31, "ce": 0.17011095909401774, "kl": 0.0017734821394697065, "context": 0.1709976934362203}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.20838657021522522, "loss_scale": 4.0, "retries": 0, "seconds": 144.9, "max_memory_gb": 17.31, "ce": 0.031308980425819755, "kl": 0.04691607545828447, "context": 0.054767016088590026}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 0.7565199732780457, "loss_scale": 4.0, "retries": 0, "seconds": 145.4, "max_memory_gb": 17.31, "ce": 0.024674114771187305, "kl": 0.43869109253864735, "context": 0.24401965702418238}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.6293576955795288, "loss_scale": 4.0, "retries": 0, "seconds": 145.9, "max_memory_gb": 17.31, "ce": 0.016727432142943144, "kl": 0.17706907718093134, "context": 0.10526197380386293}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.46045511960983276, "loss_scale": 4.0, "retries": 0, "seconds": 146.5, "max_memory_gb": 17.31, "ce": 0.12923770304769278, "kl": 0.052725394372828305, "context": 0.1556003987789154}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 1.1539233922958374, "loss_scale": 4.0, "retries": 0, "seconds": 147.0, "max_memory_gb": 17.31, "ce": 0.5343513274565339, "kl": 0.31911649531684816, "context": 0.6939095898997039}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.1286749541759491, "loss_scale": 4.0, "retries": 0, "seconds": 147.5, "max_memory_gb": 17.31, "ce": 0.010971640585921705, "kl": 0.009293175760831218, "context": 0.01561822893563658}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.34550756216049194, "loss_scale": 4.0, "retries": 0, "seconds": 148.1, "max_memory_gb": 17.31, "ce": 0.014107782393693924, "kl": 0.11852743168128654, "context": 0.07337149861268699}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 0.9776535630226135, "loss_scale": 4.0, "retries": 0, "seconds": 148.6, "max_memory_gb": 17.31, "ce": 0.17453364375978708, "kl": 0.22217989006094285, "context": 0.2856235961080529}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 0.5385403037071228, "loss_scale": 4.0, "retries": 0, "seconds": 149.1, "max_memory_gb": 17.31, "ce": 0.07468892890028656, "kl": 0.11218024487607181, "context": 0.13077905355021358}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.19901779294013977, "loss_scale": 4.0, "retries": 0, "seconds": 149.6, "max_memory_gb": 17.31, "ce": 0.019150144886225462, "kl": 0.009970518061891198, "context": 0.024135403102263808}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 1.8271805047988892, "loss_scale": 4.0, "retries": 0, "seconds": 150.2, "max_memory_gb": 17.31, "ce": 0.24311683583073318, "kl": 0.08604259765706956, "context": 0.28613812720868737}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.35863032937049866, "loss_scale": 4.0, "retries": 0, "seconds": 150.7, "max_memory_gb": 17.31, "ce": 0.08723056863527745, "kl": 0.009878498367470456, "context": 0.09216981590725482}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.42374128103256226, "loss_scale": 4.0, "retries": 0, "seconds": 151.2, "max_memory_gb": 17.31, "ce": 0.06101144291460514, "kl": 0.1868270380946342, "context": 0.1544249540893361}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 0.5913512110710144, "loss_scale": 4.0, "retries": 0, "seconds": 151.7, "max_memory_gb": 17.31, "ce": 0.06612045969814062, "kl": 0.016934401677644928, "context": 0.07458766212221235}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.3828725516796112, "loss_scale": 4.0, "retries": 0, "seconds": 152.3, "max_memory_gb": 17.31, "ce": 0.11961746797896922, "kl": 0.022749822805053554, "context": 0.130992382299155}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.16863825917243958, "loss_scale": 4.0, "retries": 0, "seconds": 152.8, "max_memory_gb": 17.31, "ce": 0.014778184238821268, "kl": 0.01649733352678595, "context": 0.023026850656606257}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.6869879961013794, "loss_scale": 4.0, "retries": 0, "seconds": 153.3, "max_memory_gb": 17.31, "ce": 0.08932938682846725, "kl": 0.03029491938650608, "context": 0.10447684512473643}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.14383520185947418, "loss_scale": 4.0, "retries": 0, "seconds": 153.9, "max_memory_gb": 17.31, "ce": 0.012077326537109911, "kl": 0.008250252387369983, "context": 0.016202453058212996}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.2572596073150635, "loss_scale": 4.0, "retries": 0, "seconds": 154.4, "max_memory_gb": 17.31, "ce": 0.04273350862786174, "kl": 0.07683929073391482, "context": 0.08115315437316895}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.5614465475082397, "loss_scale": 4.0, "retries": 0, "seconds": 154.9, "max_memory_gb": 17.31, "ce": 0.12193329783622175, "kl": 0.05864386460689275, "context": 0.15125523385358974}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.056858886033296585, "loss_scale": 4.0, "retries": 0, "seconds": 155.4, "max_memory_gb": 17.31, "ce": 0.0059915337478742, "kl": 0.002003407800657442, "context": 0.006993237533606589}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.1532334089279175, "loss_scale": 4.0, "retries": 0, "seconds": 156.0, "max_memory_gb": 17.31, "ce": 0.05812555935699493, "kl": 0.6628116825595498, "context": 0.38953140564262867}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.47628116607666016, "loss_scale": 4.0, "retries": 0, "seconds": 156.5, "max_memory_gb": 17.31, "ce": 0.053936930606141686, "kl": 0.11952950671548024, "context": 0.11370168277062476}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 0.750129759311676, "loss_scale": 4.0, "retries": 0, "seconds": 157.0, "max_memory_gb": 17.31, "ce": 0.06697545479983091, "kl": 0.024762449553236365, "context": 0.07935668062418699}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 1.3898667097091675, "loss_scale": 4.0, "retries": 0, "seconds": 157.5, "max_memory_gb": 17.31, "ce": 0.15290867490693927, "kl": 0.10153269267175347, "context": 0.20367502141743898}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 1.3503867387771606, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 17.31, "ce": 0.25745649554301053, "kl": 0.46031570970080793, "context": 0.48761435225605965}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.4729754328727722, "loss_scale": 4.0, "retries": 0, "seconds": 158.6, "max_memory_gb": 17.31, "ce": 0.10990878276061267, "kl": 0.016292190266540274, "context": 0.11805487924721092}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.0301610231399536, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 17.31, "ce": 0.45432438468560576, "kl": 0.16346258379053324, "context": 0.5360556868836284}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.3233763873577118, "loss_scale": 4.0, "retries": 0, "seconds": 159.7, "max_memory_gb": 17.31, "ce": 0.011442585149779916, "kl": 0.11059198339353316, "context": 0.06673857558052987}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.6489782333374023, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 17.31, "ce": 0.27605493797454983, "kl": 0.05048443449777551, "context": 0.3012971505522728}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 0.969120442867279, "loss_scale": 4.0, "retries": 0, "seconds": 160.8, "max_memory_gb": 17.31, "ce": 0.1344486337620765, "kl": 0.11871072872600053, "context": 0.1938039984088391}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.19615715742111206, "loss_scale": 4.0, "retries": 0, "seconds": 161.4, "max_memory_gb": 17.31, "ce": 0.026664382312446833, "kl": 0.007196474791271612, "context": 0.030262620421126485}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.3446427583694458, "loss_scale": 4.0, "retries": 0, "seconds": 162.0, "max_memory_gb": 17.31, "ce": 0.03626010974403471, "kl": 0.0230695512946113, "context": 0.04779488581698388}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.17057576775550842, "loss_scale": 4.0, "retries": 0, "seconds": 162.5, "max_memory_gb": 17.31, "ce": 0.019145326921716332, "kl": 0.005763108492828906, "context": 0.02202688087709248}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 2.6770076751708984, "loss_scale": 4.0, "retries": 0, "seconds": 163.1, "max_memory_gb": 17.31, "ce": 0.1486177092883736, "kl": 0.18814886853215285, "context": 0.24269214342348278}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 1.0186740159988403, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 17.31, "ce": 0.012118458282202482, "kl": 0.17870810336899012, "context": 0.10147251025773585}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.9613459706306458, "loss_scale": 4.0, "retries": 0, "seconds": 164.1, "max_memory_gb": 17.31, "ce": 0.1406439235433936, "kl": 0.010318574641132727, "context": 0.14580321521498263}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 0.8412837982177734, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 17.31, "ce": 0.022905138495843858, "kl": 0.6486361839051824, "context": 0.34722324274480343}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 1.0632719993591309, "loss_scale": 4.0, "retries": 0, "seconds": 165.2, "max_memory_gb": 17.31, "ce": 0.22213329729856923, "kl": 0.01871130190556869, "context": 0.2314889432163909}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.4465198516845703, "loss_scale": 4.0, "retries": 0, "seconds": 165.8, "max_memory_gb": 17.31, "ce": 0.034656121977604926, "kl": 0.22296902595553547, "context": 0.14614063780754805}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.5951316952705383, "loss_scale": 4.0, "retries": 0, "seconds": 166.3, "max_memory_gb": 17.31, "ce": 0.13218186818994582, "kl": 0.018357816410571104, "context": 0.14136076963040978}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.2316228598356247, "loss_scale": 4.0, "retries": 0, "seconds": 166.9, "max_memory_gb": 17.31, "ce": 0.027131252456456423, "kl": 0.012268788821529597, "context": 0.03326564689632505}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.970579206943512, "loss_scale": 4.0, "retries": 0, "seconds": 167.4, "max_memory_gb": 17.31, "ce": 0.1567849619896151, "kl": 0.04650635570578743, "context": 0.1800381404464133}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.5809342265129089, "loss_scale": 4.0, "retries": 0, "seconds": 168.0, "max_memory_gb": 17.31, "ce": 0.14531419309787452, "kl": 0.08222768869018182, "context": 0.18642804492264986}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.3024374544620514, "loss_scale": 4.0, "retries": 0, "seconds": 168.5, "max_memory_gb": 17.31, "ce": 0.04578668670728803, "kl": 0.06103349756449461, "context": 0.07630343502387404}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 0.7998405694961548, "loss_scale": 4.0, "retries": 0, "seconds": 169.1, "max_memory_gb": 17.31, "ce": 0.189713190542534, "kl": 0.6417160076089203, "context": 0.5105711771175265}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.7272747159004211, "loss_scale": 4.0, "retries": 0, "seconds": 169.6, "max_memory_gb": 17.31, "ce": 0.15472026355564594, "kl": 0.017809340497478843, "context": 0.16362493112683296}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.5154730677604675, "loss_scale": 4.0, "retries": 0, "seconds": 170.1, "max_memory_gb": 17.31, "ce": 0.04247464565560222, "kl": 0.4049219664011616, "context": 0.24493562383577228}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 0.6868150234222412, "loss_scale": 4.0, "retries": 0, "seconds": 170.6, "max_memory_gb": 17.31, "ce": 0.07623403146862984, "kl": 0.12333642420708202, "context": 0.13790224120020866}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 0.7218220829963684, "loss_scale": 4.0, "retries": 0, "seconds": 171.2, "max_memory_gb": 17.31, "ce": 0.13708646362647414, "kl": 0.08686215382113005, "context": 0.18051753856707364}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.5296749472618103, "loss_scale": 4.0, "retries": 0, "seconds": 171.7, "max_memory_gb": 17.31, "ce": 0.07991252653300762, "kl": 0.07209059881279245, "context": 0.11595782684162259}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.305107444524765, "loss_scale": 4.0, "retries": 0, "seconds": 172.2, "max_memory_gb": 17.31, "ce": 0.02242232277058065, "kl": 0.05381116493663285, "context": 0.049327906453981996}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.5987204909324646, "loss_scale": 4.0, "retries": 0, "seconds": 172.7, "max_memory_gb": 17.31, "ce": 0.013471356593072414, "kl": 0.08615290041780099, "context": 0.05654780799522996}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.17620107531547546, "loss_scale": 4.0, "retries": 0, "seconds": 173.2, "max_memory_gb": 17.31, "ce": 0.01918498962186277, "kl": 0.009792604243557435, "context": 0.02408129209652543}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.6565091013908386, "loss_scale": 4.0, "retries": 0, "seconds": 173.8, "max_memory_gb": 17.31, "ce": 0.22707894339691848, "kl": 0.03196439133898821, "context": 0.24306114722276106}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.4534909427165985, "loss_scale": 4.0, "retries": 0, "seconds": 174.3, "max_memory_gb": 17.31, "ce": 0.057688349625095725, "kl": 0.0506753867957741, "context": 0.08302604116033763}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 1.7434200048446655, "loss_scale": 4.0, "retries": 0, "seconds": 174.8, "max_memory_gb": 17.31, "ce": 0.18811318383086473, "kl": 0.07479666956351139, "context": 0.22551151108928025}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.6617727279663086, "loss_scale": 4.0, "retries": 0, "seconds": 175.4, "max_memory_gb": 17.31, "ce": 0.10488617536611855, "kl": 0.1409297766367672, "context": 0.17535106535069644}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.6078178882598877, "loss_scale": 4.0, "retries": 0, "seconds": 175.9, "max_memory_gb": 17.31, "ce": 0.11202233377844095, "kl": 0.1626968601703993, "context": 0.1933707562275231}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.47036030888557434, "loss_scale": 4.0, "retries": 0, "seconds": 176.4, "max_memory_gb": 17.31, "ce": 0.17326733627123758, "kl": 0.009807440888835117, "context": 0.17817105544963852}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.9947178363800049, "loss_scale": 4.0, "retries": 0, "seconds": 177.0, "max_memory_gb": 17.31, "ce": 0.2903082650154829, "kl": 0.07980104954913259, "context": 0.3302088021300733}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 2.20650577545166, "loss_scale": 4.0, "retries": 0, "seconds": 177.5, "max_memory_gb": 17.31, "ce": 0.17828290746547282, "kl": 0.05226428194146138, "context": 0.2044150393921882}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 0.8585221171379089, "loss_scale": 4.0, "retries": 0, "seconds": 178.0, "max_memory_gb": 17.31, "ce": 0.30567532405257225, "kl": 0.02465221320744604, "context": 0.31800142186693847}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 0.8691380023956299, "loss_scale": 4.0, "retries": 0, "seconds": 178.6, "max_memory_gb": 17.31, "ce": 0.16468183579854667, "kl": 0.12730631022714078, "context": 0.22833499102853239}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 2.058283567428589, "loss_scale": 4.0, "retries": 0, "seconds": 179.1, "max_memory_gb": 17.31, "ce": 0.37382712005637586, "kl": 0.23680751328356564, "context": 0.4922308827517554}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.19898997247219086, "loss_scale": 4.0, "retries": 0, "seconds": 179.6, "max_memory_gb": 17.31, "ce": 0.013119601819198579, "kl": 0.04407488899596501, "context": 0.03515704581514001}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 0.7601826190948486, "loss_scale": 4.0, "retries": 0, "seconds": 180.1, "max_memory_gb": 17.31, "ce": 0.19655379187315702, "kl": 0.10439855046570301, "context": 0.24875306338071823}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.0929088592529297, "loss_scale": 4.0, "retries": 0, "seconds": 180.7, "max_memory_gb": 17.31, "ce": 0.19798841211013496, "kl": 0.16051979607436806, "context": 0.27824830915778875}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.6427023410797119, "loss_scale": 4.0, "retries": 0, "seconds": 181.2, "max_memory_gb": 17.31, "ce": 0.15861530310939997, "kl": 0.011633504604105838, "context": 0.16443205869290978}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.8650548458099365, "loss_scale": 4.0, "retries": 0, "seconds": 181.8, "max_memory_gb": 17.31, "ce": 0.22126956400461495, "kl": 0.08920181692519691, "context": 0.26587047916837037}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 0.7466883659362793, "loss_scale": 4.0, "retries": 0, "seconds": 182.3, "max_memory_gb": 17.31, "ce": 0.13090414681937546, "kl": 0.021060938015580177, "context": 0.14143461314961314}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 1.1174136400222778, "loss_scale": 4.0, "retries": 0, "seconds": 182.8, "max_memory_gb": 17.31, "ce": 0.25453125429339707, "kl": 0.8455726842803415, "context": 0.6773175783455372}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.2352028489112854, "loss_scale": 4.0, "retries": 0, "seconds": 183.4, "max_memory_gb": 17.31, "ce": 0.01893940707668662, "kl": 0.04250049724942073, "context": 0.040189655497670174}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.20841483771800995, "loss_scale": 4.0, "retries": 0, "seconds": 183.9, "max_memory_gb": 17.31, "ce": 0.02476526191458106, "kl": 0.02060083142714575, "context": 0.03506567794829607}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.3995298147201538, "loss_scale": 4.0, "retries": 0, "seconds": 184.4, "max_memory_gb": 17.31, "ce": 0.1430471065978054, "kl": 0.17443050115252845, "context": 0.23026235535508022}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.8138923645019531, "loss_scale": 4.0, "retries": 0, "seconds": 185.0, "max_memory_gb": 17.31, "ce": 0.19626987853553146, "kl": 0.006035679311025888, "context": 0.19928771746344864}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 0.9774106740951538, "loss_scale": 4.0, "retries": 0, "seconds": 185.5, "max_memory_gb": 17.31, "ce": 0.2091837404295802, "kl": 0.02855662035290152, "context": 0.2234620489180088}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.08937428146600723, "loss_scale": 4.0, "retries": 0, "seconds": 186.0, "max_memory_gb": 17.31, "ce": 0.011527213966473937, "kl": 0.004715632003353676, "context": 0.013885029882658273}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.15202675759792328, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 17.31, "ce": 0.10102241247659549, "kl": 0.023657641097088344, "context": 0.11285123211564496}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.622860312461853, "loss_scale": 4.0, "retries": 0, "seconds": 187.1, "max_memory_gb": 17.31, "ce": 0.1549117360264063, "kl": 0.051959617412649095, "context": 0.18089154176414013}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.0363757610321045, "loss_scale": 4.0, "retries": 0, "seconds": 187.6, "max_memory_gb": 17.31, "ce": 0.042517161346040666, "kl": 0.3005982589966152, "context": 0.19281628262251616}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.7399649620056152, "loss_scale": 4.0, "retries": 0, "seconds": 188.1, "max_memory_gb": 17.31, "ce": 0.1939158730674535, "kl": 0.4189689732156694, "context": 0.40340035408735275}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.4651981294155121, "loss_scale": 4.0, "retries": 0, "seconds": 188.6, "max_memory_gb": 17.31, "ce": 0.13632916449569166, "kl": 0.052040774549823254, "context": 0.16234954982064664}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.46138066053390503, "loss_scale": 4.0, "retries": 0, "seconds": 189.2, "max_memory_gb": 17.31, "ce": 0.18944887700490654, "kl": 0.022462890807219082, "context": 0.20068031980190426}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.39515450596809387, "loss_scale": 4.0, "retries": 0, "seconds": 189.7, "max_memory_gb": 17.31, "ce": 0.13800951797747985, "kl": 0.01979657306219451, "context": 0.14790780207840726}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.4122653603553772, "loss_scale": 4.0, "retries": 0, "seconds": 190.2, "max_memory_gb": 17.31, "ce": 0.044397549354471266, "kl": 0.029139262478565797, "context": 0.05896718136500567}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.5623785257339478, "loss_scale": 4.0, "retries": 0, "seconds": 190.8, "max_memory_gb": 17.31, "ce": 0.07402316620573401, "kl": 0.024236060853581876, "context": 0.08614119770936668}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 0.7885231971740723, "loss_scale": 4.0, "retries": 0, "seconds": 191.3, "max_memory_gb": 17.31, "ce": 0.2221981710754335, "kl": 0.011236282996833324, "context": 0.22781631723046303}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.2287978082895279, "loss_scale": 4.0, "retries": 0, "seconds": 191.9, "max_memory_gb": 17.31, "ce": 0.026520179817453027, "kl": 0.022936317982384935, "context": 0.037988337920978665}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.4684109687805176, "loss_scale": 4.0, "retries": 0, "seconds": 192.4, "max_memory_gb": 17.31, "ce": 0.06872370466589928, "kl": 0.012781321012880653, "context": 0.07511436403729022}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.1746326684951782, "loss_scale": 4.0, "retries": 0, "seconds": 192.9, "max_memory_gb": 17.31, "ce": 0.07142644538544118, "kl": 0.46823710318130907, "context": 0.3055449960520491}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 1.2379860877990723, "loss_scale": 4.0, "retries": 0, "seconds": 193.4, "max_memory_gb": 17.31, "ce": 0.29529596643988043, "kl": 0.05981731664360268, "context": 0.32520462095271796}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.4232144057750702, "loss_scale": 4.0, "retries": 0, "seconds": 194.0, "max_memory_gb": 17.31, "ce": 0.05989320622757077, "kl": 0.10396154876798391, "context": 0.11187397874891758}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.5331265926361084, "loss_scale": 4.0, "retries": 0, "seconds": 194.5, "max_memory_gb": 17.31, "ce": 0.01728833490051329, "kl": 0.5432630903451354, "context": 0.288919879385503}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.7004033923149109, "loss_scale": 4.0, "retries": 0, "seconds": 195.0, "max_memory_gb": 17.31, "ce": 0.15132873016409576, "kl": 0.023525097909441683, "context": 0.16309128422290087}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.627312183380127, "loss_scale": 4.0, "retries": 0, "seconds": 195.6, "max_memory_gb": 17.31, "ce": 0.058211989933624864, "kl": 0.2607579408213496, "context": 0.18859096337109804}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.7110931873321533, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 17.31, "ce": 0.13270172418560833, "kl": 0.10514601533213863, "context": 0.1852747320663184}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.44081416726112366, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 17.31, "ce": 0.02284326625522226, "kl": 0.05447525970521383, "context": 0.050080897635780275}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.22230611741542816, "loss_scale": 4.0, "retries": 0, "seconds": 197.1, "max_memory_gb": 17.31, "ce": 0.042610535747371614, "kl": 0.01390131763764657, "context": 0.04956119472626597}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 1.68952214717865, "loss_scale": 4.0, "retries": 0, "seconds": 197.7, "max_memory_gb": 17.31, "ce": 0.3453829772770405, "kl": 0.15102461236529052, "context": 0.4208952821791172}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.037580132484436, "loss_scale": 4.0, "retries": 0, "seconds": 198.2, "max_memory_gb": 17.31, "ce": 0.18150640395469964, "kl": 0.025442997066420503, "context": 0.19422790431417525}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.3677651286125183, "loss_scale": 4.0, "retries": 0, "seconds": 198.7, "max_memory_gb": 17.31, "ce": 0.04293885221704841, "kl": 0.02601837982365396, "context": 0.05594804300926626}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.4345276653766632, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 17.31, "ce": 0.09418550948612392, "kl": 0.05120680821710266, "context": 0.11978891142643988}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.7729049324989319, "loss_scale": 4.0, "retries": 0, "seconds": 199.8, "max_memory_gb": 17.31, "ce": 0.12947783310664818, "kl": 0.10562019706048886, "context": 0.18228792445734143}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.7894666194915771, "loss_scale": 4.0, "retries": 0, "seconds": 200.3, "max_memory_gb": 17.31, "ce": 0.18561764410696924, "kl": 0.2108985335653415, "context": 0.2910669108387083}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.5426984429359436, "loss_scale": 4.0, "retries": 0, "seconds": 200.8, "max_memory_gb": 17.31, "ce": 0.05777652678079903, "kl": 0.03552672054502182, "context": 0.0755398852052167}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.4529976546764374, "loss_scale": 4.0, "retries": 0, "seconds": 201.4, "max_memory_gb": 17.31, "ce": 0.05645685759373009, "kl": 0.13602876703225775, "context": 0.12447124102618545}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.5601738691329956, "loss_scale": 4.0, "retries": 0, "seconds": 201.9, "max_memory_gb": 17.31, "ce": 0.09351804852485657, "kl": 0.046135150420013815, "context": 0.11658562766388059}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.28470781445503235, "loss_scale": 4.0, "retries": 0, "seconds": 202.4, "max_memory_gb": 17.31, "ce": 0.08288600808009505, "kl": 0.057864714646711946, "context": 0.11181836551986635}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.8656839728355408, "loss_scale": 4.0, "retries": 0, "seconds": 202.9, "max_memory_gb": 17.31, "ce": 0.030157615314237773, "kl": 0.12252438874565996, "context": 0.09141981019638479}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.43771329522132874, "loss_scale": 4.0, "retries": 0, "seconds": 203.5, "max_memory_gb": 17.31, "ce": 0.23119045852217823, "kl": 0.0842541391029954, "context": 0.27331753447651863}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 2.1641647815704346, "loss_scale": 4.0, "retries": 0, "seconds": 204.0, "max_memory_gb": 17.31, "ce": 0.34094935131724924, "kl": 0.2921482024830766, "context": 0.4870234637055546}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.26462647318840027, "loss_scale": 4.0, "retries": 0, "seconds": 204.5, "max_memory_gb": 17.31, "ce": 0.08240374340675771, "kl": 0.06525672793213744, "context": 0.11503210756927729}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.17902712523937225, "loss_scale": 4.0, "retries": 0, "seconds": 205.0, "max_memory_gb": 17.31, "ce": 0.011123534757643938, "kl": 0.02210443260264583, "context": 0.022175751044414937}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.5331514477729797, "loss_scale": 4.0, "retries": 0, "seconds": 205.6, "max_memory_gb": 17.31, "ce": 0.15347254695370793, "kl": 0.004575679762638174, "context": 0.15576038602739573}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.5833068490028381, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 17.31, "ce": 0.06992446933872998, "kl": 0.01324800053407671, "context": 0.0765484704170376}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.5965461134910583, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 17.31, "ce": 0.08970635948935524, "kl": 0.025684788823127747, "context": 0.10254875244572759}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.23022815585136414, "loss_scale": 4.0, "retries": 0, "seconds": 207.2, "max_memory_gb": 17.31, "ce": 0.03938582638511434, "kl": 0.005187864429899491, "context": 0.04197975812712684}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.4226451516151428, "loss_scale": 4.0, "retries": 0, "seconds": 207.7, "max_memory_gb": 17.31, "ce": 0.03315344266593456, "kl": 0.041137863910989836, "context": 0.05372237367555499}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 0.8137809634208679, "loss_scale": 4.0, "retries": 0, "seconds": 208.2, "max_memory_gb": 17.31, "ce": 0.06702431617304683, "kl": 0.054190998431295156, "context": 0.09411981515586376}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 0.986221432685852, "loss_scale": 4.0, "retries": 0, "seconds": 208.8, "max_memory_gb": 17.31, "ce": 0.20827149786055088, "kl": 0.18875558709260076, "context": 0.3026492912322283}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.08903282880783081, "loss_scale": 4.0, "retries": 0, "seconds": 209.3, "max_memory_gb": 17.31, "ce": 0.007050230953609571, "kl": 0.005582012207014486, "context": 0.009841236926149577}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 0.9232256412506104, "loss_scale": 4.0, "retries": 0, "seconds": 209.8, "max_memory_gb": 17.31, "ce": 0.02170577965443954, "kl": 0.6698426123475656, "context": 0.3566270829178393}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.3816762864589691, "loss_scale": 4.0, "retries": 0, "seconds": 210.4, "max_memory_gb": 17.31, "ce": 0.09186480089556426, "kl": 0.028116897447034717, "context": 0.10592325124889612}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.8012030124664307, "loss_scale": 4.0, "retries": 0, "seconds": 210.9, "max_memory_gb": 17.31, "ce": 0.04876160994172096, "kl": 0.017058019526302814, "context": 0.05729061842430383}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 0.7919465899467468, "loss_scale": 4.0, "retries": 0, "seconds": 211.4, "max_memory_gb": 17.31, "ce": 0.01379699306562543, "kl": 0.25691536644808366, "context": 0.14225467381766066}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.41384604573249817, "loss_scale": 4.0, "retries": 0, "seconds": 211.9, "max_memory_gb": 17.31, "ce": 0.11755361151881516, "kl": 0.0754660340026021, "context": 0.15528662665747106}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.5171099901199341, "loss_scale": 4.0, "retries": 0, "seconds": 212.5, "max_memory_gb": 17.31, "ce": 0.10343851847574115, "kl": 0.015279379382263869, "context": 0.11107820551842451}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.9284781813621521, "loss_scale": 4.0, "retries": 0, "seconds": 213.0, "max_memory_gb": 17.31, "ce": 0.16781632756465115, "kl": 0.0009731802019814495, "context": 0.16830292152008042}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 0.7761513590812683, "loss_scale": 4.0, "retries": 0, "seconds": 213.5, "max_memory_gb": 17.31, "ce": 0.036364297615364194, "kl": 0.08696598456299398, "context": 0.07984729006420821}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.0390734672546387, "loss_scale": 4.0, "retries": 0, "seconds": 214.1, "max_memory_gb": 17.31, "ce": 0.024611002765595913, "kl": 0.5686826963792555, "context": 0.3089523436501622}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.575989842414856, "loss_scale": 4.0, "retries": 0, "seconds": 214.6, "max_memory_gb": 17.31, "ce": 0.1458056648261845, "kl": 0.08032205933704972, "context": 0.18596669426187873}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.08626358211040497, "loss_scale": 4.0, "retries": 0, "seconds": 215.1, "max_memory_gb": 17.31, "ce": 0.007040403608698398, "kl": 0.008803075586911291, "context": 0.011441941314842552}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 1.1367511749267578, "loss_scale": 4.0, "retries": 0, "seconds": 215.7, "max_memory_gb": 17.31, "ce": 0.20655191922560334, "kl": 0.06767423384735594, "context": 0.24038902972824872}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.40347906947135925, "loss_scale": 4.0, "retries": 0, "seconds": 216.2, "max_memory_gb": 17.31, "ce": 0.029055142833385617, "kl": 0.01944268005900085, "context": 0.03877648385241628}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.563086986541748, "loss_scale": 4.0, "retries": 0, "seconds": 216.7, "max_memory_gb": 17.31, "ce": 0.16544038883876055, "kl": 0.019319786108098924, "context": 0.17510028113611042}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 1.5547802448272705, "loss_scale": 4.0, "retries": 0, "seconds": 217.3, "max_memory_gb": 17.31, "ce": 0.7744999135611579, "kl": 0.39118500816402957, "context": 0.9700923941563815}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 1.0584346055984497, "loss_scale": 4.0, "retries": 0, "seconds": 217.8, "max_memory_gb": 17.31, "ce": 0.2511384051758796, "kl": 0.2076520008558873, "context": 0.3549644062295556}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.5849858522415161, "loss_scale": 4.0, "retries": 0, "seconds": 218.3, "max_memory_gb": 17.31, "ce": 0.18366181728197262, "kl": 0.0799235324284382, "context": 0.22362358117243275}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.6782754063606262, "loss_scale": 4.0, "retries": 0, "seconds": 218.8, "max_memory_gb": 17.31, "ce": 0.05288371193455532, "kl": 0.0855918153465609, "context": 0.09567962030996569}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 1.939919114112854, "loss_scale": 4.0, "retries": 0, "seconds": 219.4, "max_memory_gb": 17.31, "ce": 0.2272919617826119, "kl": 0.24700721961562522, "context": 0.35079556971322745}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.5671954154968262, "loss_scale": 4.0, "retries": 0, "seconds": 219.9, "max_memory_gb": 17.31, "ce": 0.15249847620725632, "kl": 0.07828898966545239, "context": 0.1916429748525843}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.6464647054672241, "loss_scale": 4.0, "retries": 0, "seconds": 220.4, "max_memory_gb": 17.31, "ce": 0.05460002925246954, "kl": 0.03606017382116988, "context": 0.07263011508621275}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 1.110363483428955, "loss_scale": 4.0, "retries": 0, "seconds": 221.0, "max_memory_gb": 17.31, "ce": 0.19830372580327094, "kl": 0.009159039214864606, "context": 0.2028832493815571}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 1.295838475227356, "loss_scale": 4.0, "retries": 0, "seconds": 221.5, "max_memory_gb": 17.31, "ce": 0.2957859906600788, "kl": 0.09125510882586241, "context": 0.3414135533384979}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.12118624895811081, "loss_scale": 4.0, "retries": 0, "seconds": 222.0, "max_memory_gb": 17.31, "ce": 0.014132392243482172, "kl": 0.003175497051415732, "context": 0.015720140421763062}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 1.2519840002059937, "loss_scale": 4.0, "retries": 0, "seconds": 222.5, "max_memory_gb": 17.31, "ce": 0.40314388647675514, "kl": 0.09739952045492828, "context": 0.45184364914894104}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.20839661359786987, "loss_scale": 4.0, "retries": 0, "seconds": 223.1, "max_memory_gb": 17.31, "ce": 0.03268754982855171, "kl": 0.04707178386161104, "context": 0.056223442079499364}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.27522462606430054, "loss_scale": 4.0, "retries": 0, "seconds": 223.6, "max_memory_gb": 17.31, "ce": 0.07467691914644092, "kl": 0.04388765458134003, "context": 0.09662074560765177}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 2.0047473907470703, "loss_scale": 4.0, "retries": 0, "seconds": 224.1, "max_memory_gb": 17.31, "ce": 0.13552536754286848, "kl": 0.9745046955067664, "context": 0.6227777148596942}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.22882382571697235, "loss_scale": 4.0, "retries": 0, "seconds": 224.6, "max_memory_gb": 17.31, "ce": 0.04160066042095423, "kl": 0.011538372651557438, "context": 0.04736984637565911}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 0.7145722508430481, "loss_scale": 4.0, "retries": 0, "seconds": 225.2, "max_memory_gb": 17.31, "ce": 0.17436334720696323, "kl": 0.010129468471859582, "context": 0.17942808696534485}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 1.7712373733520508, "loss_scale": 4.0, "retries": 0, "seconds": 225.7, "max_memory_gb": 17.31, "ce": 0.37965712463483214, "kl": 0.405244879424572, "context": 0.5822795778512955}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.4652528762817383, "loss_scale": 4.0, "retries": 0, "seconds": 226.2, "max_memory_gb": 17.31, "ce": 0.04357929021352902, "kl": 0.030591574119171128, "context": 0.058875077636912465}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.8287930488586426, "loss_scale": 4.0, "retries": 0, "seconds": 226.8, "max_memory_gb": 17.31, "ce": 0.06667757226387039, "kl": 0.08028069362262613, "context": 0.10681792080868036}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 1.1349226236343384, "loss_scale": 4.0, "retries": 0, "seconds": 227.4, "max_memory_gb": 17.31, "ce": 0.19381682574748993, "kl": 0.06088698282837868, "context": 0.22426031436771154}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.7734172344207764, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 17.31, "ce": 0.07571904629003257, "kl": 0.022057140328797686, "context": 0.08674761647125706}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.3078854084014893, "loss_scale": 4.0, "retries": 0, "seconds": 228.4, "max_memory_gb": 17.31, "ce": 0.006745769176632166, "kl": 0.29264671245618956, "context": 0.15306913128006272}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.7407684326171875, "loss_scale": 4.0, "retries": 0, "seconds": 229.0, "max_memory_gb": 17.31, "ce": 0.01678245177026838, "kl": 0.0914362791008898, "context": 0.0625005904585123}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.48096227645874023, "loss_scale": 4.0, "retries": 0, "seconds": 229.5, "max_memory_gb": 17.31, "ce": 0.10678895452292636, "kl": 0.007370732764684362, "context": 0.11047431838233024}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 2.476978063583374, "loss_scale": 4.0, "retries": 0, "seconds": 230.1, "max_memory_gb": 17.31, "ce": 0.24289101618342102, "kl": 0.13142051687464118, "context": 0.30860127467894927}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.6172927021980286, "loss_scale": 4.0, "retries": 0, "seconds": 230.6, "max_memory_gb": 17.31, "ce": 0.07919021730776876, "kl": 0.06181252631358802, "context": 0.11009648162871599}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 0.7997560501098633, "loss_scale": 4.0, "retries": 0, "seconds": 231.1, "max_memory_gb": 17.31, "ce": 0.10580528539139777, "kl": 0.11703359286184423, "context": 0.16432207997422665}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.07119517773389816, "loss_scale": 4.0, "retries": 0, "seconds": 231.6, "max_memory_gb": 17.31, "ce": 0.007935886736959219, "kl": 0.00020805010171898175, "context": 0.00803991174325347}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.6560856699943542, "loss_scale": 4.0, "retries": 0, "seconds": 232.2, "max_memory_gb": 17.31, "ce": 0.10467533150222152, "kl": 0.06973219646533835, "context": 0.1395414259750396}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.5389385223388672, "loss_scale": 4.0, "retries": 0, "seconds": 232.7, "max_memory_gb": 17.31, "ce": 0.09995332779362798, "kl": 0.01854746983735822, "context": 0.1092270640656352}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 0.7917338609695435, "loss_scale": 4.0, "retries": 0, "seconds": 233.2, "max_memory_gb": 17.31, "ce": 0.16860945808002725, "kl": 0.06615504198998678, "context": 0.20168697537155822}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.7359633445739746, "loss_scale": 4.0, "retries": 0, "seconds": 233.7, "max_memory_gb": 17.31, "ce": 0.08787916228175163, "kl": 0.04013208742253482, "context": 0.10794520506169647}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 0.8040065169334412, "loss_scale": 4.0, "retries": 0, "seconds": 234.3, "max_memory_gb": 17.31, "ce": 0.14483833080157638, "kl": 0.18093820638023317, "context": 0.23530743923038244}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.35753384232521057, "loss_scale": 4.0, "retries": 0, "seconds": 234.8, "max_memory_gb": 17.31, "ce": 0.05051034362986684, "kl": 0.045115238362996024, "context": 0.07306796323973686}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 1.8170366287231445, "loss_scale": 4.0, "retries": 0, "seconds": 235.3, "max_memory_gb": 17.31, "ce": 0.23323713935678825, "kl": 0.2978420414729044, "context": 0.3821581582305953}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.29790198802948, "loss_scale": 4.0, "retries": 0, "seconds": 235.8, "max_memory_gb": 17.31, "ce": 0.08156865870114416, "kl": 0.01649206766160205, "context": 0.08981468970887363}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.43986308574676514, "loss_scale": 4.0, "retries": 0, "seconds": 236.4, "max_memory_gb": 17.31, "ce": 0.07175381388515234, "kl": 0.031870186438027304, "context": 0.08768890798091888}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.5021762251853943, "loss_scale": 4.0, "retries": 0, "seconds": 236.9, "max_memory_gb": 17.31, "ce": 0.11431948328390718, "kl": 0.03540050418814644, "context": 0.13201973331160843}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 0.8217553496360779, "loss_scale": 4.0, "retries": 0, "seconds": 237.4, "max_memory_gb": 17.31, "ce": 0.3559008296288084, "kl": 0.016572791266298736, "context": 0.36418721804511733}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 1.0860439538955688, "loss_scale": 4.0, "retries": 0, "seconds": 238.0, "max_memory_gb": 17.31, "ce": 0.10628588125109673, "kl": 0.14271500706672668, "context": 0.17764338245615363}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.5284717679023743, "loss_scale": 4.0, "retries": 0, "seconds": 238.5, "max_memory_gb": 17.31, "ce": 0.181235148338601, "kl": 0.023803103425962036, "context": 0.19313669891562313}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 1.6815688610076904, "loss_scale": 4.0, "retries": 0, "seconds": 239.0, "max_memory_gb": 17.31, "ce": 0.1792502086609602, "kl": 0.09466780768707395, "context": 0.22658411040902138}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.4315051734447479, "loss_scale": 4.0, "retries": 0, "seconds": 239.5, "max_memory_gb": 17.31, "ce": 0.020638933056034148, "kl": 0.00868192423513392, "context": 0.024979894573334605}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.6555394530296326, "loss_scale": 4.0, "retries": 0, "seconds": 240.1, "max_memory_gb": 17.31, "ce": 0.026888645748840645, "kl": 0.029803448132952326, "context": 0.041790369839873165}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.4997548758983612, "loss_scale": 4.0, "retries": 0, "seconds": 240.6, "max_memory_gb": 17.31, "ce": 0.022940574621316046, "kl": 0.07284201384754851, "context": 0.05936158017721027}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.0494458675384521, "loss_scale": 4.0, "retries": 0, "seconds": 241.1, "max_memory_gb": 17.31, "ce": 0.1532352373469621, "kl": 0.04817996491328813, "context": 0.17732522217556834}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.4599563479423523, "loss_scale": 4.0, "retries": 0, "seconds": 241.6, "max_memory_gb": 17.31, "ce": 0.05308067053556442, "kl": 0.028932540852110833, "context": 0.06754694133996964}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 1.1998947858810425, "loss_scale": 4.0, "retries": 0, "seconds": 242.2, "max_memory_gb": 17.31, "ce": 0.15980344789568335, "kl": 0.008421602607995737, "context": 0.16401424712967128}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 1.0069141387939453, "loss_scale": 4.0, "retries": 0, "seconds": 242.7, "max_memory_gb": 17.31, "ce": 0.11499012156855315, "kl": 0.1652874668361619, "context": 0.19763385318219662}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.21140316128730774, "loss_scale": 4.0, "retries": 0, "seconds": 243.2, "max_memory_gb": 17.31, "ce": 0.0235696459421888, "kl": 0.013266648165881634, "context": 0.030202970374375582}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.22820700705051422, "loss_scale": 4.0, "retries": 0, "seconds": 243.7, "max_memory_gb": 17.31, "ce": 0.025736105570103973, "kl": 0.03748695654212497, "context": 0.0444795829243958}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 0.6230515837669373, "loss_scale": 4.0, "retries": 0, "seconds": 244.3, "max_memory_gb": 17.31, "ce": 0.13735760003328323, "kl": 0.11132844806706998, "context": 0.19302182376850396}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 1.57383394241333, "loss_scale": 4.0, "retries": 0, "seconds": 244.8, "max_memory_gb": 17.31, "ce": 0.15780569799244404, "kl": 0.1123368251719512, "context": 0.21397411008365452}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.22459527850151062, "loss_scale": 4.0, "retries": 0, "seconds": 245.3, "max_memory_gb": 17.31, "ce": 0.02201431943103671, "kl": 0.02047867650981061, "context": 0.03225365711841732}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 0.8551491498947144, "loss_scale": 4.0, "retries": 0, "seconds": 245.9, "max_memory_gb": 17.31, "ce": 0.1835831084754318, "kl": 0.09318466947297566, "context": 0.23017543856985867}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.8088259696960449, "loss_scale": 4.0, "retries": 0, "seconds": 246.4, "max_memory_gb": 17.31, "ce": 0.05874063796363771, "kl": 0.00793648468243191, "context": 0.06270888086874038}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.2966333031654358, "loss_scale": 4.0, "retries": 0, "seconds": 246.9, "max_memory_gb": 17.31, "ce": 0.02938009350327775, "kl": 0.0018097618594765663, "context": 0.030284975306130946}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.06644527614116669, "loss_scale": 4.0, "retries": 0, "seconds": 247.4, "max_memory_gb": 17.31, "ce": 0.0062969623832032084, "kl": 0.0013922343241574708, "context": 0.0069930796744301915}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 1.4587676525115967, "loss_scale": 4.0, "retries": 0, "seconds": 248.0, "max_memory_gb": 17.31, "ce": 0.020101122703636065, "kl": 0.5825711521101766, "context": 0.31138671227381565}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 0.5518891215324402, "loss_scale": 4.0, "retries": 0, "seconds": 248.5, "max_memory_gb": 17.31, "ce": 0.014894038016791455, "kl": 0.2415256710955873, "context": 0.135656875470886}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.39747998118400574, "loss_scale": 4.0, "retries": 0, "seconds": 249.0, "max_memory_gb": 17.31, "ce": 0.048470022389665246, "kl": 0.013748733093962073, "context": 0.05534438695758581}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.2717917263507843, "loss_scale": 4.0, "retries": 0, "seconds": 249.5, "max_memory_gb": 17.31, "ce": 0.025798394752200693, "kl": 0.009991878483560868, "context": 0.030794333550147712}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 0.9021822214126587, "loss_scale": 4.0, "retries": 0, "seconds": 250.1, "max_memory_gb": 17.31, "ce": 0.10662063665222377, "kl": 0.03532180371985305, "context": 0.12428153469227254}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.20729660987854, "loss_scale": 4.0, "retries": 0, "seconds": 250.6, "max_memory_gb": 17.31, "ce": 0.004590832919348031, "kl": 0.3708981167146703, "context": 0.19003989710472524}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.281759262084961, "loss_scale": 4.0, "retries": 0, "seconds": 251.1, "max_memory_gb": 17.31, "ce": 0.17081722902366892, "kl": 0.16399126101168804, "context": 0.25281286134850234}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.15851567685604095, "loss_scale": 4.0, "retries": 0, "seconds": 251.6, "max_memory_gb": 17.31, "ce": 0.011794209014624357, "kl": 0.009165871939330827, "context": 0.01637714554090053}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.514805793762207, "loss_scale": 4.0, "retries": 0, "seconds": 252.2, "max_memory_gb": 17.31, "ce": 0.036213040380971506, "kl": 0.016749394817452412, "context": 0.04458773773512803}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.5166785717010498, "loss_scale": 4.0, "retries": 0, "seconds": 252.7, "max_memory_gb": 17.31, "ce": 0.09766297019086778, "kl": 0.016331874699972104, "context": 0.10582891083322465}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 1.2027976512908936, "loss_scale": 4.0, "retries": 0, "seconds": 253.3, "max_memory_gb": 17.31, "ce": 0.19631656113779172, "kl": 0.007639224995728, "context": 0.20013617194490507}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.3397985100746155, "loss_scale": 4.0, "retries": 0, "seconds": 253.8, "max_memory_gb": 17.31, "ce": 0.031130947085330263, "kl": 0.00705903087509796, "context": 0.034660463337786496}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 1.8921375274658203, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 17.31, "ce": 0.012755137315252796, "kl": 0.6145152369354037, "context": 0.32001274172216654}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.4561610221862793, "loss_scale": 4.0, "retries": 0, "seconds": 254.8, "max_memory_gb": 17.31, "ce": 0.025163792568491772, "kl": 0.03820072970211186, "context": 0.04426415741909295}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 1.7068464756011963, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 17.31, "ce": 0.3830773460213095, "kl": 0.2658348790719174, "context": 0.515994768589735}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.49974483251571655, "loss_scale": 4.0, "retries": 0, "seconds": 255.9, "max_memory_gb": 17.31, "ce": 0.16862262855283916, "kl": 0.011038304422982037, "context": 0.1741417811717838}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.36090075969696045, "loss_scale": 4.0, "retries": 0, "seconds": 256.4, "max_memory_gb": 17.31, "ce": 0.02175421838182956, "kl": 0.02211475736112334, "context": 0.03281159797916189}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.4922819137573242, "loss_scale": 4.0, "retries": 0, "seconds": 257.0, "max_memory_gb": 17.31, "ce": 0.10002782102674246, "kl": 0.023984297353308648, "context": 0.1120199728757143}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 1.076178789138794, "loss_scale": 4.0, "retries": 0, "seconds": 257.5, "max_memory_gb": 17.31, "ce": 0.14875241182744503, "kl": 0.5612515150569379, "context": 0.4293781667947769}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 1.5092185735702515, "loss_scale": 4.0, "retries": 0, "seconds": 258.0, "max_memory_gb": 17.31, "ce": 0.48547244630753994, "kl": 0.12490343628451228, "context": 0.547924148850143}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.9766793847084045, "loss_scale": 4.0, "retries": 0, "seconds": 258.6, "max_memory_gb": 17.31, "ce": 0.10276019526645541, "kl": 0.056165123474784195, "context": 0.13084275973960757}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 0.7108724117279053, "loss_scale": 4.0, "retries": 0, "seconds": 259.1, "max_memory_gb": 17.31, "ce": 0.13164706632960588, "kl": 0.030916931107640266, "context": 0.14710553211625665}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 0.718393862247467, "loss_scale": 4.0, "retries": 0, "seconds": 259.6, "max_memory_gb": 17.31, "ce": 0.045010002446360886, "kl": 0.06156385623398819, "context": 0.07579193054698408}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.3264992833137512, "loss_scale": 4.0, "retries": 0, "seconds": 260.1, "max_memory_gb": 17.31, "ce": 0.0995101552689448, "kl": 0.015204577102849726, "context": 0.10711244668345898}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 1.9929993152618408, "loss_scale": 4.0, "retries": 0, "seconds": 260.7, "max_memory_gb": 17.31, "ce": 0.3764300289331004, "kl": 0.08638528158189729, "context": 0.4196226631756872}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.247695192694664, "loss_scale": 4.0, "retries": 0, "seconds": 261.2, "max_memory_gb": 17.31, "ce": 0.02925803977996111, "kl": 0.010358862862631213, "context": 0.034437471855198964}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.62294602394104, "loss_scale": 4.0, "retries": 0, "seconds": 261.7, "max_memory_gb": 17.31, "ce": 0.08004456362687051, "kl": 0.16030265775043517, "context": 0.16019589436473325}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.3862920105457306, "loss_scale": 4.0, "retries": 0, "seconds": 262.2, "max_memory_gb": 17.31, "ce": 0.020411418983712792, "kl": 0.00869315660384018, "context": 0.02475799829699099}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 0.7679024338722229, "loss_scale": 4.0, "retries": 0, "seconds": 262.8, "max_memory_gb": 17.31, "ce": 0.08673377241939306, "kl": 0.024388261372223496, "context": 0.09892790392041206}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.175195336341858, "loss_scale": 4.0, "retries": 0, "seconds": 263.3, "max_memory_gb": 17.31, "ce": 0.01268297794740647, "kl": 0.49124509768444113, "context": 0.2583055233117193}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.5522776246070862, "loss_scale": 4.0, "retries": 0, "seconds": 263.8, "max_memory_gb": 17.31, "ce": 0.054136307910084724, "kl": 0.01046539458911866, "context": 0.05936900665983558}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.43289512395858765, "loss_scale": 4.0, "retries": 0, "seconds": 264.4, "max_memory_gb": 17.31, "ce": 0.027784997539129108, "kl": 0.03101389412768185, "context": 0.04329194547608495}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 0.9080492854118347, "loss_scale": 4.0, "retries": 0, "seconds": 264.9, "max_memory_gb": 17.31, "ce": 0.15130028812563978, "kl": 0.021886835660552606, "context": 0.16224370407871902}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.4110689163208008, "loss_scale": 4.0, "retries": 0, "seconds": 265.4, "max_memory_gb": 17.31, "ce": 0.12763950452790596, "kl": 0.027731757203582674, "context": 0.14150538708781824}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 1.0411031246185303, "loss_scale": 4.0, "retries": 0, "seconds": 265.9, "max_memory_gb": 17.31, "ce": 0.15029668164788745, "kl": 0.02628239542536903, "context": 0.1634378773451317}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 1.3209844827651978, "loss_scale": 4.0, "retries": 0, "seconds": 266.5, "max_memory_gb": 17.31, "ce": 0.2676500512752682, "kl": 0.012235701287863776, "context": 0.27376790158450603}
diff --git a/runs/single_llava/context/L14_s13/adapter/README.md b/runs/single_llava/context/L14_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L14_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L14_s13/adapter/adapter_config.json b/runs/single_llava/context/L14_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..31b2b4801fc10e0f5edfe6370c13e275c838a06a
--- /dev/null
+++ b/runs/single_llava/context/L14_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.14.self_attn.v_proj",
+    "model.language_model.layers.14.mlp.down_proj",
+    "model.language_model.layers.14.mlp.gate_proj",
+    "model.language_model.layers.14.self_attn.k_proj",
+    "model.language_model.layers.14.self_attn.o_proj",
+    "model.language_model.layers.14.mlp.up_proj",
+    "model.language_model.layers.14.self_attn.q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L14_s13/config.json b/runs/single_llava/context/L14_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..112debcf94099000301530f713e533a681f13927
--- /dev/null
+++ b/runs/single_llava/context/L14_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 14,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L14_s13/steps.jsonl b/runs/single_llava/context/L14_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..61677fe719ad7e2a9eecef6bdfcf5982c41c04c6
--- /dev/null
+++ b/runs/single_llava/context/L14_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.40095409750938416, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 16.45, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.3970106244087219, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 16.45, "ce": 2.6795505475020036, "kl": 0.010421675644465722, "context": 2.6847614366561174}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.8640565872192383, "loss_scale": 4.0, "retries": 0, "seconds": 2.5, "max_memory_gb": 16.45, "ce": 5.389914572238922, "kl": 0.17899961386137875, "context": 5.479414463043213}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.6002174615859985, "loss_scale": 4.0, "retries": 0, "seconds": 3.0, "max_memory_gb": 16.45, "ce": 3.821549555286765, "kl": 0.05122109541116515, "context": 3.8471601218916476}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.10333116352558136, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 16.46, "ce": 0.08608745178207755, "kl": 0.03672995849046856, "context": 0.10445243120193481}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.2705540657043457, "loss_scale": 4.0, "retries": 0, "seconds": 4.0, "max_memory_gb": 16.46, "ce": 1.4639207106083632, "kl": 0.1820984281366691, "context": 1.5549698695540428}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.6831176280975342, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 16.46, "ce": 3.4298244221135974, "kl": 0.10500784854957601, "context": 3.4823282533325255}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.7633374929428101, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 16.46, "ce": 3.6822783648967743, "kl": 0.16115933284163475, "context": 3.7628579810261726}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.6216635704040527, "loss_scale": 4.0, "retries": 0, "seconds": 5.5, "max_memory_gb": 16.46, "ce": 2.8270897269248962, "kl": 0.16140741179697216, "context": 2.9077935442328453}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.43183836340904236, "loss_scale": 4.0, "retries": 0, "seconds": 6.0, "max_memory_gb": 16.46, "ce": 1.6266660019755363, "kl": 0.05907746742013842, "context": 1.6562047824263573}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.6821915507316589, "loss_scale": 4.0, "retries": 0, "seconds": 6.5, "max_memory_gb": 16.46, "ce": 2.613087604753673, "kl": 0.009390223800437525, "context": 2.6177826756611466}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.47752952575683594, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 16.46, "ce": 1.5260633351281285, "kl": 0.01772516629716847, "context": 1.5349259267095476}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.7640547752380371, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 16.46, "ce": 2.6494307327084243, "kl": 0.007760223641525954, "context": 2.6533108348958194}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.0853530168533325, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 16.46, "ce": 3.859847202897072, "kl": 0.11793750454671681, "context": 3.918815977871418}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.8047823905944824, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 16.46, "ce": 2.5597151433466934, "kl": 0.2654697850521188, "context": 2.6924500922323205}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.5461804270744324, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 16.46, "ce": 1.3572616135934368, "kl": 0.009288137487601489, "context": 1.3619056390598416}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.523298978805542, "loss_scale": 4.0, "retries": 0, "seconds": 9.5, "max_memory_gb": 16.46, "ce": 1.5706396056339145, "kl": 0.0772920215385966, "context": 1.609285606071353}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.0922253131866455, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 16.47, "ce": 3.526051428169012, "kl": 0.09351023472845554, "context": 3.5728065222501755}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.7813056707382202, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 16.47, "ce": 5.255427718162537, "kl": 0.3295505925780162, "context": 5.420202970504761}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.032255228608846664, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 16.47, "ce": 0.01291770813986659, "kl": 0.002120274191838689, "context": 0.01397784543223679}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.1131919622421265, "loss_scale": 4.0, "retries": 0, "seconds": 11.5, "max_memory_gb": 16.47, "ce": 2.7169826715253294, "kl": 0.029362594745180104, "context": 2.731663978425786}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.187639832496643, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 16.47, "ce": 2.8730485886335373, "kl": 0.0534835736034438, "context": 2.899790348485112}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.1841790676116943, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 16.47, "ce": 2.5874788982328027, "kl": 0.5012225312530063, "context": 2.838090216508135}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.1163740158081055, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 16.47, "ce": 4.883448660373688, "kl": 0.014527099672704935, "context": 4.890712320804596}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 2.0607030391693115, "loss_scale": 4.0, "retries": 0, "seconds": 13.5, "max_memory_gb": 16.47, "ce": 4.105630345642567, "kl": 0.034742886666208506, "context": 4.123001761734486}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.7098028659820557, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 16.47, "ce": 5.035383522510529, "kl": 0.06383692554663867, "context": 5.067302048206329}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 2.134841203689575, "loss_scale": 4.0, "retries": 0, "seconds": 14.5, "max_memory_gb": 16.47, "ce": 4.0345066748559475, "kl": 0.08338465809356421, "context": 4.076199050992727}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.6114455461502075, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 16.47, "ce": 2.682475984096527, "kl": 0.23173975013196468, "context": 2.798345908522606}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.6016476154327393, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 16.47, "ce": 2.4990364643745124, "kl": 0.510543290496571, "context": 2.7543080446776003}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.924261212348938, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 16.47, "ce": 2.59440372698009, "kl": 0.03768770373426378, "context": 2.613247588276863}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 2.283142566680908, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 16.47, "ce": 2.6587978824973106, "kl": 0.0281366208801046, "context": 2.672866180539131}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 3.7472825050354004, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 16.47, "ce": 4.994524002075195, "kl": 0.015676290990086272, "context": 5.002362132072449}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.7281957864761353, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 16.47, "ce": 2.323904595337808, "kl": 0.1811987836845219, "context": 2.4145039431750774}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 2.2309298515319824, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 16.47, "ce": 2.4398878253996372, "kl": 0.5993131613358855, "context": 2.739544490352273}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 2.6568894386291504, "loss_scale": 4.0, "retries": 0, "seconds": 18.4, "max_memory_gb": 16.47, "ce": 2.4629994737915695, "kl": 0.028675561829004437, "context": 2.4773372653871775}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 2.2477757930755615, "loss_scale": 4.0, "retries": 0, "seconds": 18.9, "max_memory_gb": 16.47, "ce": 2.3454897701740265, "kl": 0.0070646225940436125, "context": 2.3490220718085766}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 2.5462558269500732, "loss_scale": 4.0, "retries": 0, "seconds": 19.4, "max_memory_gb": 16.47, "ce": 2.240980964154005, "kl": 0.3288935183081776, "context": 2.405427787452936}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 4.863332748413086, "loss_scale": 4.0, "retries": 0, "seconds": 19.8, "max_memory_gb": 16.47, "ce": 4.194817006587982, "kl": 0.471922296055709, "context": 4.430778086185455}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.8925278186798096, "loss_scale": 4.0, "retries": 0, "seconds": 20.3, "max_memory_gb": 16.47, "ce": 1.259476387873292, "kl": 0.03262852202169597, "context": 1.2757906108163297}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 3.0839357376098633, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 16.47, "ce": 2.407693348824978, "kl": 0.11469828826375306, "context": 2.465042471885681}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.6490496397018433, "loss_scale": 4.0, "retries": 0, "seconds": 21.2, "max_memory_gb": 16.47, "ce": 1.3132710866630077, "kl": 0.1685709897428751, "context": 1.3975565284490585}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 4.294135093688965, "loss_scale": 4.0, "retries": 0, "seconds": 21.7, "max_memory_gb": 16.47, "ce": 1.9069021893665195, "kl": 0.007263108826009557, "context": 1.9105338184162974}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 4.10496711730957, "loss_scale": 4.0, "retries": 0, "seconds": 22.1, "max_memory_gb": 16.47, "ce": 1.993170804809779, "kl": 0.012037342588882893, "context": 1.9991894881241024}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.9970600605010986, "loss_scale": 4.0, "retries": 0, "seconds": 22.6, "max_memory_gb": 16.47, "ce": 1.173780918121338, "kl": 0.24218820240639616, "context": 1.2948750420473516}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 2.435319185256958, "loss_scale": 4.0, "retries": 0, "seconds": 23.1, "max_memory_gb": 16.47, "ce": 1.1550011560320854, "kl": 0.055587744573131204, "context": 1.1827950477600098}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 4.332507610321045, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 16.47, "ce": 1.8088439926505089, "kl": 0.07368117361329496, "context": 1.8456845581531525}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 5.14188814163208, "loss_scale": 4.0, "retries": 0, "seconds": 24.0, "max_memory_gb": 16.47, "ce": 1.4755855547264218, "kl": 0.033960934495553374, "context": 1.4925660218577832}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 10.074446678161621, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 16.47, "ce": 2.662796102464199, "kl": 0.016585486780968495, "context": 2.671088859438896}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 9.495004653930664, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 16.47, "ce": 2.5065710553899407, "kl": 0.09395543625578284, "context": 2.5535487830638885}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 9.452685356140137, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 16.47, "ce": 2.5156056927517056, "kl": 0.04918546712724492, "context": 2.5401984276250005}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 3.098405599594116, "loss_scale": 4.0, "retries": 0, "seconds": 25.8, "max_memory_gb": 16.47, "ce": 1.0794640555977821, "kl": 0.12573620188049972, "context": 1.142332179006189}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 7.779293537139893, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 16.47, "ce": 1.4287497252225876, "kl": 0.11674099665833637, "context": 1.4871201813220978}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 11.725625991821289, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 16.47, "ce": 2.4877301827073097, "kl": 0.1879114112816751, "context": 2.581685855984688}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 13.471559524536133, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 16.47, "ce": 2.1784149929881096, "kl": 0.04049502790439874, "context": 2.198662519454956}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 9.317594528198242, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 16.47, "ce": 1.2352120131254196, "kl": 0.019167776248650625, "context": 1.2447958984412253}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 18.860427856445312, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 16.47, "ce": 2.335784077644348, "kl": 0.32221675640903413, "context": 2.4968923926353455}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 4.82528018951416, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 16.47, "ce": 0.6295270868577063, "kl": 0.10294500587042421, "context": 0.6809996152296662}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 8.86641788482666, "loss_scale": 4.0, "retries": 0, "seconds": 29.0, "max_memory_gb": 16.47, "ce": 0.8714459750335664, "kl": 0.00605881069895986, "context": 0.8744753897190094}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 5.11892557144165, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 16.47, "ce": 0.8128960058093071, "kl": 0.07385462603997439, "context": 0.8498233128339052}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 10.25710391998291, "loss_scale": 4.0, "retries": 0, "seconds": 29.9, "max_memory_gb": 16.47, "ce": 0.9576330217532814, "kl": 0.014686448848806322, "context": 0.9649762725457549}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 12.081525802612305, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 16.47, "ce": 0.7815849408507347, "kl": 0.09098244478809647, "context": 0.827076181769371}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 4.815271377563477, "loss_scale": 4.0, "retries": 0, "seconds": 30.9, "max_memory_gb": 16.47, "ce": 0.3935527610592544, "kl": 0.04038222099188715, "context": 0.4137438638135791}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 10.419766426086426, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 16.47, "ce": 1.05339699331671, "kl": 0.08191736074877554, "context": 1.094355653040111}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 8.092826843261719, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 16.49, "ce": 0.6125555001199245, "kl": 0.048214030102826655, "context": 0.6366625148802996}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 8.74825382232666, "loss_scale": 4.0, "retries": 0, "seconds": 32.3, "max_memory_gb": 16.49, "ce": 0.8346355482935905, "kl": 0.11522500892169774, "context": 0.8922480344772339}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 2.804847002029419, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 16.49, "ce": 0.3666313021676615, "kl": 0.06129830746795051, "context": 0.3972804570803419}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 5.222054958343506, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 16.49, "ce": 0.40013042837381363, "kl": 0.02954949182458222, "context": 0.4149051681160927}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 3.0173542499542236, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 16.49, "ce": 0.17845441680401564, "kl": 0.07371188083197922, "context": 0.2153103593736887}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 2.58091139793396, "loss_scale": 4.0, "retries": 0, "seconds": 34.1, "max_memory_gb": 16.49, "ce": 0.24567674240097404, "kl": 0.024130498233716935, "context": 0.2577419909648597}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 1.5876567363739014, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 16.49, "ce": 0.15992990648373961, "kl": 0.00436038343468681, "context": 0.1621100977063179}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 1.5864452123641968, "loss_scale": 4.0, "retries": 0, "seconds": 35.1, "max_memory_gb": 16.49, "ce": 0.17075826972723007, "kl": 0.0173432482406497, "context": 0.1794298905879259}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 1.8127243518829346, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 16.49, "ce": 0.18004296347498894, "kl": 0.10890745767392218, "context": 0.23449669405817986}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.3801928162574768, "loss_scale": 4.0, "retries": 0, "seconds": 36.0, "max_memory_gb": 16.49, "ce": 0.22247748961672187, "kl": 0.03529508039355278, "context": 0.2401250354014337}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 0.49020400643348694, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 16.49, "ce": 0.17421591095626354, "kl": 0.022046577360015363, "context": 0.1852392004802823}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 0.7431647777557373, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 16.49, "ce": 0.5847371071577072, "kl": 0.19693333934992552, "context": 0.6832037642598152}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 0.7187508940696716, "loss_scale": 4.0, "retries": 0, "seconds": 37.4, "max_memory_gb": 16.49, "ce": 0.09330590814352036, "kl": 0.40508148074150085, "context": 0.2958466522395611}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 0.9181243181228638, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 16.49, "ce": 0.31250662356615067, "kl": 0.035085528856143355, "context": 0.3300493862479925}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 0.47772935032844543, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 16.49, "ce": 0.09711822867393494, "kl": 0.05872640653979033, "context": 0.12648143339902163}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 0.36304736137390137, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 16.49, "ce": 0.1073327586054802, "kl": 0.07827959994756384, "context": 0.146472557913512}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 0.41409608721733093, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 16.49, "ce": 0.057031569769605994, "kl": 0.007657634501811117, "context": 0.06086038867942989}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 0.7148348093032837, "loss_scale": 4.0, "retries": 0, "seconds": 39.7, "max_memory_gb": 16.49, "ce": 0.16369134653359652, "kl": 0.051625875814352185, "context": 0.18950428627431393}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 0.6198093891143799, "loss_scale": 4.0, "retries": 0, "seconds": 40.1, "max_memory_gb": 16.49, "ce": 0.14456757716834545, "kl": 0.044640267034992576, "context": 0.1668877061456442}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 0.6238330602645874, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 16.49, "ce": 0.12085921783000231, "kl": 0.16505634039640427, "context": 0.20338738895952702}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 0.5538256764411926, "loss_scale": 4.0, "retries": 0, "seconds": 41.1, "max_memory_gb": 16.49, "ce": 0.15844253334216774, "kl": 0.06901614637172315, "context": 0.19295060285367072}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.3181087374687195, "loss_scale": 4.0, "retries": 0, "seconds": 41.5, "max_memory_gb": 16.49, "ce": 0.0598630104213953, "kl": 0.03715319924594951, "context": 0.07843961217440665}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 0.5334892272949219, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 16.49, "ce": 0.13183782994747162, "kl": 0.22782199492212385, "context": 0.2457488253712654}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 0.5167044401168823, "loss_scale": 4.0, "retries": 0, "seconds": 42.5, "max_memory_gb": 16.49, "ce": 0.08507393393665552, "kl": 0.05383572413120419, "context": 0.11199179105460644}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 0.48269525170326233, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 16.49, "ce": 0.2067416599020362, "kl": 0.02100751910620602, "context": 0.21724542183801532}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.2241165190935135, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 16.49, "ce": 0.11283433646894991, "kl": 0.0057992665824713185, "context": 0.11573397368192673}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 0.5007070302963257, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 16.49, "ce": 0.14836760377511382, "kl": 0.020284335827454925, "context": 0.1585097718052566}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.49261873960494995, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 16.49, "ce": 0.3122694715857506, "kl": 0.08936256286688149, "context": 0.35695076547563076}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 0.3042667508125305, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 16.49, "ce": 0.10037139081396163, "kl": 0.08467388129793108, "context": 0.14270833134651184}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 0.4628134071826935, "loss_scale": 4.0, "retries": 0, "seconds": 45.2, "max_memory_gb": 16.49, "ce": 0.06197229470126331, "kl": 0.058110798825509846, "context": 0.09102769335731864}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 0.4243711829185486, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 16.49, "ce": 0.11938055139034986, "kl": 0.1016116093378514, "context": 0.1701863557100296}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 0.5570381879806519, "loss_scale": 4.0, "retries": 0, "seconds": 46.2, "max_memory_gb": 16.49, "ce": 0.19579204730689526, "kl": 0.06178927095606923, "context": 0.22668668441474438}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 0.5934134125709534, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 16.49, "ce": 0.14887369144707918, "kl": 0.03417187725426629, "context": 0.16595963295549154}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 0.54264897108078, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 16.49, "ce": 0.16758022271096706, "kl": 0.019872752483934164, "context": 0.17751659639179707}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 0.28043311834335327, "loss_scale": 4.0, "retries": 0, "seconds": 47.5, "max_memory_gb": 16.49, "ce": 0.039052450098097324, "kl": 0.03460546779388096, "context": 0.05635518254712224}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.8525003790855408, "loss_scale": 4.0, "retries": 0, "seconds": 48.0, "max_memory_gb": 16.49, "ce": 0.19668448762968183, "kl": 0.07447649340610951, "context": 0.23392272926867008}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 0.4575943648815155, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 16.49, "ce": 0.14762125629931688, "kl": 0.1149051491956925, "context": 0.20507382601499557}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 0.6405865550041199, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 16.49, "ce": 0.08954057935625315, "kl": 0.3730528987944126, "context": 0.27606702875345945}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 1.2688987255096436, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 16.49, "ce": 0.3348759114742279, "kl": 0.15018490888178349, "context": 0.40996837243437767}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 0.5323187112808228, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 16.49, "ce": 0.047002716921269894, "kl": 0.07624780520563945, "context": 0.08512661908753216}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 0.31642940640449524, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 16.49, "ce": 0.06052226200699806, "kl": 0.042204297555144876, "context": 0.08162440860178322}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 0.5678738951683044, "loss_scale": 4.0, "retries": 0, "seconds": 50.8, "max_memory_gb": 16.49, "ce": 0.09536426421254873, "kl": 0.1989718994591385, "context": 0.19485021103173494}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 0.4268009066581726, "loss_scale": 4.0, "retries": 0, "seconds": 51.2, "max_memory_gb": 16.49, "ce": 0.10169319063425064, "kl": 0.04354024864733219, "context": 0.12346331775188446}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.30584192276000977, "loss_scale": 4.0, "retries": 0, "seconds": 51.7, "max_memory_gb": 16.49, "ce": 0.11529741901904345, "kl": 0.015573778437101282, "context": 0.12308430718258023}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.4926677644252777, "loss_scale": 4.0, "retries": 0, "seconds": 52.2, "max_memory_gb": 16.49, "ce": 0.1299993726424873, "kl": 0.022470648982562125, "context": 0.14123469381593168}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 0.7458255290985107, "loss_scale": 4.0, "retries": 0, "seconds": 52.6, "max_memory_gb": 16.49, "ce": 0.1916444320231676, "kl": 0.4172865781583823, "context": 0.40028770826756954}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 0.6807906627655029, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 16.49, "ce": 0.1342969974502921, "kl": 0.052461380284512416, "context": 0.16052768612280488}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 0.3295195996761322, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 16.49, "ce": 0.043669285252690315, "kl": 0.08607867643149802, "context": 0.08670862461440265}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 0.3842599391937256, "loss_scale": 4.0, "retries": 0, "seconds": 54.0, "max_memory_gb": 16.49, "ce": 0.08205892471596599, "kl": 0.013100325915729627, "context": 0.08860908937640488}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 0.5542569160461426, "loss_scale": 4.0, "retries": 0, "seconds": 54.5, "max_memory_gb": 16.49, "ce": 0.13225167174823582, "kl": 0.042388575384393334, "context": 0.15344596467912197}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 1.2051365375518799, "loss_scale": 4.0, "retries": 0, "seconds": 55.0, "max_memory_gb": 16.49, "ce": 0.20776316500268877, "kl": 0.7919291772413999, "context": 0.6037277383729815}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.7672551870346069, "loss_scale": 4.0, "retries": 0, "seconds": 55.4, "max_memory_gb": 16.49, "ce": 0.17910021753050387, "kl": 0.3034399007447064, "context": 0.33082017255946994}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.356219083070755, "loss_scale": 4.0, "retries": 0, "seconds": 55.9, "max_memory_gb": 16.49, "ce": 0.12379840854555368, "kl": 0.006101737642893568, "context": 0.1268492757808417}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 0.9529770016670227, "loss_scale": 4.0, "retries": 0, "seconds": 56.4, "max_memory_gb": 16.49, "ce": 0.18276701006107032, "kl": 0.011375316593330353, "context": 0.1884546612855047}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.2605549395084381, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 16.49, "ce": 0.1442178904544562, "kl": 0.03788626469031442, "context": 0.16316102957352996}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.3580184876918793, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 16.49, "ce": 0.09527663514018059, "kl": 0.06706063001183793, "context": 0.12880695145577192}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 0.8658574819564819, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 16.49, "ce": 0.16256467765197158, "kl": 0.061303463648073375, "context": 0.19321641558781266}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 0.7788321375846863, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 16.49, "ce": 0.21469864575192332, "kl": 0.08046854610074661, "context": 0.25493292324244976}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 0.19098298251628876, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 16.49, "ce": 0.11579315562266856, "kl": 0.013519436732167378, "context": 0.12255286960862577}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.035266399383545, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 16.49, "ce": 0.5000025806948543, "kl": 0.07960688063758425, "context": 0.5398060134612024}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 0.8268725275993347, "loss_scale": 4.0, "retries": 0, "seconds": 59.6, "max_memory_gb": 16.49, "ce": 0.27346643060445786, "kl": 0.03648861180408858, "context": 0.2917107343673706}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.31618398427963257, "loss_scale": 4.0, "retries": 0, "seconds": 60.1, "max_memory_gb": 16.49, "ce": 0.12236876972019672, "kl": 0.020531751593807712, "context": 0.13263464905321598}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 0.5630642175674438, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 16.49, "ce": 0.2771804016083479, "kl": 0.10536141961347312, "context": 0.3298611156642437}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.18882004916667938, "loss_scale": 4.0, "retries": 0, "seconds": 61.0, "max_memory_gb": 16.49, "ce": 0.03747683996334672, "kl": 0.004476858601265121, "context": 0.03971526981331408}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.0116578340530396, "loss_scale": 4.0, "retries": 0, "seconds": 61.5, "max_memory_gb": 16.49, "ce": 0.0793207174865529, "kl": 0.4257483850378776, "context": 0.2921949048759416}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 0.22746583819389343, "loss_scale": 4.0, "retries": 0, "seconds": 61.9, "max_memory_gb": 16.49, "ce": 0.04023323021829128, "kl": 0.09250922323553823, "context": 0.08648784132674336}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 1.2964472770690918, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 16.49, "ce": 0.2443711031228304, "kl": 0.5114317442639731, "context": 0.5000869678333402}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 1.0620867013931274, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 16.49, "ce": 0.25176119804382324, "kl": 0.043110796716064215, "context": 0.2733165957033634}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.4488937854766846, "loss_scale": 4.0, "retries": 0, "seconds": 63.3, "max_memory_gb": 16.49, "ce": 0.3348778891377151, "kl": 0.29445826206938364, "context": 0.48210701276548207}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 0.6649450063705444, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 16.49, "ce": 0.14600739814341068, "kl": 0.13289209408685565, "context": 0.2124534472823143}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.6417251825332642, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 16.49, "ce": 0.15223134076222777, "kl": 0.013465778989484534, "context": 0.15896422462537885}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 0.2860777676105499, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 16.49, "ce": 0.09620970743708313, "kl": 0.020341191586339846, "context": 0.10638030059635639}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.5224363207817078, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 16.49, "ce": 0.16064728796482086, "kl": 0.03544528083875775, "context": 0.17836992535740137}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.4979495704174042, "loss_scale": 4.0, "retries": 0, "seconds": 65.7, "max_memory_gb": 16.49, "ce": 0.13012562203221023, "kl": 0.01451662571344059, "context": 0.13738393771927804}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.47806617617607117, "loss_scale": 4.0, "retries": 0, "seconds": 66.1, "max_memory_gb": 16.49, "ce": 0.1043601012788713, "kl": 0.07117210049182177, "context": 0.13994615338742733}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 0.8315428495407104, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 16.49, "ce": 0.12772442772984505, "kl": 0.17624848245759495, "context": 0.2158486694097519}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.4708045721054077, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 16.49, "ce": 0.056501423474401236, "kl": 0.016175828641280532, "context": 0.06458933581598103}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.4842630624771118, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 16.49, "ce": 0.1475634193047881, "kl": 0.010288504650816321, "context": 0.1527076787315309}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.38238829374313354, "loss_scale": 4.0, "retries": 0, "seconds": 68.0, "max_memory_gb": 16.49, "ce": 0.13835517736151814, "kl": 0.015637190954294056, "context": 0.146173772867769}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.3225178122520447, "loss_scale": 4.0, "retries": 0, "seconds": 68.5, "max_memory_gb": 16.49, "ce": 0.12921647308394313, "kl": 0.05777007614960894, "context": 0.1581015078118071}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.9476665258407593, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 16.49, "ce": 0.1513649094849825, "kl": 0.10319786623585969, "context": 0.2029638383537531}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.16442745923995972, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 16.49, "ce": 0.025913843885064125, "kl": 0.0021958661382086575, "context": 0.027011777041479945}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.2753714621067047, "loss_scale": 4.0, "retries": 0, "seconds": 69.9, "max_memory_gb": 16.49, "ce": 0.10822304524481297, "kl": 0.010789322783239186, "context": 0.11361770471557975}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.8809874057769775, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 16.49, "ce": 0.20926233869977295, "kl": 0.3731991774402559, "context": 0.3958619326585904}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.3933931291103363, "loss_scale": 4.0, "retries": 0, "seconds": 70.8, "max_memory_gb": 16.49, "ce": 0.01778303983155638, "kl": 0.06987805059179664, "context": 0.052722065476700664}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.3492506444454193, "loss_scale": 4.0, "retries": 0, "seconds": 71.3, "max_memory_gb": 16.49, "ce": 0.08762054168619215, "kl": 0.009303019098297227, "context": 0.0922720511443913}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.6225278377532959, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 16.49, "ce": 0.1599350064061582, "kl": 0.015714399341959506, "context": 0.16779221035540104}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.5766628980636597, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 16.49, "ce": 0.12624926853459328, "kl": 0.015500207984587178, "context": 0.13399937050417066}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 1.0199998617172241, "loss_scale": 4.0, "retries": 0, "seconds": 72.7, "max_memory_gb": 16.49, "ce": 0.0488387206569314, "kl": 0.5514460512786172, "context": 0.32456173142418265}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.4413723647594452, "loss_scale": 4.0, "retries": 0, "seconds": 73.1, "max_memory_gb": 16.49, "ce": 0.04642307944595814, "kl": 0.2739824224845506, "context": 0.18341428320854902}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.559483528137207, "loss_scale": 4.0, "retries": 0, "seconds": 73.6, "max_memory_gb": 16.49, "ce": 0.46687490376643836, "kl": 0.34322065362357534, "context": 0.6384852603077888}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 1.6398963928222656, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 16.49, "ce": 0.4959320966154337, "kl": 0.1854545588430483, "context": 0.5886593733448535}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.5604804754257202, "loss_scale": 4.0, "retries": 0, "seconds": 74.5, "max_memory_gb": 16.49, "ce": 0.12289335415698588, "kl": 0.15114489345432958, "context": 0.19846580072771758}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.31989404559135437, "loss_scale": 4.0, "retries": 0, "seconds": 75.0, "max_memory_gb": 16.49, "ce": 0.12250020145438612, "kl": 0.05210294383869041, "context": 0.14855167549103498}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.6954536437988281, "loss_scale": 4.0, "retries": 0, "seconds": 75.4, "max_memory_gb": 16.49, "ce": 0.2126956507563591, "kl": 0.3159166444092989, "context": 0.3706539701670408}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.34200477600097656, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 16.49, "ce": 0.04256215924397111, "kl": 0.07838913757586852, "context": 0.08175672590732574}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.461394339799881, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 16.49, "ce": 0.10476363985799253, "kl": 0.05448986554984003, "context": 0.13200857350602746}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.5933650732040405, "loss_scale": 4.0, "retries": 0, "seconds": 76.8, "max_memory_gb": 16.49, "ce": 0.2254717079922557, "kl": 0.034028703565127216, "context": 0.24248605407774448}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.3029869496822357, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 16.49, "ce": 0.08659732830710709, "kl": 0.021680251753423363, "context": 0.09743745438754559}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.9807755947113037, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 16.49, "ce": 0.07929239934310317, "kl": 0.08941712657724565, "context": 0.12400095746852458}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.5679401755332947, "loss_scale": 4.0, "retries": 0, "seconds": 78.2, "max_memory_gb": 16.49, "ce": 0.20235061645507812, "kl": 0.019380772951990366, "context": 0.21204099617898464}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.5190773606300354, "loss_scale": 4.0, "retries": 0, "seconds": 78.7, "max_memory_gb": 16.49, "ce": 0.10827316553331912, "kl": 0.047406487571606704, "context": 0.13197641330771148}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.8122556209564209, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 16.49, "ce": 0.15073541831225157, "kl": 0.0563977601705119, "context": 0.17893429566174746}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.6266049742698669, "loss_scale": 4.0, "retries": 0, "seconds": 79.6, "max_memory_gb": 16.49, "ce": 0.14062269497662783, "kl": 0.01704491446434986, "context": 0.14914515637792647}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.38527676463127136, "loss_scale": 4.0, "retries": 0, "seconds": 80.1, "max_memory_gb": 16.49, "ce": 0.06413776287809014, "kl": 0.09079484071116894, "context": 0.10953518189489841}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.5922598838806152, "loss_scale": 4.0, "retries": 0, "seconds": 80.6, "max_memory_gb": 16.49, "ce": 0.20061036944389343, "kl": 1.0291117021115497, "context": 0.7151662320829928}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.13365577161312103, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 16.49, "ce": 0.015449804486706853, "kl": 0.0035078819819318596, "context": 0.01720374613068998}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 1.361243486404419, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 16.49, "ce": 0.033816987299360335, "kl": 0.2185593950562179, "context": 0.14309668773785233}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.4285989999771118, "loss_scale": 4.0, "retries": 0, "seconds": 82.0, "max_memory_gb": 16.49, "ce": 0.0726038923021406, "kl": 0.0053033426811452955, "context": 0.07525556301698089}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 0.5418692231178284, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 16.49, "ce": 0.10042039887048304, "kl": 0.08837265864713117, "context": 0.14460672228597105}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.4523215591907501, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 16.49, "ce": 0.07525644614361227, "kl": 0.06649326192564331, "context": 0.10850308137014508}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.7010094523429871, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 16.49, "ce": 0.11895357258617878, "kl": 0.1811119425110519, "context": 0.2095095431432128}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.1520749181509018, "loss_scale": 4.0, "retries": 0, "seconds": 83.8, "max_memory_gb": 16.49, "ce": 0.019778864923864603, "kl": 0.005787184039945714, "context": 0.022672456689178944}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.731447160243988, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 16.49, "ce": 0.1829193407902494, "kl": 0.14058482688415097, "context": 0.25321175053250045}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.3992633819580078, "loss_scale": 4.0, "retries": 0, "seconds": 84.7, "max_memory_gb": 16.49, "ce": 0.12863323581404984, "kl": 0.08658912940882146, "context": 0.1719278017990291}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.677241861820221, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 16.49, "ce": 0.3330298219807446, "kl": 0.0056184602799476124, "context": 0.3358390489593148}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 0.8961654901504517, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 16.49, "ce": 0.2922156364656985, "kl": 0.047572774506988935, "context": 0.31600201642140746}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.41975778341293335, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 16.49, "ce": 0.05773221957497299, "kl": 0.04293608118314296, "context": 0.07920025987550616}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.5677491426467896, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 16.49, "ce": 0.23466025502420962, "kl": 0.007153448561439291, "context": 0.2382369841216132}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.07362914085388184, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 16.49, "ce": 0.009470564313232899, "kl": 0.0026844010499189608, "context": 0.010812764638103545}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.6775428056716919, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 16.49, "ce": 0.041376328095793724, "kl": 0.08422684462857433, "context": 0.08348975295666605}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.1961163580417633, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 16.49, "ce": 0.02464078285265714, "kl": 0.012114842167648021, "context": 0.030698203248903155}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.2501660883426666, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 16.49, "ce": 0.041368802543729544, "kl": 0.004658752121031284, "context": 0.043698177207261324}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.9031912684440613, "loss_scale": 4.0, "retries": 0, "seconds": 88.9, "max_memory_gb": 16.49, "ce": 0.07125617982819676, "kl": 0.363751468990813, "context": 0.253131908364594}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.2850895822048187, "loss_scale": 4.0, "retries": 0, "seconds": 89.4, "max_memory_gb": 16.49, "ce": 0.11884018231648952, "kl": 0.04063716215387103, "context": 0.13915876619284973}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.2741353511810303, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 16.49, "ce": 0.03202782338485122, "kl": 0.0030350778833962977, "context": 0.033545361598953605}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.2752012014389038, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 16.49, "ce": 0.014390720636583865, "kl": 0.42016801179852337, "context": 0.2244747239165008}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.850608766078949, "loss_scale": 4.0, "retries": 0, "seconds": 90.7, "max_memory_gb": 16.49, "ce": 0.29183074133470654, "kl": 0.04596756957471371, "context": 0.3148145251907408}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.3396790027618408, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 16.49, "ce": 0.1747065489180386, "kl": 0.006687325054372195, "context": 0.17805021279491484}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.2133064568042755, "loss_scale": 4.0, "retries": 0, "seconds": 91.7, "max_memory_gb": 16.49, "ce": 0.03307036845944822, "kl": 0.036329517839476466, "context": 0.05123512726277113}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 1.6085224151611328, "loss_scale": 4.0, "retries": 0, "seconds": 92.1, "max_memory_gb": 16.49, "ce": 0.3769539329223335, "kl": 0.01888828445225954, "context": 0.3863980695605278}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.39076656103134155, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 16.49, "ce": 0.09406741813290864, "kl": 0.011611786758294329, "context": 0.0998733073938638}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.2224457561969757, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 16.49, "ce": 0.024333589943125844, "kl": 0.023853208709624596, "context": 0.03626019507646561}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.4072534441947937, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 16.49, "ce": 0.08968448103405535, "kl": 0.026098536334757227, "context": 0.10273375199176371}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.42714399099349976, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 16.49, "ce": 0.1254171160981059, "kl": 0.03845334710786119, "context": 0.14464378729462624}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.6196445226669312, "loss_scale": 4.0, "retries": 0, "seconds": 94.4, "max_memory_gb": 16.49, "ce": 0.05075641814619303, "kl": 0.254857523352257, "context": 0.1781851826235652}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.4569329619407654, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 16.49, "ce": 0.21241131285205483, "kl": 0.013244106085039675, "context": 0.21903336886316538}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.4417020380496979, "loss_scale": 4.0, "retries": 0, "seconds": 95.5, "max_memory_gb": 16.49, "ce": 0.06581430207006633, "kl": 0.009502144530415535, "context": 0.07056537619791925}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.05544797703623772, "loss_scale": 4.0, "retries": 0, "seconds": 95.9, "max_memory_gb": 16.49, "ce": 0.006638965802267194, "kl": 0.003795687298406847, "context": 0.008536809356883168}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.6153566241264343, "loss_scale": 4.0, "retries": 0, "seconds": 96.4, "max_memory_gb": 16.49, "ce": 0.14098320482298732, "kl": 0.0855614747852087, "context": 0.18376393988728523}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.48368361592292786, "loss_scale": 4.0, "retries": 0, "seconds": 96.9, "max_memory_gb": 16.49, "ce": 0.08723354793619365, "kl": 0.028912397450767457, "context": 0.10168974660336971}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.8394849300384521, "loss_scale": 4.0, "retries": 0, "seconds": 97.3, "max_memory_gb": 16.49, "ce": 0.04837074293754995, "kl": 0.41879844226059504, "context": 0.2577699712710455}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.5713576078414917, "loss_scale": 4.0, "retries": 0, "seconds": 97.8, "max_memory_gb": 16.49, "ce": 0.10576289420714602, "kl": 0.0868328066480899, "context": 0.1491793048917316}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.5017204284667969, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 16.49, "ce": 0.5917852667625993, "kl": 0.47016446413181257, "context": 0.8268675026483834}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.10767904669046402, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 16.49, "ce": 0.013221331289969385, "kl": 0.005292082110827323, "context": 0.01586737239267677}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.3894801139831543, "loss_scale": 4.0, "retries": 0, "seconds": 99.3, "max_memory_gb": 16.49, "ce": 0.1775781367905438, "kl": 0.01290350581984967, "context": 0.18402988696470857}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.5636200904846191, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 16.49, "ce": 0.21774925850331783, "kl": 0.09796212147375627, "context": 0.26673032285179943}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.5232396125793457, "loss_scale": 4.0, "retries": 0, "seconds": 100.3, "max_memory_gb": 16.49, "ce": 0.19846140744630247, "kl": 0.10675808624364436, "context": 0.2518404512666166}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 1.0366712808609009, "loss_scale": 4.0, "retries": 0, "seconds": 100.8, "max_memory_gb": 16.49, "ce": 0.1297914031893015, "kl": 0.6705842651426792, "context": 0.4650835506618023}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.7791160941123962, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 16.49, "ce": 0.3270495851757005, "kl": 0.05668986824457534, "context": 0.35539450496435165}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 0.949048638343811, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 16.49, "ce": 0.21395510667935014, "kl": 0.33394061541184783, "context": 0.38092541065998375}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 1.6897625923156738, "loss_scale": 4.0, "retries": 0, "seconds": 102.3, "max_memory_gb": 16.49, "ce": 0.3913428051164374, "kl": 0.004431257286341861, "context": 0.3935584358405322}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.1837036609649658, "loss_scale": 4.0, "retries": 0, "seconds": 102.8, "max_memory_gb": 16.49, "ce": 0.17938300338573754, "kl": 0.07744704640936106, "context": 0.21810652408748865}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.5601137280464172, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 16.49, "ce": 0.06158155400771648, "kl": 0.042981392367437365, "context": 0.0830722501850687}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.813641369342804, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 16.49, "ce": 0.16456225630827248, "kl": 0.41729519376531243, "context": 0.37320984806865454}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.4116780459880829, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 16.49, "ce": 0.11137104500085115, "kl": 0.029699606995563954, "context": 0.12622084841132164}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.6109987497329712, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 16.49, "ce": 0.0906227792147547, "kl": 0.022033402317902073, "context": 0.10163947730325162}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.5871860980987549, "loss_scale": 4.0, "retries": 0, "seconds": 105.3, "max_memory_gb": 16.49, "ce": 0.09705124306492507, "kl": 0.08651352836750448, "context": 0.14030800852924585}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.3953654766082764, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 16.49, "ce": 0.26945695374161005, "kl": 0.20316444843774661, "context": 0.37103919172659516}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.5988931059837341, "loss_scale": 4.0, "retries": 0, "seconds": 106.3, "max_memory_gb": 16.49, "ce": 0.16707144118845463, "kl": 0.047288617060985416, "context": 0.19071575161069632}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.61159348487854, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 16.49, "ce": 0.06579447933472693, "kl": 0.12985710334032774, "context": 0.13072303496301174}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.6644483208656311, "loss_scale": 4.0, "retries": 0, "seconds": 107.3, "max_memory_gb": 16.49, "ce": 0.06038550520315766, "kl": 0.044840354938060045, "context": 0.08280568104237318}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.5688025951385498, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 16.49, "ce": 0.07388035184703767, "kl": 0.019449280167464167, "context": 0.0836049928329885}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.10719085484743118, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 16.49, "ce": 0.010523117845878005, "kl": 0.0046637196792289615, "context": 0.012854977860115469}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.8296752572059631, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 16.49, "ce": 0.06664079561596736, "kl": 0.07663439027965069, "context": 0.10495799034833908}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.9075672626495361, "loss_scale": 4.0, "retries": 0, "seconds": 109.3, "max_memory_gb": 16.49, "ce": 0.23210805654525757, "kl": 0.1532438051654026, "context": 0.30872995825484395}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.6980519890785217, "loss_scale": 4.0, "retries": 0, "seconds": 109.8, "max_memory_gb": 16.49, "ce": 0.2531855134293437, "kl": 0.11449370335321873, "context": 0.31043236318510026}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.6137673258781433, "loss_scale": 4.0, "retries": 0, "seconds": 110.3, "max_memory_gb": 16.49, "ce": 0.23556938057299703, "kl": 0.02180894983757753, "context": 0.2464738606940955}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.20991896092891693, "loss_scale": 4.0, "retries": 0, "seconds": 110.8, "max_memory_gb": 16.49, "ce": 0.02506689541041851, "kl": 0.007359551542322151, "context": 0.028746671276167035}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.2190619707107544, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 16.49, "ce": 0.12139121722429991, "kl": 0.1291251581897086, "context": 0.18595379032194614}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.6480725407600403, "loss_scale": 4.0, "retries": 0, "seconds": 111.8, "max_memory_gb": 16.49, "ce": 0.13247373700141907, "kl": 0.02218609393457882, "context": 0.14356678072363138}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.972912609577179, "loss_scale": 4.0, "retries": 0, "seconds": 112.3, "max_memory_gb": 16.49, "ce": 0.3365855978336185, "kl": 0.024498108454281464, "context": 0.3488346446538344}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.435619592666626, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 16.49, "ce": 0.08611670741811395, "kl": 0.03566857361875009, "context": 0.10395099292509258}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 0.8656817674636841, "loss_scale": 4.0, "retries": 0, "seconds": 113.3, "max_memory_gb": 16.49, "ce": 0.15558273321948946, "kl": 0.055671741174137424, "context": 0.18341860291548073}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.5127785801887512, "loss_scale": 4.0, "retries": 0, "seconds": 113.7, "max_memory_gb": 16.49, "ce": 0.04678897617850453, "kl": 0.11340265793842264, "context": 0.10349030373618007}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.5797640085220337, "loss_scale": 4.0, "retries": 0, "seconds": 114.2, "max_memory_gb": 16.49, "ce": 0.1369806167203933, "kl": 0.031451526418095455, "context": 0.15270637930370867}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.7770348787307739, "loss_scale": 4.0, "retries": 0, "seconds": 114.7, "max_memory_gb": 16.49, "ce": 0.12137720407918096, "kl": 0.04202273685950786, "context": 0.142388568026945}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.49174830317497253, "loss_scale": 4.0, "retries": 0, "seconds": 115.2, "max_memory_gb": 16.49, "ce": 0.15101049095392227, "kl": 0.014607215649448335, "context": 0.15831409767270088}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.45540469884872437, "loss_scale": 4.0, "retries": 0, "seconds": 115.7, "max_memory_gb": 16.49, "ce": 0.10820796375628561, "kl": 0.04004697487107478, "context": 0.12823145010042936}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.9041573405265808, "loss_scale": 4.0, "retries": 0, "seconds": 116.2, "max_memory_gb": 16.49, "ce": 0.017620802740566432, "kl": 0.1744481768400874, "context": 0.10484489274676889}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 1.28529953956604, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 16.49, "ce": 0.27428843965753913, "kl": 0.0721424973453395, "context": 0.3103596873115748}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 1.0153982639312744, "loss_scale": 4.0, "retries": 0, "seconds": 117.2, "max_memory_gb": 16.49, "ce": 0.373710103565827, "kl": 0.007275287003722042, "context": 0.37734774756245315}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.21588808298110962, "loss_scale": 4.0, "retries": 0, "seconds": 117.7, "max_memory_gb": 16.49, "ce": 0.017365275649353862, "kl": 0.019287649796751793, "context": 0.02700910100247711}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.48498860001564026, "loss_scale": 4.0, "retries": 0, "seconds": 118.2, "max_memory_gb": 16.49, "ce": 0.06193838774925098, "kl": 0.03421547031030059, "context": 0.07904612529091537}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.5719579458236694, "loss_scale": 4.0, "retries": 0, "seconds": 118.7, "max_memory_gb": 16.49, "ce": 0.09116409532725811, "kl": 0.03852327959612012, "context": 0.1104257358238101}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.4050007164478302, "loss_scale": 4.0, "retries": 0, "seconds": 119.2, "max_memory_gb": 16.49, "ce": 0.05868100607767701, "kl": 0.024376966553973034, "context": 0.07086948957294226}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.42646524310112, "loss_scale": 4.0, "retries": 0, "seconds": 119.7, "max_memory_gb": 16.49, "ce": 0.04352322907652706, "kl": 0.01161631515424233, "context": 0.04933138540945947}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.7532181143760681, "loss_scale": 4.0, "retries": 0, "seconds": 120.2, "max_memory_gb": 16.49, "ce": 0.13278720394009724, "kl": 0.02758107634872431, "context": 0.1465777406701818}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.5844497084617615, "loss_scale": 4.0, "retries": 0, "seconds": 120.7, "max_memory_gb": 16.49, "ce": 0.034424071200191975, "kl": 0.12412409269018099, "context": 0.09648611769080162}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.7753492593765259, "loss_scale": 4.0, "retries": 0, "seconds": 121.3, "max_memory_gb": 16.49, "ce": 0.1373435031855479, "kl": 0.006014343794959132, "context": 0.14035068033263087}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.9044702649116516, "loss_scale": 4.0, "retries": 0, "seconds": 121.7, "max_memory_gb": 16.49, "ce": 0.13191690249368548, "kl": 0.25494274497032166, "context": 0.2593882717192173}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.7709348797798157, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 16.49, "ce": 0.10895634163171053, "kl": 0.2383138898294419, "context": 0.22811328619718552}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.4459632635116577, "loss_scale": 4.0, "retries": 0, "seconds": 122.7, "max_memory_gb": 16.49, "ce": 0.07303167134523392, "kl": 0.027534690207176027, "context": 0.08679901505820453}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.7318071722984314, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 16.49, "ce": 0.08007048111176118, "kl": 0.021488648006197764, "context": 0.09081480704480782}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.45209750533103943, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 16.49, "ce": 0.034221988753415644, "kl": 0.10030787100549787, "context": 0.08437592582777143}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.3246059715747833, "loss_scale": 4.0, "retries": 0, "seconds": 124.3, "max_memory_gb": 16.49, "ce": 0.012552783591672778, "kl": 0.032861613399290945, "context": 0.02898359124083072}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.568222165107727, "loss_scale": 4.0, "retries": 0, "seconds": 124.7, "max_memory_gb": 16.49, "ce": 0.06377650704234838, "kl": 0.1315537773189135, "context": 0.12955339578911662}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.7021852135658264, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 16.49, "ce": 0.11989594157785177, "kl": 0.08722903160378337, "context": 0.16351045854389668}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 1.1105847358703613, "loss_scale": 4.0, "retries": 0, "seconds": 125.8, "max_memory_gb": 16.49, "ce": 0.16255790693685412, "kl": 0.11430615984136239, "context": 0.21971099474467337}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.7339210510253906, "loss_scale": 4.0, "retries": 0, "seconds": 126.3, "max_memory_gb": 16.49, "ce": 0.14065251080319285, "kl": 0.047854717005975544, "context": 0.1645798683166504}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 1.7522225379943848, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 16.49, "ce": 0.17595380346756428, "kl": 0.08950992817699444, "context": 0.22070876380894333}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.4311321973800659, "loss_scale": 4.0, "retries": 0, "seconds": 127.3, "max_memory_gb": 16.49, "ce": 0.07189590553753078, "kl": 0.10290596392587759, "context": 0.12334888754412532}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 1.4524496793746948, "loss_scale": 4.0, "retries": 0, "seconds": 127.8, "max_memory_gb": 16.49, "ce": 0.40086260018870234, "kl": 0.09084237719071098, "context": 0.44628380227368325}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.6835477352142334, "loss_scale": 4.0, "retries": 0, "seconds": 128.3, "max_memory_gb": 16.49, "ce": 0.16243277001194656, "kl": 0.004523283161688596, "context": 0.16469440702348948}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.458018958568573, "loss_scale": 4.0, "retries": 0, "seconds": 128.8, "max_memory_gb": 16.49, "ce": 0.21059885760769248, "kl": 0.05430183815769851, "context": 0.23774977773427963}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 2.633493661880493, "loss_scale": 4.0, "retries": 0, "seconds": 129.3, "max_memory_gb": 16.49, "ce": 0.36508757155388594, "kl": 0.5524882172376238, "context": 0.6413316689431667}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 1.1873557567596436, "loss_scale": 4.0, "retries": 0, "seconds": 129.8, "max_memory_gb": 16.49, "ce": 0.20424414332956076, "kl": 0.00593382287206623, "context": 0.20721105474513024}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.2962417006492615, "loss_scale": 4.0, "retries": 0, "seconds": 130.3, "max_memory_gb": 16.49, "ce": 0.026377977337688208, "kl": 0.04894688900094479, "context": 0.050851420033723116}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 0.8157946467399597, "loss_scale": 4.0, "retries": 0, "seconds": 130.7, "max_memory_gb": 16.49, "ce": 0.020175080397166312, "kl": 0.33716683470265707, "context": 0.18875849642790854}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.5436698198318481, "loss_scale": 4.0, "retries": 0, "seconds": 131.2, "max_memory_gb": 16.49, "ce": 0.021652580704540014, "kl": 0.13934829429490492, "context": 0.0913267278810963}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.5065981149673462, "loss_scale": 4.0, "retries": 0, "seconds": 131.7, "max_memory_gb": 16.49, "ce": 0.11268955422565341, "kl": 0.04307521024020389, "context": 0.13422716222703457}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 1.3032722473144531, "loss_scale": 4.0, "retries": 0, "seconds": 132.2, "max_memory_gb": 16.49, "ce": 0.538528339297045, "kl": 0.3437541265739128, "context": 0.7104053805815056}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.10588734596967697, "loss_scale": 4.0, "retries": 0, "seconds": 132.7, "max_memory_gb": 16.49, "ce": 0.00819621013943106, "kl": 0.005515574448509142, "context": 0.010953997145406902}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.4016583263874054, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 16.49, "ce": 0.01433095766697079, "kl": 0.12875221079593757, "context": 0.07870706683024764}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 1.2378809452056885, "loss_scale": 4.0, "retries": 0, "seconds": 133.7, "max_memory_gb": 16.49, "ce": 0.14533270185347646, "kl": 0.2225750836150837, "context": 0.25662024039775133}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 0.8298504948616028, "loss_scale": 4.0, "retries": 0, "seconds": 134.2, "max_memory_gb": 16.49, "ce": 0.07553158304654062, "kl": 0.1166628782011685, "context": 0.13386302068829536}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.2687474489212036, "loss_scale": 4.0, "retries": 0, "seconds": 134.7, "max_memory_gb": 16.49, "ce": 0.0204991273349151, "kl": 0.007126580647309311, "context": 0.0240624169819057}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 1.1361163854599, "loss_scale": 4.0, "retries": 0, "seconds": 135.2, "max_memory_gb": 16.49, "ce": 0.27609644446056336, "kl": 0.09952375368447974, "context": 0.32585831987671554}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.4786723852157593, "loss_scale": 4.0, "retries": 0, "seconds": 135.7, "max_memory_gb": 16.49, "ce": 0.09024053066968918, "kl": 0.01023748105399136, "context": 0.09535926976241171}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.4512394964694977, "loss_scale": 4.0, "retries": 0, "seconds": 136.2, "max_memory_gb": 16.49, "ce": 0.08063431567279622, "kl": 0.1675361949528451, "context": 0.1644024213310331}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 0.5467017889022827, "loss_scale": 4.0, "retries": 0, "seconds": 136.7, "max_memory_gb": 16.49, "ce": 0.05640638677868992, "kl": 0.013414800206192012, "context": 0.0631137863965705}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.48511865735054016, "loss_scale": 4.0, "retries": 0, "seconds": 137.2, "max_memory_gb": 16.49, "ce": 0.10311902617104352, "kl": 0.018714078891207464, "context": 0.11247606249526143}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.12139830738306046, "loss_scale": 4.0, "retries": 0, "seconds": 137.7, "max_memory_gb": 16.49, "ce": 0.00925806793384254, "kl": 0.016019366019463632, "context": 0.017267750692553818}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 1.2622650861740112, "loss_scale": 4.0, "retries": 0, "seconds": 138.2, "max_memory_gb": 16.49, "ce": 0.10202008299529552, "kl": 0.03488647626363672, "context": 0.1194633188424632}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.10863882303237915, "loss_scale": 4.0, "retries": 0, "seconds": 138.7, "max_memory_gb": 16.49, "ce": 0.00901548215188086, "kl": 0.0057742005155887455, "context": 0.01190258248243481}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.3520859479904175, "loss_scale": 4.0, "retries": 0, "seconds": 139.2, "max_memory_gb": 16.49, "ce": 0.04045608872547746, "kl": 0.07021493998036021, "context": 0.07556355826091021}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.6737887859344482, "loss_scale": 4.0, "retries": 0, "seconds": 139.7, "max_memory_gb": 16.49, "ce": 0.10552578401984647, "kl": 0.05244901780679356, "context": 0.13175028911791742}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.07282106578350067, "loss_scale": 4.0, "retries": 0, "seconds": 140.2, "max_memory_gb": 16.49, "ce": 0.006305783928837627, "kl": 0.0013763420283794403, "context": 0.006993954943027347}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.481512427330017, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 16.49, "ce": 0.05636959953699261, "kl": 0.6175430319271982, "context": 0.3651411118917167}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.4765529930591583, "loss_scale": 4.0, "retries": 0, "seconds": 141.2, "max_memory_gb": 16.49, "ce": 0.045416723820380867, "kl": 0.07967960802125162, "context": 0.08525652519892901}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 0.7342100739479065, "loss_scale": 4.0, "retries": 0, "seconds": 141.7, "max_memory_gb": 16.49, "ce": 0.07991500245407224, "kl": 0.027988156769424677, "context": 0.09390908107161522}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 1.5002977848052979, "loss_scale": 4.0, "retries": 0, "seconds": 142.2, "max_memory_gb": 16.49, "ce": 0.15047356323339045, "kl": 0.0938182712852722, "context": 0.1973826951580122}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 2.0892467498779297, "loss_scale": 4.0, "retries": 0, "seconds": 142.7, "max_memory_gb": 16.49, "ce": 0.25847597426036373, "kl": 0.4703956535086036, "context": 0.49367379769682884}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.3879042863845825, "loss_scale": 4.0, "retries": 0, "seconds": 143.2, "max_memory_gb": 16.49, "ce": 0.0828956050099805, "kl": 0.023393638650304638, "context": 0.09459242760203779}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.2352460622787476, "loss_scale": 4.0, "retries": 0, "seconds": 143.7, "max_memory_gb": 16.49, "ce": 0.43770938320085406, "kl": 0.16645276022609323, "context": 0.5209357591811568}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.3021293878555298, "loss_scale": 4.0, "retries": 0, "seconds": 144.2, "max_memory_gb": 16.49, "ce": 0.012890508922282606, "kl": 0.10882326585124247, "context": 0.06730214273557067}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 1.1757469177246094, "loss_scale": 4.0, "retries": 0, "seconds": 144.7, "max_memory_gb": 16.49, "ce": 0.2618827238911763, "kl": 0.057225068972911686, "context": 0.29049525782465935}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 1.6243383884429932, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 16.49, "ce": 0.13953568297438323, "kl": 0.1430507963123091, "context": 0.21106108179083094}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.23792672157287598, "loss_scale": 4.0, "retries": 0, "seconds": 145.7, "max_memory_gb": 16.49, "ce": 0.02790953393559903, "kl": 0.00719230708637042, "context": 0.03150568692944944}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.42500069737434387, "loss_scale": 4.0, "retries": 0, "seconds": 146.2, "max_memory_gb": 16.49, "ce": 0.052979803178459406, "kl": 0.015755010448629037, "context": 0.06085730710765347}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.21031078696250916, "loss_scale": 4.0, "retries": 0, "seconds": 146.7, "max_memory_gb": 16.49, "ce": 0.015027879737317562, "kl": 0.006659151942585595, "context": 0.01835745549760759}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 2.9986751079559326, "loss_scale": 4.0, "retries": 0, "seconds": 147.2, "max_memory_gb": 16.49, "ce": 0.1349087037378922, "kl": 0.2053450939420145, "context": 0.23758125060703605}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 0.7185783386230469, "loss_scale": 4.0, "retries": 0, "seconds": 147.6, "max_memory_gb": 16.49, "ce": 0.012481531244702637, "kl": 0.10558710678014904, "context": 0.06527508341241628}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 1.055113673210144, "loss_scale": 4.0, "retries": 0, "seconds": 148.1, "max_memory_gb": 16.49, "ce": 0.13503321935422719, "kl": 0.018019328395894263, "context": 0.1440428839996457}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 1.099784016609192, "loss_scale": 4.0, "retries": 0, "seconds": 148.7, "max_memory_gb": 16.49, "ce": 0.012505955412052572, "kl": 0.654852432970074, "context": 0.33993217488750815}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 1.5981265306472778, "loss_scale": 4.0, "retries": 0, "seconds": 149.1, "max_memory_gb": 16.49, "ce": 0.2027362243970856, "kl": 0.020524626539554447, "context": 0.2129985383944586}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.40582993626594543, "loss_scale": 4.0, "retries": 0, "seconds": 149.6, "max_memory_gb": 16.49, "ce": 0.043537845835089684, "kl": 0.19903580815298483, "context": 0.14305575285106897}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.6790255904197693, "loss_scale": 4.0, "retries": 0, "seconds": 150.1, "max_memory_gb": 16.49, "ce": 0.1486606881953776, "kl": 0.017529948929222883, "context": 0.15742565819527954}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.23831352591514587, "loss_scale": 4.0, "retries": 0, "seconds": 150.6, "max_memory_gb": 16.49, "ce": 0.029861846298445016, "kl": 0.009805070876609534, "context": 0.03476438170764595}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 1.2135847806930542, "loss_scale": 4.0, "retries": 0, "seconds": 151.1, "max_memory_gb": 16.49, "ce": 0.14287625171709806, "kl": 0.040854003324056976, "context": 0.16330325626768172}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.9871915578842163, "loss_scale": 4.0, "retries": 0, "seconds": 151.6, "max_memory_gb": 16.49, "ce": 0.13881068001501262, "kl": 0.10334285208955407, "context": 0.1904820988420397}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.4245723485946655, "loss_scale": 4.0, "retries": 0, "seconds": 152.1, "max_memory_gb": 16.49, "ce": 0.0474325847462751, "kl": 0.04256079241167754, "context": 0.06871298211626709}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 1.0464351177215576, "loss_scale": 4.0, "retries": 0, "seconds": 152.6, "max_memory_gb": 16.49, "ce": 0.16867721802555025, "kl": 0.6379817060660571, "context": 0.4876680797897279}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.6583315134048462, "loss_scale": 4.0, "retries": 0, "seconds": 153.1, "max_memory_gb": 16.49, "ce": 0.12304715672507882, "kl": 0.013746365788392723, "context": 0.12992033595219254}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.7214298844337463, "loss_scale": 4.0, "retries": 0, "seconds": 153.6, "max_memory_gb": 16.49, "ce": 0.027610770892351866, "kl": 0.4082111061852629, "context": 0.23171632597222924}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 0.8718123435974121, "loss_scale": 4.0, "retries": 0, "seconds": 154.1, "max_memory_gb": 16.49, "ce": 0.08635064674308524, "kl": 0.1060068137885537, "context": 0.1393540520220995}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 0.8192521929740906, "loss_scale": 4.0, "retries": 0, "seconds": 154.6, "max_memory_gb": 16.49, "ce": 0.13746501109562814, "kl": 0.09140019848564407, "context": 0.18316511041484773}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.9089828729629517, "loss_scale": 4.0, "retries": 0, "seconds": 155.1, "max_memory_gb": 16.49, "ce": 0.0787562916520983, "kl": 0.09774406674841885, "context": 0.12762832874432206}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.28655266761779785, "loss_scale": 4.0, "retries": 0, "seconds": 155.6, "max_memory_gb": 16.49, "ce": 0.018272335233632475, "kl": 0.03298922476096777, "context": 0.03476694750133902}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.3008876442909241, "loss_scale": 4.0, "retries": 0, "seconds": 156.1, "max_memory_gb": 16.49, "ce": 0.009710890240967274, "kl": 0.028392712934873998, "context": 0.023907246300950646}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.1542600691318512, "loss_scale": 4.0, "retries": 0, "seconds": 156.6, "max_memory_gb": 16.49, "ce": 0.014991875039413571, "kl": 0.007576423067803262, "context": 0.01878008688800037}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.779497504234314, "loss_scale": 4.0, "retries": 0, "seconds": 157.1, "max_memory_gb": 16.49, "ce": 0.18674083845689893, "kl": 0.017159695671580266, "context": 0.19532068853732198}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.48425450921058655, "loss_scale": 4.0, "retries": 0, "seconds": 157.6, "max_memory_gb": 16.49, "ce": 0.044727032305672765, "kl": 0.03927286621183157, "context": 0.0643634648877196}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 0.9183388948440552, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 16.49, "ce": 0.09053969391970895, "kl": 0.01820676797069609, "context": 0.09964307610061951}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.6145930886268616, "loss_scale": 4.0, "retries": 0, "seconds": 158.6, "max_memory_gb": 16.49, "ce": 0.09559716691728681, "kl": 0.12251649213430937, "context": 0.15685541252605617}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.8089218139648438, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 16.49, "ce": 0.09270083566661924, "kl": 0.1816255462326808, "context": 0.18351360852830112}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.6548569798469543, "loss_scale": 4.0, "retries": 0, "seconds": 159.6, "max_memory_gb": 16.49, "ce": 0.16815356840379536, "kl": 0.014250841937609948, "context": 0.1752789905294776}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 1.5521256923675537, "loss_scale": 4.0, "retries": 0, "seconds": 160.1, "max_memory_gb": 16.49, "ce": 0.2821451807394624, "kl": 0.08676777593791485, "context": 0.32552906265482306}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 0.9601732492446899, "loss_scale": 4.0, "retries": 0, "seconds": 160.6, "max_memory_gb": 16.49, "ce": 0.13980637153144926, "kl": 0.04751563301397255, "context": 0.16356418398208916}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 1.3348767757415771, "loss_scale": 4.0, "retries": 0, "seconds": 161.1, "max_memory_gb": 16.49, "ce": 0.313073797384277, "kl": 0.03136661695316434, "context": 0.3287571056280285}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 1.0162473917007446, "loss_scale": 4.0, "retries": 0, "seconds": 161.6, "max_memory_gb": 16.49, "ce": 0.1600492752622813, "kl": 0.13451511994935572, "context": 0.22730683139525354}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 2.921605348587036, "loss_scale": 4.0, "retries": 0, "seconds": 162.1, "max_memory_gb": 16.49, "ce": 0.3865890473825857, "kl": 0.29276996676344424, "context": 0.5329740529414266}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.2411516159772873, "loss_scale": 4.0, "retries": 0, "seconds": 162.6, "max_memory_gb": 16.49, "ce": 0.011316087213344872, "kl": 0.05232330796388851, "context": 0.037477740610484034}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 0.7376086115837097, "loss_scale": 4.0, "retries": 0, "seconds": 163.1, "max_memory_gb": 16.49, "ce": 0.15832328191027045, "kl": 0.08677443768829107, "context": 0.20171049609780312}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 0.7460811138153076, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 16.49, "ce": 0.19211957207880914, "kl": 0.19095357228070498, "context": 0.287596357986331}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.7971405386924744, "loss_scale": 4.0, "retries": 0, "seconds": 164.1, "max_memory_gb": 16.49, "ce": 0.15552964591188356, "kl": 0.010379516432294622, "context": 0.16071940236724913}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.954414427280426, "loss_scale": 4.0, "retries": 0, "seconds": 164.6, "max_memory_gb": 16.49, "ce": 0.18741099629551172, "kl": 0.08010031175217591, "context": 0.22746114782057703}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 1.2483419179916382, "loss_scale": 4.0, "retries": 0, "seconds": 165.1, "max_memory_gb": 16.49, "ce": 0.18380964372772723, "kl": 0.03822501446120441, "context": 0.20292215468361974}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 1.3068135976791382, "loss_scale": 4.0, "retries": 0, "seconds": 165.6, "max_memory_gb": 16.49, "ce": 0.2556749687064439, "kl": 0.8787843399913982, "context": 0.6950671412050724}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.31914857029914856, "loss_scale": 4.0, "retries": 0, "seconds": 166.1, "max_memory_gb": 16.49, "ce": 0.01817988627590239, "kl": 0.05636819583014585, "context": 0.046363985165953636}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.33582499623298645, "loss_scale": 4.0, "retries": 0, "seconds": 166.6, "max_memory_gb": 16.49, "ce": 0.030362065182998776, "kl": 0.020640570321120322, "context": 0.04068235168233514}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.46276018023490906, "loss_scale": 4.0, "retries": 0, "seconds": 167.1, "max_memory_gb": 16.49, "ce": 0.11629799695219845, "kl": 0.20607501290214714, "context": 0.21933550341054797}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 1.143947958946228, "loss_scale": 4.0, "retries": 0, "seconds": 167.6, "max_memory_gb": 16.49, "ce": 0.18437596992589533, "kl": 0.004796072898898274, "context": 0.18677400157321244}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 0.753383457660675, "loss_scale": 4.0, "retries": 0, "seconds": 168.1, "max_memory_gb": 16.49, "ce": 0.16681870445609093, "kl": 0.032830704702064395, "context": 0.18323405552655458}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.11634337902069092, "loss_scale": 4.0, "retries": 0, "seconds": 168.6, "max_memory_gb": 16.49, "ce": 0.010502153949346393, "kl": 0.004900179026662954, "context": 0.012952243618201464}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.18308627605438232, "loss_scale": 4.0, "retries": 0, "seconds": 169.1, "max_memory_gb": 16.49, "ce": 0.0840290046762675, "kl": 0.02166387926263269, "context": 0.09486094221938401}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.56085205078125, "loss_scale": 4.0, "retries": 0, "seconds": 169.6, "max_memory_gb": 16.49, "ce": 0.1084100790321827, "kl": 0.0582074336707592, "context": 0.13751379773020744}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.3955386877059937, "loss_scale": 4.0, "retries": 0, "seconds": 170.1, "max_memory_gb": 16.49, "ce": 0.04918229265604168, "kl": 0.2829638622060884, "context": 0.19066423026379198}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 2.3081560134887695, "loss_scale": 4.0, "retries": 0, "seconds": 170.6, "max_memory_gb": 16.49, "ce": 0.16996321978513151, "kl": 0.3003302877768874, "context": 0.3201283607631922}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.5733405351638794, "loss_scale": 4.0, "retries": 0, "seconds": 171.1, "max_memory_gb": 16.49, "ce": 0.11970049515366554, "kl": 0.07565984483517241, "context": 0.15753042500000447}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.5733271837234497, "loss_scale": 4.0, "retries": 0, "seconds": 171.5, "max_memory_gb": 16.49, "ce": 0.1726303247269243, "kl": 0.0362939772894606, "context": 0.19077731017023325}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.4625111520290375, "loss_scale": 4.0, "retries": 0, "seconds": 172.0, "max_memory_gb": 16.49, "ce": 0.121204026741907, "kl": 0.017092623485950753, "context": 0.12975033605471253}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.766829252243042, "loss_scale": 4.0, "retries": 0, "seconds": 172.6, "max_memory_gb": 16.49, "ce": 0.05384117329958826, "kl": 0.04197914444375783, "context": 0.07483074470655993}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.736072838306427, "loss_scale": 4.0, "retries": 0, "seconds": 173.1, "max_memory_gb": 16.49, "ce": 0.07857405953109264, "kl": 0.03537280214368366, "context": 0.09626046055927873}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 1.0536741018295288, "loss_scale": 4.0, "retries": 0, "seconds": 173.6, "max_memory_gb": 16.49, "ce": 0.2255657035857439, "kl": 0.016841402044519782, "context": 0.23398640379309654}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.23159939050674438, "loss_scale": 4.0, "retries": 0, "seconds": 174.1, "max_memory_gb": 16.49, "ce": 0.018157577607780695, "kl": 0.02558776765363291, "context": 0.030951460590586066}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.7305313348770142, "loss_scale": 4.0, "retries": 0, "seconds": 174.6, "max_memory_gb": 16.49, "ce": 0.07114991365233436, "kl": 0.01639525347854942, "context": 0.07934753713198006}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.502366542816162, "loss_scale": 4.0, "retries": 0, "seconds": 175.1, "max_memory_gb": 16.49, "ce": 0.08181010105181485, "kl": 0.43052321633149404, "context": 0.29707171814516187}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 1.5390006303787231, "loss_scale": 4.0, "retries": 0, "seconds": 175.6, "max_memory_gb": 16.49, "ce": 0.25619034544797614, "kl": 0.06883549366466468, "context": 0.29060808510985225}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.37943410873413086, "loss_scale": 4.0, "retries": 0, "seconds": 176.0, "max_memory_gb": 16.49, "ce": 0.041875481663737446, "kl": 0.07945625437423587, "context": 0.08160360739566386}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.5119826793670654, "loss_scale": 4.0, "retries": 0, "seconds": 176.5, "max_memory_gb": 16.49, "ce": 0.01601339955232106, "kl": 0.32122773724586295, "context": 0.17662727084825747}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.7804110050201416, "loss_scale": 4.0, "retries": 0, "seconds": 177.0, "max_memory_gb": 16.49, "ce": 0.10931071802042425, "kl": 0.014427903181058355, "context": 0.11652467062231153}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.8057841062545776, "loss_scale": 4.0, "retries": 0, "seconds": 177.6, "max_memory_gb": 16.49, "ce": 0.045945679768919945, "kl": 0.2987727675936185, "context": 0.1953320698812604}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.8671268224716187, "loss_scale": 4.0, "retries": 0, "seconds": 178.0, "max_memory_gb": 16.49, "ce": 0.13555009861011058, "kl": 0.10999448673101142, "context": 0.19054733822122216}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.28418678045272827, "loss_scale": 4.0, "retries": 0, "seconds": 178.5, "max_memory_gb": 16.49, "ce": 0.013465039432048798, "kl": 0.042199197534500854, "context": 0.03456463851034641}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.22921863198280334, "loss_scale": 4.0, "retries": 0, "seconds": 179.0, "max_memory_gb": 16.49, "ce": 0.03755212563555688, "kl": 0.011839259757834952, "context": 0.04347175429575145}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 2.9125888347625732, "loss_scale": 4.0, "retries": 0, "seconds": 179.5, "max_memory_gb": 16.49, "ce": 0.44077774323523045, "kl": 0.16978255985304713, "context": 0.5256690150126815}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 0.8832525014877319, "loss_scale": 4.0, "retries": 0, "seconds": 180.0, "max_memory_gb": 16.49, "ce": 0.1367519911727868, "kl": 0.02536933143346687, "context": 0.14943665504688397}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.2980116307735443, "loss_scale": 4.0, "retries": 0, "seconds": 180.5, "max_memory_gb": 16.49, "ce": 0.037389173870906234, "kl": 0.02126661391957896, "context": 0.04802248231135309}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.47358793020248413, "loss_scale": 4.0, "retries": 0, "seconds": 181.0, "max_memory_gb": 16.49, "ce": 0.08575641631614417, "kl": 0.058348379214294255, "context": 0.11493060924112797}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.9367878437042236, "loss_scale": 4.0, "retries": 0, "seconds": 181.6, "max_memory_gb": 16.49, "ce": 0.10737957793753594, "kl": 0.10874771128510474, "context": 0.1617534263059497}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.7757900357246399, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 16.49, "ce": 0.17309943260625005, "kl": 0.18138494703634933, "context": 0.2637919057160616}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.5782924890518188, "loss_scale": 4.0, "retries": 0, "seconds": 182.6, "max_memory_gb": 16.49, "ce": 0.04838459211168811, "kl": 0.03173373370373156, "context": 0.0642514571081847}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.7385059595108032, "loss_scale": 4.0, "retries": 0, "seconds": 183.0, "max_memory_gb": 16.49, "ce": 0.06339256040519103, "kl": 0.1528736246509652, "context": 0.13982937374385074}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.9476425051689148, "loss_scale": 4.0, "retries": 0, "seconds": 183.6, "max_memory_gb": 16.49, "ce": 0.10893564112484455, "kl": 0.06022617456619628, "context": 0.13904872350394726}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.7207528352737427, "loss_scale": 4.0, "retries": 0, "seconds": 184.0, "max_memory_gb": 16.49, "ce": 0.08343237993540242, "kl": 0.07123379272525199, "context": 0.11904927634168416}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 1.1847320795059204, "loss_scale": 4.0, "retries": 0, "seconds": 184.5, "max_memory_gb": 16.49, "ce": 0.03236173855839297, "kl": 0.17814786215603817, "context": 0.12143567414022982}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.5608683228492737, "loss_scale": 4.0, "retries": 0, "seconds": 185.0, "max_memory_gb": 16.49, "ce": 0.21813952433876693, "kl": 0.08764141146093607, "context": 0.26196023751981556}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 1.4793421030044556, "loss_scale": 4.0, "retries": 0, "seconds": 185.5, "max_memory_gb": 16.49, "ce": 0.32624155044322833, "kl": 0.13632807855901774, "context": 0.39440559333888814}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.25497502088546753, "loss_scale": 4.0, "retries": 0, "seconds": 186.0, "max_memory_gb": 16.49, "ce": 0.06842278013937175, "kl": 0.0740742563175445, "context": 0.10545990813989192}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.16364619135856628, "loss_scale": 4.0, "retries": 0, "seconds": 186.5, "max_memory_gb": 16.49, "ce": 0.010654866811819375, "kl": 0.012574904540088028, "context": 0.01694231911096722}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.696416974067688, "loss_scale": 4.0, "retries": 0, "seconds": 187.0, "max_memory_gb": 16.49, "ce": 0.15597507380880415, "kl": 0.00629232260689605, "context": 0.15912123816087842}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.6749343872070312, "loss_scale": 4.0, "retries": 0, "seconds": 187.5, "max_memory_gb": 16.49, "ce": 0.07492834294680506, "kl": 0.016578517159359762, "context": 0.0832176033873111}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.594086766242981, "loss_scale": 4.0, "retries": 0, "seconds": 188.0, "max_memory_gb": 16.49, "ce": 0.07888243639899883, "kl": 0.014806634644628502, "context": 0.08628575180773623}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.3652600347995758, "loss_scale": 4.0, "retries": 0, "seconds": 188.5, "max_memory_gb": 16.49, "ce": 0.042980199243174866, "kl": 0.0016935548046603799, "context": 0.04382697504479438}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.4486660361289978, "loss_scale": 4.0, "retries": 0, "seconds": 189.0, "max_memory_gb": 16.49, "ce": 0.030865743523463607, "kl": 0.03746070701163262, "context": 0.04959609801881015}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 1.5711352825164795, "loss_scale": 4.0, "retries": 0, "seconds": 189.6, "max_memory_gb": 16.49, "ce": 0.07778302812948823, "kl": 0.07486000203061849, "context": 0.11521302536129951}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 1.2771090269088745, "loss_scale": 4.0, "retries": 0, "seconds": 190.1, "max_memory_gb": 16.49, "ce": 0.17827856540679932, "kl": 0.23766010027611628, "context": 0.29710861667990685}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.04596599563956261, "loss_scale": 4.0, "retries": 0, "seconds": 190.6, "max_memory_gb": 16.49, "ce": 0.0037901546456851065, "kl": 0.0014858587819617242, "context": 0.004533083993010223}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 0.9798190593719482, "loss_scale": 4.0, "retries": 0, "seconds": 191.1, "max_memory_gb": 16.49, "ce": 0.018341832561418414, "kl": 0.6812947988510132, "context": 0.3589892245363444}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.5280826687812805, "loss_scale": 4.0, "retries": 0, "seconds": 191.6, "max_memory_gb": 16.49, "ce": 0.08329934772336856, "kl": 0.020829103072173893, "context": 0.09371389844454825}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.5304405093193054, "loss_scale": 4.0, "retries": 0, "seconds": 192.1, "max_memory_gb": 16.49, "ce": 0.02635419147554785, "kl": 0.005396615728386678, "context": 0.02905249915784225}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 0.760327935218811, "loss_scale": 4.0, "retries": 0, "seconds": 192.5, "max_memory_gb": 16.49, "ce": 0.008172595174983144, "kl": 0.22943735334047233, "context": 0.12289126851828769}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.44168946146965027, "loss_scale": 4.0, "retries": 0, "seconds": 193.0, "max_memory_gb": 16.49, "ce": 0.09531740722013637, "kl": 0.08999131098971702, "context": 0.14031305990647525}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.6878989934921265, "loss_scale": 4.0, "retries": 0, "seconds": 193.5, "max_memory_gb": 16.49, "ce": 0.09776817495003343, "kl": 0.015922960126772523, "context": 0.10572965536266565}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 1.3603590726852417, "loss_scale": 4.0, "retries": 0, "seconds": 194.1, "max_memory_gb": 16.49, "ce": 0.15072755835717544, "kl": 0.0011893063492607325, "context": 0.1513222170760855}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 1.057875394821167, "loss_scale": 4.0, "retries": 0, "seconds": 194.6, "max_memory_gb": 16.49, "ce": 0.03153058874886483, "kl": 0.06557005393551663, "context": 0.06431561743374914}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.4184507131576538, "loss_scale": 4.0, "retries": 0, "seconds": 195.1, "max_memory_gb": 16.49, "ce": 0.019652791670523584, "kl": 0.6564554462383967, "context": 0.34788051480427384}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.5561053156852722, "loss_scale": 4.0, "retries": 0, "seconds": 195.6, "max_memory_gb": 16.49, "ce": 0.12462249712552875, "kl": 0.10066700144670904, "context": 0.1749559931922704}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.09915147721767426, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 16.49, "ce": 0.005548614950384945, "kl": 0.007761346920233336, "context": 0.00942928809672594}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 1.4270285367965698, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 16.49, "ce": 0.18582809146028012, "kl": 0.07212610359420069, "context": 0.22189113911008462}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.6582164168357849, "loss_scale": 4.0, "retries": 0, "seconds": 197.1, "max_memory_gb": 16.49, "ce": 0.03441088681574911, "kl": 0.027469567023217678, "context": 0.048145669512450695}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.5273903012275696, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 16.49, "ce": 0.13217258313670754, "kl": 0.026057305105496198, "context": 0.14520123333204538}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 1.7610093355178833, "loss_scale": 4.0, "retries": 0, "seconds": 198.1, "max_memory_gb": 16.49, "ce": 0.8062670927029103, "kl": 0.43291264545405284, "context": 1.0227234188932925}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 1.3170613050460815, "loss_scale": 4.0, "retries": 0, "seconds": 198.6, "max_memory_gb": 16.49, "ce": 0.2519641261314973, "kl": 0.23082394152879715, "context": 0.36737611307762563}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.508346676826477, "loss_scale": 4.0, "retries": 0, "seconds": 199.1, "max_memory_gb": 16.49, "ce": 0.1587802884168923, "kl": 0.0462292389729555, "context": 0.1818949130247347}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.5733360648155212, "loss_scale": 4.0, "retries": 0, "seconds": 199.6, "max_memory_gb": 16.49, "ce": 0.03552022817893885, "kl": 0.06447325181943597, "context": 0.06775685280445032}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 2.381291627883911, "loss_scale": 4.0, "retries": 0, "seconds": 200.1, "max_memory_gb": 16.49, "ce": 0.17873697204049677, "kl": 0.20217472603690112, "context": 0.2798243350116536}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.2714650630950928, "loss_scale": 4.0, "retries": 0, "seconds": 200.6, "max_memory_gb": 16.49, "ce": 0.13162660598754883, "kl": 0.0796241573698353, "context": 0.17143868084531277}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.8249456882476807, "loss_scale": 4.0, "retries": 0, "seconds": 201.1, "max_memory_gb": 16.49, "ce": 0.04669255414046347, "kl": 0.033723539556376636, "context": 0.06355432607233524}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 1.290130376815796, "loss_scale": 4.0, "retries": 0, "seconds": 201.6, "max_memory_gb": 16.49, "ce": 0.22501388343516737, "kl": 0.015077525833476102, "context": 0.23255264264298603}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 2.025238037109375, "loss_scale": 4.0, "retries": 0, "seconds": 202.1, "max_memory_gb": 16.49, "ce": 0.3035530698252842, "kl": 0.07967262109741569, "context": 0.3433893728069961}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.10515132546424866, "loss_scale": 4.0, "retries": 0, "seconds": 202.6, "max_memory_gb": 16.49, "ce": 0.010447627981193364, "kl": 0.0016992599394143326, "context": 0.011297257791738957}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 1.3723855018615723, "loss_scale": 4.0, "retries": 0, "seconds": 203.1, "max_memory_gb": 16.49, "ce": 0.4061299953609705, "kl": 0.11399221466854215, "context": 0.46312611550092697}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.25422796607017517, "loss_scale": 4.0, "retries": 0, "seconds": 203.6, "max_memory_gb": 16.49, "ce": 0.02171748277032748, "kl": 0.05159246121183969, "context": 0.04751371452584863}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.33442753553390503, "loss_scale": 4.0, "retries": 0, "seconds": 204.0, "max_memory_gb": 16.49, "ce": 0.06423961470136419, "kl": 0.05323586308804806, "context": 0.09085755026899278}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 2.2415318489074707, "loss_scale": 4.0, "retries": 0, "seconds": 204.5, "max_memory_gb": 16.49, "ce": 0.12206798615807202, "kl": 0.936834926949814, "context": 0.5904854508116841}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.3181338608264923, "loss_scale": 4.0, "retries": 0, "seconds": 205.0, "max_memory_gb": 16.49, "ce": 0.03398924367502332, "kl": 0.014176629832945764, "context": 0.041077558998949826}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 0.7608566880226135, "loss_scale": 4.0, "retries": 0, "seconds": 205.5, "max_memory_gb": 16.49, "ce": 0.18168202153174207, "kl": 0.011594099341891706, "context": 0.18747907906072214}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 3.104982852935791, "loss_scale": 4.0, "retries": 0, "seconds": 206.0, "max_memory_gb": 16.49, "ce": 0.4373149902676232, "kl": 0.502677348209545, "context": 0.6886536628007889}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.6390178203582764, "loss_scale": 4.0, "retries": 0, "seconds": 206.5, "max_memory_gb": 16.49, "ce": 0.048424100037664175, "kl": 0.027562725874304306, "context": 0.062205461552366614}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 1.032021403312683, "loss_scale": 4.0, "retries": 0, "seconds": 207.0, "max_memory_gb": 16.49, "ce": 0.0774283220525831, "kl": 0.09504991046560463, "context": 0.12495327170472592}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 1.158995270729065, "loss_scale": 4.0, "retries": 0, "seconds": 207.5, "max_memory_gb": 16.49, "ce": 0.15616168407723308, "kl": 0.05237148480955511, "context": 0.18234742898494005}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.681455135345459, "loss_scale": 4.0, "retries": 0, "seconds": 208.0, "max_memory_gb": 16.49, "ce": 0.05958514945814386, "kl": 0.014664120910310885, "context": 0.06691720974049531}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.7214324474334717, "loss_scale": 4.0, "retries": 0, "seconds": 208.5, "max_memory_gb": 16.49, "ce": 0.006340062216622755, "kl": 0.2149709702407563, "context": 0.11382554983720183}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.7174956202507019, "loss_scale": 4.0, "retries": 0, "seconds": 209.0, "max_memory_gb": 16.49, "ce": 0.011725230840966105, "kl": 0.13174678286304697, "context": 0.07759862032253295}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.6375983953475952, "loss_scale": 4.0, "retries": 0, "seconds": 209.5, "max_memory_gb": 16.49, "ce": 0.09451208630343899, "kl": 0.011381876072846353, "context": 0.10020302346674725}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 2.3519532680511475, "loss_scale": 4.0, "retries": 0, "seconds": 210.0, "max_memory_gb": 16.49, "ce": 0.33101668540621176, "kl": 0.037488499787286855, "context": 0.34976092411670834}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.5932696461677551, "loss_scale": 4.0, "retries": 0, "seconds": 210.5, "max_memory_gb": 16.49, "ce": 0.06620390385796782, "kl": 0.037623023497872055, "context": 0.08501541521400213}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 1.0038344860076904, "loss_scale": 4.0, "retries": 0, "seconds": 211.0, "max_memory_gb": 16.49, "ce": 0.08545886777574196, "kl": 0.11275922483764589, "context": 0.1418384801945649}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.06508470326662064, "loss_scale": 4.0, "retries": 0, "seconds": 211.5, "max_memory_gb": 16.49, "ce": 0.00622883765026927, "kl": 8.784395140537526e-05, "context": 0.006272759579587728}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.8358617424964905, "loss_scale": 4.0, "retries": 0, "seconds": 212.0, "max_memory_gb": 16.49, "ce": 0.12039410404395312, "kl": 0.05252061667488306, "context": 0.14665440958924592}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.8146231770515442, "loss_scale": 4.0, "retries": 0, "seconds": 212.5, "max_memory_gb": 16.49, "ce": 0.09137108316645026, "kl": 0.020881229560473002, "context": 0.10181169933639467}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 1.0180904865264893, "loss_scale": 4.0, "retries": 0, "seconds": 213.0, "max_memory_gb": 16.49, "ce": 0.13794435042655095, "kl": 0.09941816743594245, "context": 0.18765343609265983}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.9454132914543152, "loss_scale": 4.0, "retries": 0, "seconds": 213.5, "max_memory_gb": 16.49, "ce": 0.11335185408825055, "kl": 0.036745303266798146, "context": 0.13172450091224164}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.011070966720581, "loss_scale": 4.0, "retries": 0, "seconds": 214.0, "max_memory_gb": 16.49, "ce": 0.11186662688851357, "kl": 0.19605997088365257, "context": 0.20989660918712616}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.4827899932861328, "loss_scale": 4.0, "retries": 0, "seconds": 214.5, "max_memory_gb": 16.49, "ce": 0.038089237874373794, "kl": 0.03795788255229127, "context": 0.05706817889586091}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 2.5652782917022705, "loss_scale": 4.0, "retries": 0, "seconds": 215.0, "max_memory_gb": 16.49, "ce": 0.19854440621566027, "kl": 0.2649825330154272, "context": 0.3310356783331372}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.36645087599754333, "loss_scale": 4.0, "retries": 0, "seconds": 215.5, "max_memory_gb": 16.49, "ce": 0.06791363330557942, "kl": 0.02027785812970251, "context": 0.07805256196297705}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.41453906893730164, "loss_scale": 4.0, "retries": 0, "seconds": 216.0, "max_memory_gb": 16.49, "ce": 0.059781525284051895, "kl": 0.02344882263423642, "context": 0.07150593516416848}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.5092474222183228, "loss_scale": 4.0, "retries": 0, "seconds": 216.5, "max_memory_gb": 16.49, "ce": 0.07831852464005351, "kl": 0.02957962383516133, "context": 0.09310833713971078}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 1.5718350410461426, "loss_scale": 4.0, "retries": 0, "seconds": 217.0, "max_memory_gb": 16.49, "ce": 0.385455722163897, "kl": 0.02437201548127632, "context": 0.39764172805007547}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 0.8662590980529785, "loss_scale": 4.0, "retries": 0, "seconds": 217.5, "max_memory_gb": 16.49, "ce": 0.07437431812286377, "kl": 0.09370406634843675, "context": 0.12122635007835925}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.40791916847229004, "loss_scale": 4.0, "retries": 0, "seconds": 218.0, "max_memory_gb": 16.49, "ce": 0.16687980748247355, "kl": 0.014573673840459378, "context": 0.17416664666961879}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 1.9630506038665771, "loss_scale": 4.0, "retries": 0, "seconds": 218.5, "max_memory_gb": 16.49, "ce": 0.14993547461926937, "kl": 0.08446997962892056, "context": 0.19217046536505222}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 1.0592191219329834, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 16.49, "ce": 0.038411087880376726, "kl": 0.014944206985092023, "context": 0.04588319326285273}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.6114047169685364, "loss_scale": 4.0, "retries": 0, "seconds": 219.5, "max_memory_gb": 16.49, "ce": 0.03085145991644822, "kl": 0.038300733663163555, "context": 0.050001826719380915}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.6549032926559448, "loss_scale": 4.0, "retries": 0, "seconds": 220.0, "max_memory_gb": 16.49, "ce": 0.017392916604876518, "kl": 0.08452677773675532, "context": 0.0596563036669977}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.784287691116333, "loss_scale": 4.0, "retries": 0, "seconds": 220.5, "max_memory_gb": 16.49, "ce": 0.14777540881186724, "kl": 0.07567985678906552, "context": 0.18561533570755273}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.5498036742210388, "loss_scale": 4.0, "retries": 0, "seconds": 221.0, "max_memory_gb": 16.49, "ce": 0.04331322479993105, "kl": 0.03664610622217879, "context": 0.06163627956993878}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 1.075693130493164, "loss_scale": 4.0, "retries": 0, "seconds": 221.5, "max_memory_gb": 16.49, "ce": 0.1538822629954666, "kl": 0.012831907701183809, "context": 0.1602982166223228}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 1.0484142303466797, "loss_scale": 4.0, "retries": 0, "seconds": 222.0, "max_memory_gb": 16.49, "ce": 0.08992561063496396, "kl": 0.1687875498610083, "context": 0.17431938694790006}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.28444477915763855, "loss_scale": 4.0, "retries": 0, "seconds": 222.5, "max_memory_gb": 16.49, "ce": 0.01828141207806766, "kl": 0.01229853177210316, "context": 0.024430678342469037}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.20800994336605072, "loss_scale": 4.0, "retries": 0, "seconds": 223.0, "max_memory_gb": 16.49, "ce": 0.018825757841113955, "kl": 0.01369303789397236, "context": 0.025672276911791414}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 1.1271008253097534, "loss_scale": 4.0, "retries": 0, "seconds": 223.5, "max_memory_gb": 16.49, "ce": 0.14269558887463063, "kl": 0.11858455759647768, "context": 0.2019878727151081}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 2.8026928901672363, "loss_scale": 4.0, "retries": 0, "seconds": 224.0, "max_memory_gb": 16.49, "ce": 0.16159475583117455, "kl": 0.10976007782301167, "context": 0.21647479332750663}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.19526933133602142, "loss_scale": 4.0, "retries": 0, "seconds": 224.5, "max_memory_gb": 16.49, "ce": 0.01312590594170615, "kl": 0.019177727659553057, "context": 0.022714769816957414}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 1.2361997365951538, "loss_scale": 4.0, "retries": 0, "seconds": 225.0, "max_memory_gb": 16.49, "ce": 0.21444687814801, "kl": 0.12556244638835778, "context": 0.27722810284467414}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 1.036210536956787, "loss_scale": 4.0, "retries": 0, "seconds": 225.5, "max_memory_gb": 16.49, "ce": 0.05730979167856276, "kl": 0.006257811463001417, "context": 0.06043869644054212}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.35558345913887024, "loss_scale": 4.0, "retries": 0, "seconds": 226.0, "max_memory_gb": 16.49, "ce": 0.024645403609611094, "kl": 0.0021480906470969785, "context": 0.02571944845840335}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.06466732919216156, "loss_scale": 4.0, "retries": 0, "seconds": 226.5, "max_memory_gb": 16.49, "ce": 0.005018477269914001, "kl": 0.001714270685624797, "context": 0.005875612609088421}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 1.6569726467132568, "loss_scale": 4.0, "retries": 0, "seconds": 227.0, "max_memory_gb": 16.49, "ce": 0.012997024838114157, "kl": 0.5803377806987555, "context": 0.3031659114349168}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 0.9367495179176331, "loss_scale": 4.0, "retries": 0, "seconds": 227.5, "max_memory_gb": 16.49, "ce": 0.004510981380008161, "kl": 0.30985996673916816, "context": 0.15944096173916478}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.5513280034065247, "loss_scale": 4.0, "retries": 0, "seconds": 228.0, "max_memory_gb": 16.49, "ce": 0.040250983904115856, "kl": 0.02307394645686145, "context": 0.05178795714164153}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.3424277603626251, "loss_scale": 4.0, "retries": 0, "seconds": 228.5, "max_memory_gb": 16.49, "ce": 0.02556964772520587, "kl": 0.01134898476266244, "context": 0.031244139769114554}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 0.7536520957946777, "loss_scale": 4.0, "retries": 0, "seconds": 229.0, "max_memory_gb": 16.49, "ce": 0.0848900496494025, "kl": 0.04019953281385824, "context": 0.10498981579439715}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.939456820487976, "loss_scale": 4.0, "retries": 0, "seconds": 229.5, "max_memory_gb": 16.49, "ce": 0.0028154020837973803, "kl": 0.3125286420035991, "context": 0.1590797167737037}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.3967280387878418, "loss_scale": 4.0, "retries": 0, "seconds": 230.0, "max_memory_gb": 16.49, "ce": 0.1428936968150083, "kl": 0.15818227121053496, "context": 0.22198483324609697}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.1386961042881012, "loss_scale": 4.0, "retries": 0, "seconds": 230.5, "max_memory_gb": 16.49, "ce": 0.007677843561396003, "kl": 0.009169385688437615, "context": 0.012262536212801933}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.31491175293922424, "loss_scale": 4.0, "retries": 0, "seconds": 231.0, "max_memory_gb": 16.49, "ce": 0.027913129481021315, "kl": 0.008198791118047666, "context": 0.03201252481085248}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.6331847906112671, "loss_scale": 4.0, "retries": 0, "seconds": 231.5, "max_memory_gb": 16.49, "ce": 0.07591426197905093, "kl": 0.011528277544130106, "context": 0.08167840301757678}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 1.515465497970581, "loss_scale": 4.0, "retries": 0, "seconds": 232.0, "max_memory_gb": 16.49, "ce": 0.19229652822832577, "kl": 0.007637175960553577, "context": 0.1961151118739508}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.28718551993370056, "loss_scale": 4.0, "retries": 0, "seconds": 232.5, "max_memory_gb": 16.49, "ce": 0.021066403103759512, "kl": 0.00634770406759344, "context": 0.024240254948381335}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 2.5448529720306396, "loss_scale": 4.0, "retries": 0, "seconds": 233.0, "max_memory_gb": 16.49, "ce": 0.006092538998927921, "kl": 0.37853794529655715, "context": 0.19536151541979052}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.6835139393806458, "loss_scale": 4.0, "retries": 0, "seconds": 233.5, "max_memory_gb": 16.49, "ce": 0.020045624289195985, "kl": 0.0643134220418915, "context": 0.052202334991307}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 2.3936238288879395, "loss_scale": 4.0, "retries": 0, "seconds": 234.0, "max_memory_gb": 16.49, "ce": 0.4063973451266065, "kl": 0.30655079724965617, "context": 0.5596727444790304}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.7622284293174744, "loss_scale": 4.0, "retries": 0, "seconds": 234.5, "max_memory_gb": 16.49, "ce": 0.14898826892022043, "kl": 0.011035690818971489, "context": 0.15450611244887114}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.22306621074676514, "loss_scale": 4.0, "retries": 0, "seconds": 235.0, "max_memory_gb": 16.49, "ce": 0.007186856470070779, "kl": 0.020300662392401136, "context": 0.01733718765899539}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.7839047312736511, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 16.49, "ce": 0.09072050836402923, "kl": 0.009215705562382936, "context": 0.09532836498692632}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 1.595866084098816, "loss_scale": 4.0, "retries": 0, "seconds": 236.0, "max_memory_gb": 16.49, "ce": 0.1265816020895727, "kl": 0.5616858899593353, "context": 0.40742454677820206}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 1.693346619606018, "loss_scale": 4.0, "retries": 0, "seconds": 236.5, "max_memory_gb": 16.49, "ce": 0.48454155470244586, "kl": 0.14563190890476108, "context": 0.55735749960877}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 1.765744924545288, "loss_scale": 4.0, "retries": 0, "seconds": 237.0, "max_memory_gb": 16.49, "ce": 0.09977058786898851, "kl": 0.087384813639801, "context": 0.14346299530006945}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 1.0720288753509521, "loss_scale": 4.0, "retries": 0, "seconds": 237.4, "max_memory_gb": 16.49, "ce": 0.14337308605900034, "kl": 0.01880107566830702, "context": 0.15277362911729142}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 0.9947617650032043, "loss_scale": 4.0, "retries": 0, "seconds": 237.9, "max_memory_gb": 16.49, "ce": 0.05329651336069219, "kl": 0.05544502240445581, "context": 0.08101902838097885}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.4088543951511383, "loss_scale": 4.0, "retries": 0, "seconds": 238.4, "max_memory_gb": 16.49, "ce": 0.09086704946821555, "kl": 0.017913389230670873, "context": 0.09982374537503347}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 1.6079622507095337, "loss_scale": 4.0, "retries": 0, "seconds": 238.9, "max_memory_gb": 16.49, "ce": 0.26935403794050217, "kl": 0.093525905336719, "context": 0.3161169817321934}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.3311980068683624, "loss_scale": 4.0, "retries": 0, "seconds": 239.4, "max_memory_gb": 16.49, "ce": 0.02650826922035776, "kl": 0.015413236524182139, "context": 0.03421488724416122}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.6842026114463806, "loss_scale": 4.0, "retries": 0, "seconds": 239.9, "max_memory_gb": 16.49, "ce": 0.06584891022066586, "kl": 0.18534271728640306, "context": 0.15852026562788524}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.1611093282699585, "loss_scale": 4.0, "retries": 0, "seconds": 240.4, "max_memory_gb": 16.49, "ce": 0.010934894060483202, "kl": 0.0032551281774431118, "context": 0.012562458126922138}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 1.0357286930084229, "loss_scale": 4.0, "retries": 0, "seconds": 240.9, "max_memory_gb": 16.49, "ce": 0.06829768093302846, "kl": 0.028780795459169894, "context": 0.08268807549029589}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.2560617923736572, "loss_scale": 4.0, "retries": 0, "seconds": 241.4, "max_memory_gb": 16.49, "ce": 0.008286280673928559, "kl": 0.6034847759110562, "context": 0.3100286790577229}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.7578326463699341, "loss_scale": 4.0, "retries": 0, "seconds": 241.9, "max_memory_gb": 16.49, "ce": 0.06196598941460252, "kl": 0.008591383491875604, "context": 0.06626168265938759}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.3049182891845703, "loss_scale": 4.0, "retries": 0, "seconds": 242.4, "max_memory_gb": 16.49, "ce": 0.019928029942093417, "kl": 0.020320162584539503, "context": 0.03008811129257083}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 1.2418839931488037, "loss_scale": 4.0, "retries": 0, "seconds": 242.9, "max_memory_gb": 16.49, "ce": 0.15955507461330853, "kl": 0.01298761455109343, "context": 0.16604888782603666}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.6934264302253723, "loss_scale": 4.0, "retries": 0, "seconds": 243.4, "max_memory_gb": 16.49, "ce": 0.11099817813374102, "kl": 0.04239760831114836, "context": 0.1321969794225879}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 1.3088105916976929, "loss_scale": 4.0, "retries": 0, "seconds": 243.9, "max_memory_gb": 16.49, "ce": 0.12721355422399938, "kl": 0.0220917787446524, "context": 0.1382594403985422}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 1.804886817932129, "loss_scale": 4.0, "retries": 0, "seconds": 244.4, "max_memory_gb": 16.49, "ce": 0.27492136077489704, "kl": 0.013241279637441039, "context": 0.2815420137485489}
diff --git a/runs/single_llava/context/L18_s13/adapter/README.md b/runs/single_llava/context/L18_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L18_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L18_s13/adapter/adapter_config.json b/runs/single_llava/context/L18_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..43a58dce9e6195e48d49ffb7c2f515b49114cf78
--- /dev/null
+++ b/runs/single_llava/context/L18_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.18.mlp.up_proj",
+    "model.language_model.layers.18.mlp.gate_proj",
+    "model.language_model.layers.18.self_attn.o_proj",
+    "model.language_model.layers.18.self_attn.k_proj",
+    "model.language_model.layers.18.self_attn.v_proj",
+    "model.language_model.layers.18.self_attn.q_proj",
+    "model.language_model.layers.18.mlp.down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L18_s13/config.json b/runs/single_llava/context/L18_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..dd7ed210f2e751af5fa66fdc2993189c95e16c00
--- /dev/null
+++ b/runs/single_llava/context/L18_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 18,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L18_s13/steps.jsonl b/runs/single_llava/context/L18_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5428631776dcf79accf8f8a01b2674fba6c0b243
--- /dev/null
+++ b/runs/single_llava/context/L18_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.2289048582315445, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 15.91, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.29885566234588623, "loss_scale": 4.0, "retries": 0, "seconds": 1.9, "max_memory_gb": 15.92, "ce": 2.681319466792047, "kl": 0.010371771204518154, "context": 2.686505377292633}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.6648598313331604, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 15.92, "ce": 5.3879687786102295, "kl": 0.17893695785460295, "context": 5.477437317371368}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.4840867221355438, "loss_scale": 4.0, "retries": 0, "seconds": 2.9, "max_memory_gb": 15.92, "ce": 3.82074888702482, "kl": 0.051233461239462486, "context": 3.8463655756786466}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.0559731051325798, "loss_scale": 4.0, "retries": 0, "seconds": 3.4, "max_memory_gb": 15.92, "ce": 0.08584192162379622, "kl": 0.03710726834833622, "context": 0.10439555533230305}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.22849850356578827, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 15.92, "ce": 1.4654138423502445, "kl": 0.18369842553511262, "context": 1.557263009250164}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.4660804271697998, "loss_scale": 4.0, "retries": 0, "seconds": 4.4, "max_memory_gb": 15.92, "ce": 3.4368390180170536, "kl": 0.10629439603508217, "context": 3.489986272994429}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.5578485727310181, "loss_scale": 4.0, "retries": 0, "seconds": 4.9, "max_memory_gb": 15.92, "ce": 3.689108692109585, "kl": 0.16095312125980854, "context": 3.769585356116295}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.45375022292137146, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 15.92, "ce": 2.8356877379119396, "kl": 0.16256996907759458, "context": 2.9169727489352226}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.30080872774124146, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 15.92, "ce": 1.628368929028511, "kl": 0.05988646048353985, "context": 1.6583121046423912}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.5031123161315918, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 15.92, "ce": 2.6207335321232677, "kl": 0.009467968455282971, "context": 2.625467559322715}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.3264741897583008, "loss_scale": 4.0, "retries": 0, "seconds": 6.9, "max_memory_gb": 15.92, "ce": 1.5301275979727507, "kl": 0.017919285863172263, "context": 1.539087243610993}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.5745254755020142, "loss_scale": 4.0, "retries": 0, "seconds": 7.4, "max_memory_gb": 15.92, "ce": 2.660954928956926, "kl": 0.007509707196732052, "context": 2.664709777571261}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.8201872706413269, "loss_scale": 4.0, "retries": 0, "seconds": 7.9, "max_memory_gb": 15.92, "ce": 3.8780581802129745, "kl": 0.11845680070109665, "context": 3.9372866600751877}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.5845848917961121, "loss_scale": 4.0, "retries": 0, "seconds": 8.4, "max_memory_gb": 15.92, "ce": 2.575086048687808, "kl": 0.26766588509781286, "context": 2.7089190662372857}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.35244619846343994, "loss_scale": 4.0, "retries": 0, "seconds": 8.9, "max_memory_gb": 15.92, "ce": 1.366654270561412, "kl": 0.009370076237246394, "context": 1.3713393204379827}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.3644045293331146, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 15.92, "ce": 1.5840982901863754, "kl": 0.07801306212786585, "context": 1.6231048591434956}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.8619370460510254, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 15.93, "ce": 3.543769933283329, "kl": 0.09384382772259414, "context": 3.590691827237606}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.4660577774047852, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 15.93, "ce": 5.285069942474365, "kl": 0.33353153977077454, "context": 5.451835751533508}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.01727542281150818, "loss_scale": 4.0, "retries": 0, "seconds": 10.9, "max_memory_gb": 15.93, "ce": 0.012969426810741425, "kl": 0.002192620995629113, "context": 0.014065737253986299}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.8881804943084717, "loss_scale": 4.0, "retries": 0, "seconds": 11.4, "max_memory_gb": 15.93, "ce": 2.7382389495614916, "kl": 0.030068486266827676, "context": 2.753273298498243}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.8931665420532227, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 15.93, "ce": 2.900220163166523, "kl": 0.05451986426487565, "context": 2.9274800904095173}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.9820845723152161, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 15.93, "ce": 2.6096114311367273, "kl": 0.5089551696437411, "context": 2.8640889022499323}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.553464651107788, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 15.93, "ce": 4.942479491233826, "kl": 0.01442680845502764, "context": 4.949692904949188}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.464390754699707, "loss_scale": 4.0, "retries": 0, "seconds": 13.5, "max_memory_gb": 15.93, "ce": 4.167667560279369, "kl": 0.03562981972936541, "context": 4.185482442378998}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.9495964050292969, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 15.93, "ce": 5.122886061668396, "kl": 0.06452279724180698, "context": 5.155147433280945}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.6238306760787964, "loss_scale": 4.0, "retries": 0, "seconds": 14.5, "max_memory_gb": 15.93, "ce": 4.095782320946455, "kl": 0.08338667382486165, "context": 4.137475706636906}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.2333581447601318, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 15.93, "ce": 2.723148211836815, "kl": 0.23368831351399422, "context": 2.8399922847747803}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.1244584321975708, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 15.93, "ce": 2.547336501535028, "kl": 0.5170740120811388, "context": 2.8058735611848533}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.2929848432540894, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 15.93, "ce": 2.6671854741871357, "kl": 0.037543865269981325, "context": 2.6859574485570192}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.3650761842727661, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 15.93, "ce": 2.7630454674363136, "kl": 0.027422876562923193, "context": 2.776756837964058}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.604830741882324, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 15.93, "ce": 5.136703848838806, "kl": 0.015464514028280973, "context": 5.1444361209869385}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.2559922933578491, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 15.93, "ce": 2.3773223282769322, "kl": 0.17854234063997865, "context": 2.466593423858285}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.4681117534637451, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 15.93, "ce": 2.52056257519871, "kl": 0.6183339115232229, "context": 2.829729560762644}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.6423381567001343, "loss_scale": 4.0, "retries": 0, "seconds": 18.5, "max_memory_gb": 15.93, "ce": 2.58612205600366, "kl": 0.029206687278929166, "context": 2.6007254156284034}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.3920037746429443, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 15.93, "ce": 2.451843351125717, "kl": 0.006816849228926003, "context": 2.4552517645061016}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.6799383163452148, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 15.93, "ce": 2.34669878333807, "kl": 0.3397455927915871, "context": 2.516571555286646}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.0307676792144775, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 15.93, "ce": 4.4205485582351685, "kl": 0.48488283160259016, "context": 4.662990093231201}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.0974570512771606, "loss_scale": 4.0, "retries": 0, "seconds": 20.5, "max_memory_gb": 15.93, "ce": 1.3465993530116975, "kl": 0.03401026129722595, "context": 1.3636045237071812}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.8311392068862915, "loss_scale": 4.0, "retries": 0, "seconds": 21.0, "max_memory_gb": 15.93, "ce": 2.5635761842131615, "kl": 0.11775587173178792, "context": 2.62245412170887}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.9048670530319214, "loss_scale": 4.0, "retries": 0, "seconds": 21.5, "max_memory_gb": 15.93, "ce": 1.391946479678154, "kl": 0.17209304263815284, "context": 1.4779929593205452}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.150712013244629, "loss_scale": 4.0, "retries": 0, "seconds": 22.1, "max_memory_gb": 15.93, "ce": 2.1519326185807586, "kl": 0.007051996537484229, "context": 2.1554586132988334}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.095949411392212, "loss_scale": 4.0, "retries": 0, "seconds": 22.6, "max_memory_gb": 15.93, "ce": 2.2378248060122132, "kl": 0.012019701520330273, "context": 2.2438346976414323}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.1390985250473022, "loss_scale": 4.0, "retries": 0, "seconds": 23.1, "max_memory_gb": 15.93, "ce": 1.271413211710751, "kl": 0.26216305408161134, "context": 1.4024947080761194}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.1439303159713745, "loss_scale": 4.0, "retries": 0, "seconds": 23.6, "max_memory_gb": 15.93, "ce": 1.3027715384960175, "kl": 0.055821881629526615, "context": 1.3306825198233128}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.328346014022827, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 15.93, "ce": 2.051923505961895, "kl": 0.07369556694175117, "context": 2.088771313428879}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.0483126640319824, "loss_scale": 4.0, "retries": 0, "seconds": 24.6, "max_memory_gb": 15.93, "ce": 1.82555618067272, "kl": 0.0348472565819975, "context": 1.8429798339493573}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 4.584033489227295, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 15.93, "ce": 3.2659099847078323, "kl": 0.016698144259862602, "context": 3.2742590457201004}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.691105842590332, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 15.93, "ce": 3.168057758361101, "kl": 0.09852081816643476, "context": 3.2173181157559156}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 4.051675796508789, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 15.93, "ce": 3.177700053434819, "kl": 0.051687810104340315, "context": 3.2035438707098365}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.3515604734420776, "loss_scale": 4.0, "retries": 0, "seconds": 26.6, "max_memory_gb": 15.93, "ce": 1.2628681333735585, "kl": 0.1288260251749307, "context": 1.3272811784408987}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 2.8710668087005615, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 15.93, "ce": 2.014474853873253, "kl": 0.1175939048989676, "context": 2.0732718035578728}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 4.332571029663086, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 15.93, "ce": 3.267915055155754, "kl": 0.18615612713620067, "context": 3.3609929978847504}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 5.010819435119629, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 15.93, "ce": 3.226568654179573, "kl": 0.04321283326135017, "context": 3.24817506223917}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 3.2851247787475586, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 15.93, "ce": 2.0455338270403445, "kl": 0.018500431986467447, "context": 2.054784053005278}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 7.002745628356934, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 15.93, "ce": 3.9189800024032593, "kl": 0.31917549355421215, "context": 4.078567683696747}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.6981077194213867, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 15.93, "ce": 1.0230817617848516, "kl": 0.10461425827816129, "context": 1.0753888664767146}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 4.341963291168213, "loss_scale": 4.0, "retries": 0, "seconds": 30.1, "max_memory_gb": 15.93, "ce": 1.815660054096952, "kl": 0.004412879254232394, "context": 1.8178664990700781}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.119624137878418, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 15.93, "ce": 1.224443031474948, "kl": 0.07442378159612417, "context": 1.2616549525409937}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 3.9720988273620605, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 15.93, "ce": 1.927640228997916, "kl": 0.013651011657202616, "context": 1.934465709142387}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 6.83502721786499, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 15.93, "ce": 2.507977068424225, "kl": 0.06343573905178346, "context": 2.5396948978304863}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 2.349813222885132, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 15.93, "ce": 1.0309702344238758, "kl": 0.03931682091206312, "context": 1.050628650933504}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 6.160894870758057, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 15.93, "ce": 2.952483043074608, "kl": 0.08717575281116297, "context": 2.9960708715952933}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 8.501119613647461, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 15.95, "ce": 2.456934377551079, "kl": 0.03448812710121274, "context": 2.474178444594145}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 7.585987091064453, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 15.95, "ce": 2.6105261892080307, "kl": 0.13915050809737295, "context": 2.680101439356804}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.542790174484253, "loss_scale": 4.0, "retries": 0, "seconds": 34.2, "max_memory_gb": 15.95, "ce": 0.7288763786200434, "kl": 0.061594690050696954, "context": 0.7596737366402522}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 8.067464828491211, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 15.95, "ce": 2.037448510527611, "kl": 0.018565469654276967, "context": 2.046731263399124}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 3.3522799015045166, "loss_scale": 4.0, "retries": 0, "seconds": 35.2, "max_memory_gb": 15.95, "ce": 0.77453528996557, "kl": 0.044760373421013355, "context": 0.7969154892489314}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 5.962077617645264, "loss_scale": 4.0, "retries": 0, "seconds": 35.7, "max_memory_gb": 15.95, "ce": 1.3356899097561836, "kl": 0.013024714717175812, "context": 1.342202213127166}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 5.943675994873047, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 15.95, "ce": 1.3212600843980908, "kl": 0.004111645059310831, "context": 1.3233158886432648}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 9.60506534576416, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 15.95, "ce": 1.8691497053951025, "kl": 0.011920221091713756, "context": 1.875109827145934}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 11.888760566711426, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 15.95, "ce": 2.0533522367477417, "kl": 0.14478776534087956, "context": 2.1257461309432983}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.2033577263355255, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 15.95, "ce": 0.22542471159249544, "kl": 0.03542906022630632, "context": 0.24313923716545105}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 6.918008804321289, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 15.95, "ce": 1.014093229547143, "kl": 0.028132823790656403, "context": 1.0281596621498466}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.539700984954834, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 15.95, "ce": 0.9759313352406025, "kl": 0.1959832590073347, "context": 1.073922947049141}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 6.781098365783691, "loss_scale": 4.0, "retries": 0, "seconds": 39.3, "max_memory_gb": 15.95, "ce": 0.8495253166183829, "kl": 0.4500753190368414, "context": 1.0745629332959652}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 10.025630950927734, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 15.95, "ce": 1.3842331767082214, "kl": 0.03879110049456358, "context": 1.4036287069320679}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 5.252883434295654, "loss_scale": 4.0, "retries": 0, "seconds": 40.3, "max_memory_gb": 15.95, "ce": 0.6870924457907677, "kl": 0.07238277612486854, "context": 0.7232838366180658}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 3.1133041381835938, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 15.95, "ce": 0.3287651687860489, "kl": 0.07422769181903277, "context": 0.3658790267072618}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 5.925323486328125, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 15.95, "ce": 0.5058364283759147, "kl": 0.008880230045178905, "context": 0.5102765656774864}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 8.770895957946777, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 15.95, "ce": 0.7272735312581062, "kl": 0.05279174796305597, "context": 0.7536693960428238}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 8.517799377441406, "loss_scale": 4.0, "retries": 0, "seconds": 42.2, "max_memory_gb": 15.95, "ce": 0.6531839817762375, "kl": 0.054755852557718754, "context": 0.6805619224905968}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 7.303871154785156, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 15.95, "ce": 0.6300693433731794, "kl": 0.19765512831509113, "context": 0.7288969196379185}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 3.3659698963165283, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 15.95, "ce": 0.3741390004288405, "kl": 0.07395021661068313, "context": 0.411114102229476}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.15532630681991577, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 15.95, "ce": 0.07112242048606277, "kl": 0.03570988780120388, "context": 0.08897736482322216}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 7.536762237548828, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 15.95, "ce": 0.535781916230917, "kl": 0.2135336077772081, "context": 0.6425487212836742}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 5.500988006591797, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 15.95, "ce": 0.3241260312497616, "kl": 0.06405378319323063, "context": 0.35615292377769947}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.3638370037078857, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 15.95, "ce": 0.29786637844517827, "kl": 0.022899486968526617, "context": 0.3093161128927022}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.09465263038873672, "loss_scale": 4.0, "retries": 0, "seconds": 45.8, "max_memory_gb": 15.95, "ce": 0.11183747823815793, "kl": 0.005711955607694108, "context": 0.11469345260411501}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.175581455230713, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 15.95, "ce": 0.24144969321787357, "kl": 0.019067337037995458, "context": 0.25098336208611727}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.25094714760780334, "loss_scale": 4.0, "retries": 0, "seconds": 46.8, "max_memory_gb": 15.95, "ce": 0.33079550229012966, "kl": 0.09105009259656072, "context": 0.3763205464929342}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 0.8943642973899841, "loss_scale": 4.0, "retries": 0, "seconds": 47.3, "max_memory_gb": 15.95, "ce": 0.13696896599140018, "kl": 0.07897844025865197, "context": 0.17645818460732698}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 2.1083834171295166, "loss_scale": 4.0, "retries": 0, "seconds": 47.8, "max_memory_gb": 15.95, "ce": 0.14492696220986545, "kl": 0.05959126411471516, "context": 0.17472259094938636}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 2.1303861141204834, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 15.95, "ce": 0.18535771779716015, "kl": 0.10486783273518085, "context": 0.23779163137078285}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 1.1401057243347168, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 15.95, "ce": 0.22464904561638832, "kl": 0.06451594131067395, "context": 0.25690701976418495}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 1.2123892307281494, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 15.95, "ce": 0.16471758857369423, "kl": 0.0347407769295387, "context": 0.18208797369152308}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 0.5409706234931946, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 15.95, "ce": 0.19083978608250618, "kl": 0.022835307056084275, "context": 0.202257439494133}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 0.9333552718162537, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 15.95, "ce": 0.05902717658318579, "kl": 0.06661674451970612, "context": 0.09233554592356086}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.4343056380748749, "loss_scale": 4.0, "retries": 0, "seconds": 50.8, "max_memory_gb": 15.95, "ce": 0.22559611545875669, "kl": 0.06568761880043894, "context": 0.25843992666341364}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 1.031822919845581, "loss_scale": 4.0, "retries": 0, "seconds": 51.3, "max_memory_gb": 15.95, "ce": 0.17624583467841148, "kl": 0.10808031521446537, "context": 0.23028599424287677}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 0.791434645652771, "loss_scale": 4.0, "retries": 0, "seconds": 51.8, "max_memory_gb": 15.95, "ce": 0.10279706306755543, "kl": 0.3817205724772066, "context": 0.29365735687315464}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 1.3172218799591064, "loss_scale": 4.0, "retries": 0, "seconds": 52.3, "max_memory_gb": 15.95, "ce": 0.3695416608825326, "kl": 0.1414281353354454, "context": 0.440255731344223}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 0.6505367159843445, "loss_scale": 4.0, "retries": 0, "seconds": 52.8, "max_memory_gb": 15.95, "ce": 0.052888025413267314, "kl": 0.1190703350584954, "context": 0.11242319364100695}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 0.3146333694458008, "loss_scale": 4.0, "retries": 0, "seconds": 53.3, "max_memory_gb": 15.95, "ce": 0.07204920798540115, "kl": 0.039630746585316956, "context": 0.09186458156909794}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 0.624392569065094, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 15.95, "ce": 0.10252712480723858, "kl": 0.18612177789327689, "context": 0.19558801921084523}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 0.31108760833740234, "loss_scale": 4.0, "retries": 0, "seconds": 54.3, "max_memory_gb": 15.95, "ce": 0.10075286217033863, "kl": 0.04802421713247895, "context": 0.12476497143507004}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.21664153039455414, "loss_scale": 4.0, "retries": 0, "seconds": 54.8, "max_memory_gb": 15.95, "ce": 0.12930400390177965, "kl": 0.0151501742657274, "context": 0.1368790902197361}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.3628832697868347, "loss_scale": 4.0, "retries": 0, "seconds": 55.3, "max_memory_gb": 15.95, "ce": 0.14367753127589822, "kl": 0.023963449464645237, "context": 0.15565926092676818}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 0.8253493905067444, "loss_scale": 4.0, "retries": 0, "seconds": 55.8, "max_memory_gb": 15.95, "ce": 0.17769026011228561, "kl": 0.3841632570256479, "context": 0.36977188289165497}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 0.5281094908714294, "loss_scale": 4.0, "retries": 0, "seconds": 56.3, "max_memory_gb": 15.95, "ce": 0.173001516610384, "kl": 0.052293493354227394, "context": 0.19914826191961765}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 0.24773363769054413, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 15.95, "ce": 0.04846229404211044, "kl": 0.07258149611880071, "context": 0.08475303975865245}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 0.35981079936027527, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 15.95, "ce": 0.10286203399300575, "kl": 0.01587900283630006, "context": 0.11080153682269156}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 0.3717632591724396, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 15.95, "ce": 0.12641913071274757, "kl": 0.04163222888018936, "context": 0.1472352489363402}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 0.8483380675315857, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 15.95, "ce": 0.22480953112244606, "kl": 0.7465536418894771, "context": 0.5980863720178604}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.6290879845619202, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 15.95, "ce": 0.18665440334007144, "kl": 0.2571871739346534, "context": 0.3152479901909828}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.1131637766957283, "loss_scale": 4.0, "retries": 0, "seconds": 59.3, "max_memory_gb": 15.95, "ce": 0.11887983232736588, "kl": 0.005701563481125049, "context": 0.12173061352223158}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 0.5266680717468262, "loss_scale": 4.0, "retries": 0, "seconds": 59.8, "max_memory_gb": 15.95, "ce": 0.2344009152147919, "kl": 0.013220371503848583, "context": 0.24101110617630184}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.12287668138742447, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 15.95, "ce": 0.14544832054525614, "kl": 0.04468738161813235, "context": 0.1677920175716281}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.34640371799468994, "loss_scale": 4.0, "retries": 0, "seconds": 60.8, "max_memory_gb": 15.95, "ce": 0.11035636533051729, "kl": 0.06360268825665116, "context": 0.1421577138826251}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 0.5346861481666565, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 15.95, "ce": 0.199596188031137, "kl": 0.06406330456957221, "context": 0.2316278386861086}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 0.5262724161148071, "loss_scale": 4.0, "retries": 0, "seconds": 61.9, "max_memory_gb": 15.95, "ce": 0.24716049898415804, "kl": 0.07869229470088612, "context": 0.28650665236636996}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 0.15411405265331268, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 15.95, "ce": 0.12893410422839224, "kl": 0.013171467173378915, "context": 0.1355198377277702}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.064426064491272, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 15.95, "ce": 0.4851493835449219, "kl": 0.06610039749648422, "context": 0.5181995807215571}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 0.30609920620918274, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 15.95, "ce": 0.2396172070875764, "kl": 0.031866772391367704, "context": 0.2555505963973701}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.18323516845703125, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 15.95, "ce": 0.13353035040199757, "kl": 0.02124582350370474, "context": 0.14415326016023755}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 0.5320181846618652, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 15.95, "ce": 0.29334274120628834, "kl": 0.08664787735324353, "context": 0.33666668739169836}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.15432575345039368, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 15.95, "ce": 0.04328009602613747, "kl": 0.004086121543878107, "context": 0.045323156053200364}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.0185317993164062, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 15.95, "ce": 0.09735971922054887, "kl": 0.41570050273730885, "context": 0.3052099756896496}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 0.290465772151947, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 15.95, "ce": 0.04090350866317749, "kl": 0.09740956019959413, "context": 0.08960828511044383}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 0.9101408123970032, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 15.95, "ce": 0.275765108410269, "kl": 0.4418484982161317, "context": 0.4966893792152405}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 0.47455811500549316, "loss_scale": 4.0, "retries": 0, "seconds": 66.9, "max_memory_gb": 15.95, "ce": 0.22985175251960754, "kl": 0.04364005080424249, "context": 0.25167177617549896}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.6241469383239746, "loss_scale": 4.0, "retries": 0, "seconds": 67.4, "max_memory_gb": 15.95, "ce": 0.340960732428357, "kl": 0.2702257781638764, "context": 0.4760736250318587}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 0.6108813881874084, "loss_scale": 4.0, "retries": 0, "seconds": 67.9, "max_memory_gb": 15.95, "ce": 0.16411346849054098, "kl": 0.1313850819133222, "context": 0.22980600781738758}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.3690890073776245, "loss_scale": 4.0, "retries": 0, "seconds": 68.4, "max_memory_gb": 15.95, "ce": 0.18537491979077458, "kl": 0.014056198648177087, "context": 0.19240302313119173}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 0.33758512139320374, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 15.95, "ce": 0.1118977153673768, "kl": 0.022163369052577764, "context": 0.12297939788550138}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.40371352434158325, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 15.95, "ce": 0.16236933320760727, "kl": 0.03845118288882077, "context": 0.18159492500126362}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.19344103336334229, "loss_scale": 4.0, "retries": 0, "seconds": 70.0, "max_memory_gb": 15.95, "ce": 0.1132649127393961, "kl": 0.021373747978941537, "context": 0.12395178840961307}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.3856547474861145, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 15.95, "ce": 0.1100397352129221, "kl": 0.06156482652295381, "context": 0.14082215074449778}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 0.8977702260017395, "loss_scale": 4.0, "retries": 0, "seconds": 71.0, "max_memory_gb": 15.95, "ce": 0.1519895438104868, "kl": 0.13590105343610048, "context": 0.21994007006287575}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.28942519426345825, "loss_scale": 4.0, "retries": 0, "seconds": 71.5, "max_memory_gb": 15.95, "ce": 0.07135349931195378, "kl": 0.019969028857303783, "context": 0.08133801119402051}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.3133666515350342, "loss_scale": 4.0, "retries": 0, "seconds": 72.1, "max_memory_gb": 15.95, "ce": 0.14441746985539794, "kl": 0.012850865488871932, "context": 0.15084289945662022}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.38360995054244995, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 15.95, "ce": 0.14499383722431958, "kl": 0.014693882956635207, "context": 0.15234077908098698}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.1739102303981781, "loss_scale": 4.0, "retries": 0, "seconds": 73.1, "max_memory_gb": 15.95, "ce": 0.13482589134946465, "kl": 0.06337098000221886, "context": 0.16651138267479837}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.7191547155380249, "loss_scale": 4.0, "retries": 0, "seconds": 73.6, "max_memory_gb": 15.95, "ce": 0.2097382228821516, "kl": 0.11425863113254309, "context": 0.26686753146350384}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.11097820103168488, "loss_scale": 4.0, "retries": 0, "seconds": 74.2, "max_memory_gb": 15.95, "ce": 0.02826118574012071, "kl": 0.002259378885355545, "context": 0.02939087455160916}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.32336190342903137, "loss_scale": 4.0, "retries": 0, "seconds": 74.7, "max_memory_gb": 15.95, "ce": 0.1299268826842308, "kl": 0.015282989130355418, "context": 0.13756837649270892}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.8588861227035522, "loss_scale": 4.0, "retries": 0, "seconds": 75.2, "max_memory_gb": 15.95, "ce": 0.19548205006867647, "kl": 0.33997088769683614, "context": 0.3654674938879907}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.20869223773479462, "loss_scale": 4.0, "retries": 0, "seconds": 75.7, "max_memory_gb": 15.95, "ce": 0.019335349556058645, "kl": 0.09125512884929776, "context": 0.06496291630901396}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.3042381703853607, "loss_scale": 4.0, "retries": 0, "seconds": 76.2, "max_memory_gb": 15.95, "ce": 0.11151729780249298, "kl": 0.010129875903658103, "context": 0.11658223462291062}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.39155346155166626, "loss_scale": 4.0, "retries": 0, "seconds": 76.7, "max_memory_gb": 15.95, "ce": 0.16833378095179796, "kl": 0.02166111796395853, "context": 0.17916434723883867}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.5059188604354858, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 15.95, "ce": 0.13344169943593442, "kl": 0.01980598966474645, "context": 0.1433446987066418}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 1.045404076576233, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 15.95, "ce": 0.06358883273787796, "kl": 0.4969692617887631, "context": 0.31207345705479383}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.5380133986473083, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 15.95, "ce": 0.05119899567216635, "kl": 0.23148955474607646, "context": 0.16694377548992634}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.4589660167694092, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 15.95, "ce": 0.44840818061493337, "kl": 0.3177715116762556, "context": 0.6072939289733768}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.3849852085113525, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 15.95, "ce": 0.48208441282622516, "kl": 0.19415376230608672, "context": 0.579161292174831}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.4566863179206848, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 15.95, "ce": 0.13757909077685326, "kl": 0.15917160384560702, "context": 0.21716489444952458}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.26303771138191223, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 15.95, "ce": 0.1430335696786642, "kl": 0.0507754954014672, "context": 0.16842131596058607}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.5284109115600586, "loss_scale": 4.0, "retries": 0, "seconds": 80.9, "max_memory_gb": 15.95, "ce": 0.24397743865847588, "kl": 0.28373573208227754, "context": 0.3858453016728163}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.20333917438983917, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 15.95, "ce": 0.050375175196677446, "kl": 0.07247886180994101, "context": 0.0866146027110517}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.43159234523773193, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 15.95, "ce": 0.11816241138149053, "kl": 0.05377867445349693, "context": 0.14505174942314625}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.5540552139282227, "loss_scale": 4.0, "retries": 0, "seconds": 82.5, "max_memory_gb": 15.95, "ce": 0.23519641254097223, "kl": 0.039484891487518325, "context": 0.25493885995820165}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.20620660483837128, "loss_scale": 4.0, "retries": 0, "seconds": 83.0, "max_memory_gb": 15.95, "ce": 0.10262623615562916, "kl": 0.024499665421899408, "context": 0.11487607005983591}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.9359586834907532, "loss_scale": 4.0, "retries": 0, "seconds": 83.5, "max_memory_gb": 15.95, "ce": 0.08083166414871812, "kl": 0.0954015064307896, "context": 0.12853242293931544}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.6100876927375793, "loss_scale": 4.0, "retries": 0, "seconds": 84.0, "max_memory_gb": 15.95, "ce": 0.23397588357329369, "kl": 0.019216762972064316, "context": 0.2435842640697956}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.4306398332118988, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 15.95, "ce": 0.13997908169403672, "kl": 0.045034731491796265, "context": 0.16249644570052624}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.7790601253509521, "loss_scale": 4.0, "retries": 0, "seconds": 85.1, "max_memory_gb": 15.95, "ce": 0.1805943539366126, "kl": 0.07036561140557751, "context": 0.21577715687453747}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.4218543767929077, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 15.95, "ce": 0.17222784692421556, "kl": 0.01152083047782071, "context": 0.17798825539648533}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.33109116554260254, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 15.95, "ce": 0.07116542709991336, "kl": 0.08457872015424073, "context": 0.11345478519797325}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.7120014429092407, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 15.95, "ce": 0.20466402918100357, "kl": 0.9494803656561999, "context": 0.6794042300898582}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.10716524720191956, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 15.95, "ce": 0.017981384648010135, "kl": 0.00397806751425378, "context": 0.01997041841968894}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 0.8729650378227234, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 15.95, "ce": 0.04668260761536658, "kl": 0.22813153016613796, "context": 0.1607483762782067}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.4268653094768524, "loss_scale": 4.0, "retries": 0, "seconds": 88.2, "max_memory_gb": 15.95, "ce": 0.06969424930866808, "kl": 0.007799948914907873, "context": 0.073594223940745}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 0.556922197341919, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 15.95, "ce": 0.11089609190821648, "kl": 0.07707660383312032, "context": 0.1494343956001103}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.47903794050216675, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 15.95, "ce": 0.06701355776749551, "kl": 0.07950475346297026, "context": 0.10676592960953712}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.49606624245643616, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 15.95, "ce": 0.161424710880965, "kl": 0.16307036904618144, "context": 0.24295989144593477}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.28319597244262695, "loss_scale": 4.0, "retries": 0, "seconds": 90.2, "max_memory_gb": 15.95, "ce": 0.03202814282849431, "kl": 0.009420951188076288, "context": 0.0367386182770133}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.5507784485816956, "loss_scale": 4.0, "retries": 0, "seconds": 90.7, "max_memory_gb": 15.95, "ce": 0.18400276196189225, "kl": 0.12230330955935642, "context": 0.24515441781841218}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.3132356107234955, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 15.95, "ce": 0.15647687017917633, "kl": 0.07975892157992348, "context": 0.19635633379220963}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.317716509103775, "loss_scale": 4.0, "retries": 0, "seconds": 91.7, "max_memory_gb": 15.95, "ce": 0.3366987416520715, "kl": 0.004675566793594044, "context": 0.33903652359731495}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 0.7090888023376465, "loss_scale": 4.0, "retries": 0, "seconds": 92.2, "max_memory_gb": 15.95, "ce": 0.24954045843333006, "kl": 0.05413656064774841, "context": 0.2766087343916297}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.6305882930755615, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 15.95, "ce": 0.07717893761582673, "kl": 0.05158707546070218, "context": 0.10297247627750039}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 0.6249387264251709, "loss_scale": 4.0, "retries": 0, "seconds": 93.2, "max_memory_gb": 15.95, "ce": 0.18015256337821484, "kl": 0.005917940681683831, "context": 0.18311153422109783}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.08192458003759384, "loss_scale": 4.0, "retries": 0, "seconds": 93.7, "max_memory_gb": 15.95, "ce": 0.011518053943291306, "kl": 0.003873885259963572, "context": 0.013454996282234788}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.5130390524864197, "loss_scale": 4.0, "retries": 0, "seconds": 94.2, "max_memory_gb": 15.95, "ce": 0.04155093047302216, "kl": 0.0908604849682888, "context": 0.08698117698077112}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.14081324636936188, "loss_scale": 4.0, "retries": 0, "seconds": 94.7, "max_memory_gb": 15.95, "ce": 0.02893748297356069, "kl": 0.013534010788134765, "context": 0.03570448758546263}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.2655067443847656, "loss_scale": 4.0, "retries": 0, "seconds": 95.2, "max_memory_gb": 15.95, "ce": 0.052846312057226896, "kl": 0.004516003653407097, "context": 0.05510431295260787}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.8353333473205566, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 15.95, "ce": 0.07246554945595562, "kl": 0.361205454391893, "context": 0.25306827831082046}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.1909361183643341, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 15.95, "ce": 0.13192993443226442, "kl": 0.03527028304961277, "context": 0.14956507028546184}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.23515120148658752, "loss_scale": 4.0, "retries": 0, "seconds": 96.8, "max_memory_gb": 15.95, "ce": 0.04441977059468627, "kl": 0.003312718457891606, "context": 0.04607612919062376}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.2041685581207275, "loss_scale": 4.0, "retries": 0, "seconds": 97.3, "max_memory_gb": 15.95, "ce": 0.01625329675152898, "kl": 0.40997354610590264, "context": 0.22124006086960435}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.41262054443359375, "loss_scale": 4.0, "retries": 0, "seconds": 97.8, "max_memory_gb": 15.95, "ce": 0.26581892743706703, "kl": 0.04721181024797261, "context": 0.28942482732236385}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.2150839865207672, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 15.95, "ce": 0.18141297064721584, "kl": 0.006090528084314428, "context": 0.18445823364891112}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.2481170892715454, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 15.95, "ce": 0.03358002414461225, "kl": 0.03222569666104391, "context": 0.04969287384301424}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 0.8170526027679443, "loss_scale": 4.0, "retries": 0, "seconds": 99.3, "max_memory_gb": 15.95, "ce": 0.33117822278290987, "kl": 0.014565389486961067, "context": 0.3384609157219529}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.2219407707452774, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 15.95, "ce": 0.09387757745571434, "kl": 0.009071343811228871, "context": 0.09841324808076024}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.291402667760849, "loss_scale": 4.0, "retries": 0, "seconds": 100.4, "max_memory_gb": 15.95, "ce": 0.02417700202204287, "kl": 0.021827563032275066, "context": 0.035090783378109336}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.2727888822555542, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 15.95, "ce": 0.12353833788074553, "kl": 0.02389275171299232, "context": 0.1354847161564976}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.28819093108177185, "loss_scale": 4.0, "retries": 0, "seconds": 101.4, "max_memory_gb": 15.95, "ce": 0.16258630296215415, "kl": 0.03804560223943554, "context": 0.18160910066217184}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.7633166909217834, "loss_scale": 4.0, "retries": 0, "seconds": 101.9, "max_memory_gb": 15.95, "ce": 0.06648065708577633, "kl": 0.19660242638201453, "context": 0.1647818712517619}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.31189626455307007, "loss_scale": 4.0, "retries": 0, "seconds": 102.4, "max_memory_gb": 15.95, "ce": 0.23236998543143272, "kl": 0.012037859298288822, "context": 0.2383889127522707}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.29826000332832336, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 15.95, "ce": 0.09239146229811013, "kl": 0.011315304480376653, "context": 0.09804911212995648}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.048707373440265656, "loss_scale": 4.0, "retries": 0, "seconds": 103.4, "max_memory_gb": 15.95, "ce": 0.007338333525694907, "kl": 0.004962104270816781, "context": 0.009819385479204357}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.5039463043212891, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 15.95, "ce": 0.157310935202986, "kl": 0.07671477273106575, "context": 0.1956683248281479}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.6273978352546692, "loss_scale": 4.0, "retries": 0, "seconds": 104.4, "max_memory_gb": 15.95, "ce": 0.10565565270371735, "kl": 0.02955469733569771, "context": 0.12043299991637468}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.8971896171569824, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 15.95, "ce": 0.07124029390979558, "kl": 0.3786014438373968, "context": 0.26054100989131257}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.8355220556259155, "loss_scale": 4.0, "retries": 0, "seconds": 105.4, "max_memory_gb": 15.95, "ce": 0.09581160236848518, "kl": 0.07680594206976821, "context": 0.13421457051299512}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 0.880581259727478, "loss_scale": 4.0, "retries": 0, "seconds": 106.0, "max_memory_gb": 15.95, "ce": 0.5587105746380985, "kl": 0.42875170477782376, "context": 0.773086458677426}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.12072141468524933, "loss_scale": 4.0, "retries": 0, "seconds": 106.5, "max_memory_gb": 15.95, "ce": 0.01451492146588862, "kl": 0.006281286703597289, "context": 0.0176555645884946}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.2968360483646393, "loss_scale": 4.0, "retries": 0, "seconds": 107.0, "max_memory_gb": 15.95, "ce": 0.20935672777704895, "kl": 0.016467015026137233, "context": 0.21759023843333125}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.4271309971809387, "loss_scale": 4.0, "retries": 0, "seconds": 107.5, "max_memory_gb": 15.95, "ce": 0.22441005508881062, "kl": 0.0900682673927804, "context": 0.2694441976491362}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.4757001996040344, "loss_scale": 4.0, "retries": 0, "seconds": 108.0, "max_memory_gb": 15.95, "ce": 0.20033953269012272, "kl": 0.09311362903099507, "context": 0.24689635122194886}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.9960716366767883, "loss_scale": 4.0, "retries": 0, "seconds": 108.6, "max_memory_gb": 15.95, "ce": 0.15379909705370665, "kl": 0.6335514744278044, "context": 0.47057484090328217}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.4691871702671051, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 15.95, "ce": 0.304010953521356, "kl": 0.04397695674560964, "context": 0.32599944644607604}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 1.1143763065338135, "loss_scale": 4.0, "retries": 0, "seconds": 109.6, "max_memory_gb": 15.95, "ce": 0.25395193870645016, "kl": 0.27414438780397177, "context": 0.3910241383127868}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 0.8645281791687012, "loss_scale": 4.0, "retries": 0, "seconds": 110.1, "max_memory_gb": 15.95, "ce": 0.393050946528092, "kl": 0.004009773751022294, "context": 0.3950558321084827}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 0.9029553532600403, "loss_scale": 4.0, "retries": 0, "seconds": 110.6, "max_memory_gb": 15.95, "ce": 0.21965217823162675, "kl": 0.11841904069297016, "context": 0.2788617075420916}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.25765717029571533, "loss_scale": 4.0, "retries": 0, "seconds": 111.1, "max_memory_gb": 15.95, "ce": 0.06007270817644894, "kl": 0.032622394915961195, "context": 0.07638390874490142}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.7581710815429688, "loss_scale": 4.0, "retries": 0, "seconds": 111.6, "max_memory_gb": 15.95, "ce": 0.180350475711748, "kl": 0.38432922516949475, "context": 0.37251508329063654}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.41116639971733093, "loss_scale": 4.0, "retries": 0, "seconds": 112.1, "max_memory_gb": 15.95, "ce": 0.12550680991262197, "kl": 0.024831642804201692, "context": 0.13792263297364116}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.6151845455169678, "loss_scale": 4.0, "retries": 0, "seconds": 112.6, "max_memory_gb": 15.95, "ce": 0.0747355897910893, "kl": 0.02047361436416395, "context": 0.08497239835560322}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.833057165145874, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 15.95, "ce": 0.10782253183424473, "kl": 0.06367165921255946, "context": 0.1396583616733551}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.896510601043701, "loss_scale": 4.0, "retries": 0, "seconds": 113.6, "max_memory_gb": 15.95, "ce": 0.37288892455399036, "kl": 0.3085777281958144, "context": 0.5271777901798487}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.3362807333469391, "loss_scale": 4.0, "retries": 0, "seconds": 114.1, "max_memory_gb": 15.95, "ce": 0.20665709115564823, "kl": 0.03948298635077663, "context": 0.22639858024194837}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.4734548032283783, "loss_scale": 4.0, "retries": 0, "seconds": 114.6, "max_memory_gb": 15.95, "ce": 0.07359461323358119, "kl": 0.07855668684351258, "context": 0.11287295492365956}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.3750429153442383, "loss_scale": 4.0, "retries": 0, "seconds": 115.1, "max_memory_gb": 15.95, "ce": 0.0723569099791348, "kl": 0.03547116834670305, "context": 0.09009249322116375}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.38282886147499084, "loss_scale": 4.0, "retries": 0, "seconds": 115.6, "max_memory_gb": 15.95, "ce": 0.09582439856603742, "kl": 0.017727088241372257, "context": 0.10468794126063585}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.1319492757320404, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 15.95, "ce": 0.015443485812284052, "kl": 0.006609735370147973, "context": 0.018748353119008243}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.627704381942749, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 15.95, "ce": 0.08621340547688305, "kl": 0.09616770222783089, "context": 0.13429725915193558}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.5243690609931946, "loss_scale": 4.0, "retries": 0, "seconds": 117.1, "max_memory_gb": 15.95, "ce": 0.2563195419497788, "kl": 0.12512672715820372, "context": 0.3188829096034169}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.6190624833106995, "loss_scale": 4.0, "retries": 0, "seconds": 117.6, "max_memory_gb": 15.95, "ce": 0.25955821480602026, "kl": 0.09081140416674316, "context": 0.30496390839107335}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.3365190625190735, "loss_scale": 4.0, "retries": 0, "seconds": 118.0, "max_memory_gb": 15.95, "ce": 0.220353779848665, "kl": 0.0184336221718695, "context": 0.22957058902829885}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.2729451656341553, "loss_scale": 4.0, "retries": 0, "seconds": 118.5, "max_memory_gb": 15.95, "ce": 0.028296139556914568, "kl": 0.008615477912826464, "context": 0.03260387876071036}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.3701801300048828, "loss_scale": 4.0, "retries": 0, "seconds": 119.0, "max_memory_gb": 15.95, "ce": 0.15444042533636093, "kl": 0.1138467449854943, "context": 0.21136379800736904}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.37401410937309265, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 15.95, "ce": 0.15911474754102528, "kl": 0.023229755599459168, "context": 0.17072962108068168}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.5935570001602173, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 15.95, "ce": 0.3103497119154781, "kl": 0.03179819119395688, "context": 0.32624880969524384}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.30470556020736694, "loss_scale": 4.0, "retries": 0, "seconds": 120.5, "max_memory_gb": 15.95, "ce": 0.0894788505975157, "kl": 0.026448359931237064, "context": 0.10270303254947066}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 0.9583643078804016, "loss_scale": 4.0, "retries": 0, "seconds": 121.0, "max_memory_gb": 15.95, "ce": 0.160876109264791, "kl": 0.05378604245925089, "context": 0.18776913010515273}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.44504332542419434, "loss_scale": 4.0, "retries": 0, "seconds": 121.5, "max_memory_gb": 15.95, "ce": 0.05774216732243076, "kl": 0.11986798968791845, "context": 0.1176761647220701}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.41948652267456055, "loss_scale": 4.0, "retries": 0, "seconds": 122.0, "max_memory_gb": 15.95, "ce": 0.17820369824767113, "kl": 0.033891489438246936, "context": 0.1951494379900396}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.5787099599838257, "loss_scale": 4.0, "retries": 0, "seconds": 122.5, "max_memory_gb": 15.95, "ce": 0.10487299412488937, "kl": 0.04932321031810716, "context": 0.12953460263088346}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.34815236926078796, "loss_scale": 4.0, "retries": 0, "seconds": 123.0, "max_memory_gb": 15.95, "ce": 0.18031814508140087, "kl": 0.015291948453523219, "context": 0.18796411901712418}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.38477060198783875, "loss_scale": 4.0, "retries": 0, "seconds": 123.5, "max_memory_gb": 15.95, "ce": 0.12773318099789321, "kl": 0.03655900130979717, "context": 0.14601268083788455}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.4664730131626129, "loss_scale": 4.0, "retries": 0, "seconds": 124.0, "max_memory_gb": 15.95, "ce": 0.018259513890370727, "kl": 0.10412065374839585, "context": 0.07031984254717827}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 0.6471186876296997, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 15.95, "ce": 0.2763007682515308, "kl": 0.05973370364336006, "context": 0.30616762244608253}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.4289173483848572, "loss_scale": 4.0, "retries": 0, "seconds": 125.0, "max_memory_gb": 15.95, "ce": 0.3755641725147143, "kl": 0.009660845855250955, "context": 0.38039459846913815}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.187254399061203, "loss_scale": 4.0, "retries": 0, "seconds": 125.5, "max_memory_gb": 15.95, "ce": 0.020309363259002566, "kl": 0.022373032959876582, "context": 0.031495881150476635}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.4019409120082855, "loss_scale": 4.0, "retries": 0, "seconds": 126.0, "max_memory_gb": 15.95, "ce": 0.07814291771501303, "kl": 0.04999172408133745, "context": 0.10313877766020596}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.48457053303718567, "loss_scale": 4.0, "retries": 0, "seconds": 126.5, "max_memory_gb": 15.95, "ce": 0.07998528145253658, "kl": 0.03557088109664619, "context": 0.09777072072029114}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.45016756653785706, "loss_scale": 4.0, "retries": 0, "seconds": 127.0, "max_memory_gb": 15.95, "ce": 0.07347918814048171, "kl": 0.03185521788691403, "context": 0.08940679673105478}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.45746204257011414, "loss_scale": 4.0, "retries": 0, "seconds": 127.5, "max_memory_gb": 15.95, "ce": 0.048017550725489855, "kl": 0.00922218769119354, "context": 0.052628646371886134}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.30107784271240234, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 15.95, "ce": 0.14101406221743673, "kl": 0.02144141426833812, "context": 0.1517347702756524}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.4805828928947449, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 15.95, "ce": 0.03736383281648159, "kl": 0.1333574689924717, "context": 0.10404256545007229}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.33688411116600037, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 15.95, "ce": 0.11433225125074387, "kl": 0.006761530057701748, "context": 0.11771301529370248}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.7898370027542114, "loss_scale": 4.0, "retries": 0, "seconds": 129.5, "max_memory_gb": 15.95, "ce": 0.16151254600845277, "kl": 0.19748264749068767, "context": 0.2602538727223873}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.469684362411499, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 15.95, "ce": 0.1458080057054758, "kl": 0.21115628769621253, "context": 0.251386146992445}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.24517381191253662, "loss_scale": 4.0, "retries": 0, "seconds": 130.5, "max_memory_gb": 15.95, "ce": 0.07715926435776055, "kl": 0.021575921440671664, "context": 0.08794722729362547}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.4112851023674011, "loss_scale": 4.0, "retries": 0, "seconds": 131.0, "max_memory_gb": 15.95, "ce": 0.09264465520391241, "kl": 0.013678069874003995, "context": 0.09948368815821595}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.6371095180511475, "loss_scale": 4.0, "retries": 0, "seconds": 131.5, "max_memory_gb": 15.95, "ce": 0.04486498178448528, "kl": 0.08389694686047733, "context": 0.08681345614604652}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.166581392288208, "loss_scale": 4.0, "retries": 0, "seconds": 132.0, "max_memory_gb": 15.95, "ce": 0.019742193864658475, "kl": 0.0330044216098031, "context": 0.03624440496787429}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.5611520409584045, "loss_scale": 4.0, "retries": 0, "seconds": 132.5, "max_memory_gb": 15.95, "ce": 0.09177990211173892, "kl": 0.0975652597262524, "context": 0.14056252921000123}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.7997952699661255, "loss_scale": 4.0, "retries": 0, "seconds": 133.0, "max_memory_gb": 15.95, "ce": 0.14426227426156402, "kl": 0.11638602952007204, "context": 0.2024552896618843}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.7811775803565979, "loss_scale": 4.0, "retries": 0, "seconds": 133.5, "max_memory_gb": 15.95, "ce": 0.18553112167865038, "kl": 0.0869766469404567, "context": 0.22901943745091558}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.6524301767349243, "loss_scale": 4.0, "retries": 0, "seconds": 134.0, "max_memory_gb": 15.95, "ce": 0.17859137430787086, "kl": 0.04429160582367331, "context": 0.2007371811196208}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 1.0127211809158325, "loss_scale": 4.0, "retries": 0, "seconds": 134.5, "max_memory_gb": 15.95, "ce": 0.14019135397393256, "kl": 0.06704967132827733, "context": 0.1737161842174828}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.41473349928855896, "loss_scale": 4.0, "retries": 0, "seconds": 135.0, "max_memory_gb": 15.95, "ce": 0.07945562002714723, "kl": 0.09574527241056785, "context": 0.12732825602870435}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 1.8844549655914307, "loss_scale": 4.0, "retries": 0, "seconds": 135.5, "max_memory_gb": 15.95, "ce": 0.3957741769263521, "kl": 0.09333047550171614, "context": 0.4424394083907828}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.2510358393192291, "loss_scale": 4.0, "retries": 0, "seconds": 136.0, "max_memory_gb": 15.95, "ce": 0.1566306392196566, "kl": 0.005459200299810618, "context": 0.15936023369431496}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.26347875595092773, "loss_scale": 4.0, "retries": 0, "seconds": 136.5, "max_memory_gb": 15.95, "ce": 0.24091015988960862, "kl": 0.0625669035653118, "context": 0.2721936139278114}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 2.8424794673919678, "loss_scale": 4.0, "retries": 0, "seconds": 137.0, "max_memory_gb": 15.95, "ce": 0.31835599057376385, "kl": 0.4139685383779579, "context": 0.5253402600064874}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.6401751041412354, "loss_scale": 4.0, "retries": 0, "seconds": 137.5, "max_memory_gb": 15.95, "ce": 0.16999128088355064, "kl": 0.007175231228757184, "context": 0.17357889120467007}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.2021438181400299, "loss_scale": 4.0, "retries": 0, "seconds": 138.0, "max_memory_gb": 15.95, "ce": 0.038901485968381166, "kl": 0.028994231251999736, "context": 0.05339859938248992}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 0.9694870114326477, "loss_scale": 4.0, "retries": 0, "seconds": 138.5, "max_memory_gb": 15.95, "ce": 0.034736186382360756, "kl": 0.28166933112515835, "context": 0.17557084746658802}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.8418493866920471, "loss_scale": 4.0, "retries": 0, "seconds": 139.0, "max_memory_gb": 15.95, "ce": 0.023644532077014446, "kl": 0.17603304403019138, "context": 0.11166105302982032}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.4186769723892212, "loss_scale": 4.0, "retries": 0, "seconds": 139.5, "max_memory_gb": 15.95, "ce": 0.12101317197084427, "kl": 0.04891864091041498, "context": 0.14547249022871256}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 0.8951378464698792, "loss_scale": 4.0, "retries": 0, "seconds": 140.0, "max_memory_gb": 15.95, "ce": 0.5249963919923175, "kl": 0.3670445726602338, "context": 0.7085186718031764}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.1323932409286499, "loss_scale": 4.0, "retries": 0, "seconds": 140.6, "max_memory_gb": 15.95, "ce": 0.008143092156387866, "kl": 0.0031057246815180406, "context": 0.00969595432979986}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.6351711750030518, "loss_scale": 4.0, "retries": 0, "seconds": 141.1, "max_memory_gb": 15.95, "ce": 0.013649803586304188, "kl": 0.09526679432019591, "context": 0.06128319934941828}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 0.9968785643577576, "loss_scale": 4.0, "retries": 0, "seconds": 141.6, "max_memory_gb": 15.95, "ce": 0.174233622092288, "kl": 0.21437470926321112, "context": 0.28142098372336477}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 1.4566758871078491, "loss_scale": 4.0, "retries": 0, "seconds": 142.1, "max_memory_gb": 15.95, "ce": 0.10849887784570456, "kl": 0.12767515349878522, "context": 0.17233645170927048}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 1.0952237844467163, "loss_scale": 4.0, "retries": 0, "seconds": 142.6, "max_memory_gb": 15.95, "ce": 0.05811751598957926, "kl": 0.021346427223761566, "context": 0.06879072776064277}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 1.488905906677246, "loss_scale": 4.0, "retries": 0, "seconds": 143.2, "max_memory_gb": 15.95, "ce": 0.25421574898064137, "kl": 0.07967518405348528, "context": 0.2940533421933651}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.2902434170246124, "loss_scale": 4.0, "retries": 0, "seconds": 143.7, "max_memory_gb": 15.95, "ce": 0.09287467075046152, "kl": 0.009563022424117662, "context": 0.09765618399251252}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.4745367765426636, "loss_scale": 4.0, "retries": 0, "seconds": 144.2, "max_memory_gb": 15.95, "ce": 0.07384768954943866, "kl": 0.14720863504771842, "context": 0.14745199971366674}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 0.7098067998886108, "loss_scale": 4.0, "retries": 0, "seconds": 144.7, "max_memory_gb": 15.95, "ce": 0.07322041399311274, "kl": 0.023786470092090894, "context": 0.08511364739388227}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.4523392617702484, "loss_scale": 4.0, "retries": 0, "seconds": 145.3, "max_memory_gb": 15.95, "ce": 0.1265433868393302, "kl": 0.020243543141987175, "context": 0.13666515611112118}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.1467592865228653, "loss_scale": 4.0, "retries": 0, "seconds": 145.8, "max_memory_gb": 15.95, "ce": 0.008492389228194952, "kl": 0.013357048588659381, "context": 0.015170913131441921}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.5263454914093018, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 15.95, "ce": 0.11010761640500277, "kl": 0.04516557804890908, "context": 0.13269040873274207}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.11455536633729935, "loss_scale": 4.0, "retries": 0, "seconds": 146.8, "max_memory_gb": 15.95, "ce": 0.007381293398793787, "kl": 0.007697092452872312, "context": 0.011229839525185525}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.257778525352478, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 15.95, "ce": 0.052432523225434124, "kl": 0.058667973054980393, "context": 0.08176651119720191}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.45953309535980225, "loss_scale": 4.0, "retries": 0, "seconds": 147.8, "max_memory_gb": 15.95, "ce": 0.11978248850209638, "kl": 0.047876910353807034, "context": 0.14372094918508083}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.10878945142030716, "loss_scale": 4.0, "retries": 0, "seconds": 148.2, "max_memory_gb": 15.95, "ce": 0.0066273271513637155, "kl": 0.0031048777636897285, "context": 0.008179766009561718}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.7322083711624146, "loss_scale": 4.0, "retries": 0, "seconds": 148.7, "max_memory_gb": 15.95, "ce": 0.06101532356115058, "kl": 0.5109404399991035, "context": 0.3164855409413576}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.4415806829929352, "loss_scale": 4.0, "retries": 0, "seconds": 149.2, "max_memory_gb": 15.95, "ce": 0.05314270639792085, "kl": 0.057671181712066755, "context": 0.08197829616256058}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 0.8966484069824219, "loss_scale": 4.0, "retries": 0, "seconds": 149.7, "max_memory_gb": 15.95, "ce": 0.09648085664957762, "kl": 0.04474455094896257, "context": 0.1188531294465065}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 1.0463567972183228, "loss_scale": 4.0, "retries": 0, "seconds": 150.2, "max_memory_gb": 15.95, "ce": 0.19844547333195806, "kl": 0.07256030867574736, "context": 0.23472563363611698}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 1.6899611949920654, "loss_scale": 4.0, "retries": 0, "seconds": 150.7, "max_memory_gb": 15.95, "ce": 0.254251528007444, "kl": 0.39984167262446135, "context": 0.4541723635047674}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.20735858380794525, "loss_scale": 4.0, "retries": 0, "seconds": 151.2, "max_memory_gb": 15.95, "ce": 0.09519225161056966, "kl": 0.019715999951586127, "context": 0.1050502557773143}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.7892082929611206, "loss_scale": 4.0, "retries": 0, "seconds": 151.7, "max_memory_gb": 15.95, "ce": 0.48029901599511504, "kl": 0.12237576171173714, "context": 0.5414869005326182}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.4624674320220947, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 15.95, "ce": 0.010526880738325417, "kl": 0.1098739019362256, "context": 0.06546383234672248}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.566057562828064, "loss_scale": 4.0, "retries": 0, "seconds": 152.7, "max_memory_gb": 15.95, "ce": 0.2848226462956518, "kl": 0.04195817939762492, "context": 0.3058017317671329}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 0.6117011308670044, "loss_scale": 4.0, "retries": 0, "seconds": 153.1, "max_memory_gb": 15.95, "ce": 0.13142341177444905, "kl": 0.1111437432555249, "context": 0.18699528434081003}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.4540451765060425, "loss_scale": 4.0, "retries": 0, "seconds": 153.7, "max_memory_gb": 15.95, "ce": 0.03278584929648787, "kl": 0.007518003840232268, "context": 0.036544851784128696}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.5028929114341736, "loss_scale": 4.0, "retries": 0, "seconds": 154.2, "max_memory_gb": 15.95, "ce": 0.04334999434649944, "kl": 0.013397846527368529, "context": 0.0500489171827212}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.3215756416320801, "loss_scale": 4.0, "retries": 0, "seconds": 154.7, "max_memory_gb": 15.95, "ce": 0.02262354246340692, "kl": 0.006474407564383, "context": 0.025860745925456285}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 1.8984874486923218, "loss_scale": 4.0, "retries": 0, "seconds": 155.2, "max_memory_gb": 15.95, "ce": 0.1121174693107605, "kl": 0.13806465751986252, "context": 0.1811497979797423}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 0.7822816371917725, "loss_scale": 4.0, "retries": 0, "seconds": 155.6, "max_memory_gb": 15.95, "ce": 0.016056030755862594, "kl": 0.07202099851565436, "context": 0.05206652986817062}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.5988327860832214, "loss_scale": 4.0, "retries": 0, "seconds": 156.2, "max_memory_gb": 15.95, "ce": 0.1075263946549967, "kl": 0.0208488544303691, "context": 0.11795082106254995}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 1.2570196390151978, "loss_scale": 4.0, "retries": 0, "seconds": 156.7, "max_memory_gb": 15.95, "ce": 0.015036735916510224, "kl": 0.5930786760654883, "context": 0.31157608167268336}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 0.5306105017662048, "loss_scale": 4.0, "retries": 0, "seconds": 157.2, "max_memory_gb": 15.95, "ce": 0.17917783092707396, "kl": 0.017885287350509316, "context": 0.18812047492247075}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.6352756023406982, "loss_scale": 4.0, "retries": 0, "seconds": 157.7, "max_memory_gb": 15.95, "ce": 0.028021883801557124, "kl": 0.22879982111044228, "context": 0.1424217945896089}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.3862377107143402, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 15.95, "ce": 0.14281185332220048, "kl": 0.018491450016881572, "context": 0.1520575798349455}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.19272741675376892, "loss_scale": 4.0, "retries": 0, "seconds": 158.6, "max_memory_gb": 15.95, "ce": 0.023939142643939704, "kl": 0.007035157133941539, "context": 0.02745672094170004}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.7390848994255066, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 15.95, "ce": 0.16112336132209748, "kl": 0.034688997169723734, "context": 0.1784678652184084}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.5428082942962646, "loss_scale": 4.0, "retries": 0, "seconds": 159.7, "max_memory_gb": 15.95, "ce": 0.1385990041308105, "kl": 0.08893704030197114, "context": 0.18306751735508442}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.5355097055435181, "loss_scale": 4.0, "retries": 0, "seconds": 160.2, "max_memory_gb": 15.95, "ce": 0.07685000525088981, "kl": 0.054912912426516414, "context": 0.10430646169697866}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 0.942710280418396, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 15.95, "ce": 0.18119779031258076, "kl": 0.6706727992277592, "context": 0.516534196678549}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.6205706000328064, "loss_scale": 4.0, "retries": 0, "seconds": 161.1, "max_memory_gb": 15.95, "ce": 0.1532303737476468, "kl": 0.015568499220535159, "context": 0.16101462580263615}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.7468451261520386, "loss_scale": 4.0, "retries": 0, "seconds": 161.6, "max_memory_gb": 15.95, "ce": 0.03616146370768547, "kl": 0.4003941071059671, "context": 0.23635851312428713}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 0.8319004774093628, "loss_scale": 4.0, "retries": 0, "seconds": 162.1, "max_memory_gb": 15.95, "ce": 0.082296896900516, "kl": 0.05756115622352809, "context": 0.11107747629284859}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.03347909450531, "loss_scale": 4.0, "retries": 0, "seconds": 162.6, "max_memory_gb": 15.95, "ce": 0.14064599701669067, "kl": 0.08299559771512577, "context": 0.18214379309210926}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.4937300980091095, "loss_scale": 4.0, "retries": 0, "seconds": 163.1, "max_memory_gb": 15.95, "ce": 0.08777043083682656, "kl": 0.0812266695429571, "context": 0.1283837640658021}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.22886796295642853, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 15.95, "ce": 0.013340349891223013, "kl": 0.01969963973533595, "context": 0.023190169944427907}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.39727699756622314, "loss_scale": 4.0, "retries": 0, "seconds": 164.1, "max_memory_gb": 15.95, "ce": 0.015485406504012644, "kl": 0.027551017294172198, "context": 0.029260915704071522}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.160470649600029, "loss_scale": 4.0, "retries": 0, "seconds": 164.6, "max_memory_gb": 15.95, "ce": 0.014565361198037863, "kl": 0.005962579314655159, "context": 0.017546650487929583}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.5494588017463684, "loss_scale": 4.0, "retries": 0, "seconds": 165.1, "max_memory_gb": 15.95, "ce": 0.2314469403354451, "kl": 0.013775171566521749, "context": 0.23833452601684257}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.27029937505722046, "loss_scale": 4.0, "retries": 0, "seconds": 165.6, "max_memory_gb": 15.95, "ce": 0.051214570295996964, "kl": 0.03886538364167791, "context": 0.0706472602323629}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 0.7629236578941345, "loss_scale": 4.0, "retries": 0, "seconds": 166.1, "max_memory_gb": 15.95, "ce": 0.07234220716054551, "kl": 0.017417541181202978, "context": 0.08105097428779118}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.5620025992393494, "loss_scale": 4.0, "retries": 0, "seconds": 166.6, "max_memory_gb": 15.95, "ce": 0.12200850649969652, "kl": 0.12926936495205155, "context": 0.18664318858645856}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.4626564383506775, "loss_scale": 4.0, "retries": 0, "seconds": 167.1, "max_memory_gb": 15.95, "ce": 0.10599052626639605, "kl": 0.1591499796777498, "context": 0.1855655089020729}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.5688292384147644, "loss_scale": 4.0, "retries": 0, "seconds": 167.6, "max_memory_gb": 15.95, "ce": 0.1673801628057845, "kl": 0.013738126675889362, "context": 0.17424923210637644}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.8773792386054993, "loss_scale": 4.0, "retries": 0, "seconds": 168.1, "max_memory_gb": 15.95, "ce": 0.27373985201120377, "kl": 0.08185086725279689, "context": 0.31466527841985226}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.2045822143554688, "loss_scale": 4.0, "retries": 0, "seconds": 168.6, "max_memory_gb": 15.95, "ce": 0.13745445664972067, "kl": 0.04149199640232837, "context": 0.15820045536383986}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 0.8048267960548401, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 15.95, "ce": 0.3272516035940498, "kl": 0.02482286561280489, "context": 0.33966302778571844}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 0.6582077741622925, "loss_scale": 4.0, "retries": 0, "seconds": 169.5, "max_memory_gb": 15.95, "ce": 0.18250221654307097, "kl": 0.10229315754259005, "context": 0.23364879982545972}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 3.4742720127105713, "loss_scale": 4.0, "retries": 0, "seconds": 170.0, "max_memory_gb": 15.95, "ce": 0.38076936814468354, "kl": 0.24941795994527638, "context": 0.5054783490486443}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.3689865469932556, "loss_scale": 4.0, "retries": 0, "seconds": 170.5, "max_memory_gb": 15.95, "ce": 0.012445642845705152, "kl": 0.05414221142200404, "context": 0.03951674874406308}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 1.3325576782226562, "loss_scale": 4.0, "retries": 0, "seconds": 171.0, "max_memory_gb": 15.95, "ce": 0.2146381539059803, "kl": 0.09569487429689616, "context": 0.2624855926260352}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 0.7604095339775085, "loss_scale": 4.0, "retries": 0, "seconds": 171.5, "max_memory_gb": 15.95, "ce": 0.18871950125321746, "kl": 0.14710110030137002, "context": 0.26227004965767264}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.5265153646469116, "loss_scale": 4.0, "retries": 0, "seconds": 172.1, "max_memory_gb": 15.95, "ce": 0.1645107590011321, "kl": 0.0073658468027133495, "context": 0.16819367505377159}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.531806230545044, "loss_scale": 4.0, "retries": 0, "seconds": 172.6, "max_memory_gb": 15.95, "ce": 0.16228801966644824, "kl": 0.0566880134283565, "context": 0.1906320246635005}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 1.8355180025100708, "loss_scale": 4.0, "retries": 0, "seconds": 173.1, "max_memory_gb": 15.95, "ce": 0.2554253089474514, "kl": 0.04278840683400631, "context": 0.2768195057287812}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 1.4045902490615845, "loss_scale": 4.0, "retries": 0, "seconds": 173.6, "max_memory_gb": 15.95, "ce": 0.2509148378157988, "kl": 0.9003133890219033, "context": 0.7010715268552303}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.5319455862045288, "loss_scale": 4.0, "retries": 0, "seconds": 174.1, "max_memory_gb": 15.95, "ce": 0.016398509731516242, "kl": 0.07146475231274962, "context": 0.05213088681921363}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.4914867877960205, "loss_scale": 4.0, "retries": 0, "seconds": 174.7, "max_memory_gb": 15.95, "ce": 0.035211680456995964, "kl": 0.019584296736866236, "context": 0.04500382859259844}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.30604198575019836, "loss_scale": 4.0, "retries": 0, "seconds": 175.2, "max_memory_gb": 15.95, "ce": 0.13754120023804717, "kl": 0.16466654597752495, "context": 0.2198744751221966}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.7013235688209534, "loss_scale": 4.0, "retries": 0, "seconds": 175.7, "max_memory_gb": 15.95, "ce": 0.17745406704489142, "kl": 0.005192467593587935, "context": 0.18005030881613493}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 0.8467338681221008, "loss_scale": 4.0, "retries": 0, "seconds": 176.2, "max_memory_gb": 15.95, "ce": 0.177434999961406, "kl": 0.01875347224995494, "context": 0.1868117363192141}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.06268632411956787, "loss_scale": 4.0, "retries": 0, "seconds": 176.7, "max_memory_gb": 15.95, "ce": 0.008831861370708793, "kl": 0.0035296530186315067, "context": 0.010596687789075077}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.12335272133350372, "loss_scale": 4.0, "retries": 0, "seconds": 177.3, "max_memory_gb": 15.95, "ce": 0.10031909571262076, "kl": 0.017816666229919065, "context": 0.10922742879483849}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.4173136353492737, "loss_scale": 4.0, "retries": 0, "seconds": 177.8, "max_memory_gb": 15.95, "ce": 0.15395507961511612, "kl": 0.051378716714680195, "context": 0.17964443936944008}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.592674970626831, "loss_scale": 4.0, "retries": 0, "seconds": 178.3, "max_memory_gb": 15.95, "ce": 0.03905780427157879, "kl": 0.2556320750154555, "context": 0.16687383758835495}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.8516590595245361, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 15.95, "ce": 0.20883938646875322, "kl": 0.24762984085828066, "context": 0.3326543066650629}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.3507815897464752, "loss_scale": 4.0, "retries": 0, "seconds": 179.3, "max_memory_gb": 15.95, "ce": 0.13389604061376303, "kl": 0.062010673122131266, "context": 0.1649013771675527}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.694624662399292, "loss_scale": 4.0, "retries": 0, "seconds": 179.8, "max_memory_gb": 15.95, "ce": 0.19045302469749004, "kl": 0.056072385210427456, "context": 0.21848921629134566}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.4694792330265045, "loss_scale": 4.0, "retries": 0, "seconds": 180.3, "max_memory_gb": 15.95, "ce": 0.13936789621948265, "kl": 0.0194162527914159, "context": 0.14907602168386802}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.293063759803772, "loss_scale": 4.0, "retries": 0, "seconds": 180.8, "max_memory_gb": 15.95, "ce": 0.04288380348589271, "kl": 0.02561986536056793, "context": 0.05569373565958813}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.36072733998298645, "loss_scale": 4.0, "retries": 0, "seconds": 181.3, "max_memory_gb": 15.95, "ce": 0.06856490834616125, "kl": 0.027312414065818302, "context": 0.08222111721988767}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 0.6326504945755005, "loss_scale": 4.0, "retries": 0, "seconds": 181.8, "max_memory_gb": 15.95, "ce": 0.2413253183476627, "kl": 0.014898540102876723, "context": 0.24877458531409502}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.2149968445301056, "loss_scale": 4.0, "retries": 0, "seconds": 182.3, "max_memory_gb": 15.95, "ce": 0.01831074059009552, "kl": 0.017733222397509962, "context": 0.02717735175974667}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.3979392349720001, "loss_scale": 4.0, "retries": 0, "seconds": 182.8, "max_memory_gb": 15.95, "ce": 0.07104189248639159, "kl": 0.014278351329267025, "context": 0.07818106759805232}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.9178228378295898, "loss_scale": 4.0, "retries": 0, "seconds": 183.3, "max_memory_gb": 15.95, "ce": 0.07926831603981555, "kl": 0.2897019496522262, "context": 0.22411928535439074}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 2.2888643741607666, "loss_scale": 4.0, "retries": 0, "seconds": 183.8, "max_memory_gb": 15.95, "ce": 0.30205254518659785, "kl": 0.07032330795482267, "context": 0.33721420471556485}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.4244821071624756, "loss_scale": 4.0, "retries": 0, "seconds": 184.3, "max_memory_gb": 15.95, "ce": 0.06807823985582218, "kl": 0.07826910680159926, "context": 0.10721279308199883}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.788150429725647, "loss_scale": 4.0, "retries": 0, "seconds": 184.8, "max_memory_gb": 15.95, "ce": 0.017955351242562756, "kl": 0.31936684995707765, "context": 0.1776387843710836}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.4532676637172699, "loss_scale": 4.0, "retries": 0, "seconds": 185.3, "max_memory_gb": 15.95, "ce": 0.13828432350419462, "kl": 0.016388556075980887, "context": 0.14647860114928335}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 1.0534305572509766, "loss_scale": 4.0, "retries": 0, "seconds": 185.8, "max_memory_gb": 15.95, "ce": 0.05637076124548912, "kl": 0.33507282682694495, "context": 0.22390717500820756}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.5848124027252197, "loss_scale": 4.0, "retries": 0, "seconds": 186.3, "max_memory_gb": 15.95, "ce": 0.1485791791928932, "kl": 0.11425142534426413, "context": 0.2057048954302445}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.4879307746887207, "loss_scale": 4.0, "retries": 0, "seconds": 186.8, "max_memory_gb": 15.95, "ce": 0.02489021886140108, "kl": 0.04620258152317547, "context": 0.04799150803592056}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.2343515157699585, "loss_scale": 4.0, "retries": 0, "seconds": 187.3, "max_memory_gb": 15.95, "ce": 0.046542848634999245, "kl": 0.0108528331911657, "context": 0.05196926335338503}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 1.2917710542678833, "loss_scale": 4.0, "retries": 0, "seconds": 187.8, "max_memory_gb": 15.95, "ce": 0.3705235216766596, "kl": 0.1514238299569115, "context": 0.4462354234419763}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.020151138305664, "loss_scale": 4.0, "retries": 0, "seconds": 188.3, "max_memory_gb": 15.95, "ce": 0.10655469732591882, "kl": 0.022619417090027127, "context": 0.11786440596915781}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.2011636197566986, "loss_scale": 4.0, "retries": 0, "seconds": 188.8, "max_memory_gb": 15.95, "ce": 0.03694756998447701, "kl": 0.01977809345407877, "context": 0.046836616937071085}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.28250303864479065, "loss_scale": 4.0, "retries": 0, "seconds": 189.3, "max_memory_gb": 15.95, "ce": 0.09114197734743357, "kl": 0.05101929389638826, "context": 0.1166516279336065}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.6611819267272949, "loss_scale": 4.0, "retries": 0, "seconds": 189.8, "max_memory_gb": 15.95, "ce": 0.12736144859809428, "kl": 0.10206185019887926, "context": 0.1783923776820302}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.884861171245575, "loss_scale": 4.0, "retries": 0, "seconds": 190.3, "max_memory_gb": 15.95, "ce": 0.18883419642224908, "kl": 0.1591412709094584, "context": 0.26840483606792986}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.5557407140731812, "loss_scale": 4.0, "retries": 0, "seconds": 190.8, "max_memory_gb": 15.95, "ce": 0.07792185933794826, "kl": 0.03564766842464451, "context": 0.09574569726828486}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.35574421286582947, "loss_scale": 4.0, "retries": 0, "seconds": 191.3, "max_memory_gb": 15.95, "ce": 0.06515932409092784, "kl": 0.11374326525401557, "context": 0.12203095527365804}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.8647810816764832, "loss_scale": 4.0, "retries": 0, "seconds": 191.8, "max_memory_gb": 15.95, "ce": 0.09139565378427505, "kl": 0.05055118340533227, "context": 0.11667124554514885}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.16935871541500092, "loss_scale": 4.0, "retries": 0, "seconds": 192.3, "max_memory_gb": 15.95, "ce": 0.08982652681879699, "kl": 0.04817301392904483, "context": 0.11391303653363138}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.7724935412406921, "loss_scale": 4.0, "retries": 0, "seconds": 192.9, "max_memory_gb": 15.95, "ce": 0.029959187610074878, "kl": 0.13790421356679872, "context": 0.09891129122115672}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.4383062422275543, "loss_scale": 4.0, "retries": 0, "seconds": 193.4, "max_memory_gb": 15.95, "ce": 0.25452452711761, "kl": 0.08679950927034952, "context": 0.2979242855217308}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 1.3748284578323364, "loss_scale": 4.0, "retries": 0, "seconds": 193.9, "max_memory_gb": 15.95, "ce": 0.3102515801729169, "kl": 0.1447986149614735, "context": 0.38265087926993147}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.22501271963119507, "loss_scale": 4.0, "retries": 0, "seconds": 194.4, "max_memory_gb": 15.95, "ce": 0.07657779240980744, "kl": 0.0661318271595519, "context": 0.10964370606234297}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.16963231563568115, "loss_scale": 4.0, "retries": 0, "seconds": 194.9, "max_memory_gb": 15.95, "ce": 0.010019993322202936, "kl": 0.005765828856965527, "context": 0.012902907910756767}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.34030187129974365, "loss_scale": 4.0, "retries": 0, "seconds": 195.4, "max_memory_gb": 15.95, "ce": 0.18421330326236784, "kl": 0.006028699819580652, "context": 0.18722764845006168}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.34980326890945435, "loss_scale": 4.0, "retries": 0, "seconds": 195.9, "max_memory_gb": 15.95, "ce": 0.07080013351514935, "kl": 0.020780032078619115, "context": 0.08119015034753829}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.3668226897716522, "loss_scale": 4.0, "retries": 0, "seconds": 196.4, "max_memory_gb": 15.95, "ce": 0.08378645739867352, "kl": 0.0119904372113524, "context": 0.08978167644818313}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.19264967739582062, "loss_scale": 4.0, "retries": 0, "seconds": 196.9, "max_memory_gb": 15.95, "ce": 0.03956692584324628, "kl": 0.002633481635712087, "context": 0.04088366770884022}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.5484243035316467, "loss_scale": 4.0, "retries": 0, "seconds": 197.5, "max_memory_gb": 15.95, "ce": 0.025189028354361653, "kl": 0.03392668752348982, "context": 0.04215237207245082}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 1.847423791885376, "loss_scale": 4.0, "retries": 0, "seconds": 198.0, "max_memory_gb": 15.95, "ce": 0.09910973347723484, "kl": 0.07900343043729663, "context": 0.13861145079135895}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 0.9705374836921692, "loss_scale": 4.0, "retries": 0, "seconds": 198.5, "max_memory_gb": 15.95, "ce": 0.20724190957844257, "kl": 0.17381973407464102, "context": 0.2941517811268568}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.06252531707286835, "loss_scale": 4.0, "retries": 0, "seconds": 199.0, "max_memory_gb": 15.95, "ce": 0.004988357133697718, "kl": 0.0019030298863071948, "context": 0.005939871887676418}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 1.2612826824188232, "loss_scale": 4.0, "retries": 0, "seconds": 199.5, "max_memory_gb": 15.95, "ce": 0.024468479969073087, "kl": 0.7531608977587894, "context": 0.40104893455281854}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 1.13796067237854, "loss_scale": 4.0, "retries": 0, "seconds": 200.1, "max_memory_gb": 15.95, "ce": 0.1289124534232542, "kl": 0.023352572228759527, "context": 0.14058874046895653}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.5264687538146973, "loss_scale": 4.0, "retries": 0, "seconds": 200.6, "max_memory_gb": 15.95, "ce": 0.0523640476167202, "kl": 0.015317520796088502, "context": 0.060022806166671216}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 0.9801123738288879, "loss_scale": 4.0, "retries": 0, "seconds": 201.0, "max_memory_gb": 15.95, "ce": 0.01236376550514251, "kl": 0.1345910648615245, "context": 0.07965929957572371}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.36638379096984863, "loss_scale": 4.0, "retries": 0, "seconds": 201.5, "max_memory_gb": 15.95, "ce": 0.11602646089158952, "kl": 0.07432988300570287, "context": 0.15319140191422775}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.6372391581535339, "loss_scale": 4.0, "retries": 0, "seconds": 202.0, "max_memory_gb": 15.95, "ce": 0.10396228870376945, "kl": 0.017486652359366417, "context": 0.11270561767742038}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.6278826594352722, "loss_scale": 4.0, "retries": 0, "seconds": 202.6, "max_memory_gb": 15.95, "ce": 0.14647710327699315, "kl": 0.0010012221300712554, "context": 0.14697771694045514}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 1.2940313816070557, "loss_scale": 4.0, "retries": 0, "seconds": 203.1, "max_memory_gb": 15.95, "ce": 0.05388891033362597, "kl": 0.06040067943104077, "context": 0.08408925356343389}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.6132858991622925, "loss_scale": 4.0, "retries": 0, "seconds": 203.6, "max_memory_gb": 15.95, "ce": 0.021871374920010567, "kl": 0.7845805211400148, "context": 0.4141616211272776}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.29341253638267517, "loss_scale": 4.0, "retries": 0, "seconds": 204.1, "max_memory_gb": 15.95, "ce": 0.12550562294200063, "kl": 0.07910970982629806, "context": 0.165060481056571}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.08472525328397751, "loss_scale": 4.0, "retries": 0, "seconds": 204.6, "max_memory_gb": 15.95, "ce": 0.0074661270482465625, "kl": 0.0108437455346575, "context": 0.012887999473605305}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 0.5305556058883667, "loss_scale": 4.0, "retries": 0, "seconds": 205.1, "max_memory_gb": 15.95, "ce": 0.16562363039702177, "kl": 0.06427653517675935, "context": 0.1977618972887285}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.4649367928504944, "loss_scale": 4.0, "retries": 0, "seconds": 205.6, "max_memory_gb": 15.95, "ce": 0.05092326330486685, "kl": 0.03372795903123915, "context": 0.06778724153991789}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.41955089569091797, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 15.95, "ce": 0.15254510147497058, "kl": 0.023017465442535467, "context": 0.1640538316569291}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 1.1654255390167236, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 15.95, "ce": 0.7707657898426987, "kl": 0.3846427599200979, "context": 0.9630871353438124}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 0.6429393291473389, "loss_scale": 4.0, "retries": 0, "seconds": 207.1, "max_memory_gb": 15.95, "ce": 0.2269605090841651, "kl": 0.2529693834367208, "context": 0.35344520094804466}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.5418516397476196, "loss_scale": 4.0, "retries": 0, "seconds": 207.6, "max_memory_gb": 15.95, "ce": 0.1765893279807642, "kl": 0.04724304113915423, "context": 0.20021084853215143}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.7774319052696228, "loss_scale": 4.0, "retries": 0, "seconds": 208.1, "max_memory_gb": 15.95, "ce": 0.04766108395415358, "kl": 0.042188134451862425, "context": 0.06875514978310093}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 3.9412662982940674, "loss_scale": 4.0, "retries": 0, "seconds": 208.6, "max_memory_gb": 15.95, "ce": 0.16462653072085232, "kl": 0.24246802883862983, "context": 0.2858605544315651}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.14389577507972717, "loss_scale": 4.0, "retries": 0, "seconds": 209.1, "max_memory_gb": 15.95, "ce": 0.13878244906663895, "kl": 0.06477297595120035, "context": 0.17116892943158746}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.59125816822052, "loss_scale": 4.0, "retries": 0, "seconds": 209.6, "max_memory_gb": 15.95, "ce": 0.0704846375156194, "kl": 0.05051165330223739, "context": 0.09574046428315341}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 0.7286556959152222, "loss_scale": 4.0, "retries": 0, "seconds": 210.1, "max_memory_gb": 15.95, "ce": 0.21666534879477695, "kl": 0.007214438810478896, "context": 0.2202725627576001}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 1.242743968963623, "loss_scale": 4.0, "retries": 0, "seconds": 210.6, "max_memory_gb": 15.95, "ce": 0.30004953185562044, "kl": 0.0659731577616185, "context": 0.33303609618451446}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.13538646697998047, "loss_scale": 4.0, "retries": 0, "seconds": 211.1, "max_memory_gb": 15.95, "ce": 0.011797031562309712, "kl": 0.0030886231797921937, "context": 0.013341343263164163}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 0.5261528491973877, "loss_scale": 4.0, "retries": 0, "seconds": 211.6, "max_memory_gb": 15.95, "ce": 0.36114413663744926, "kl": 0.08226535096764565, "context": 0.4022768083959818}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.12776823341846466, "loss_scale": 4.0, "retries": 0, "seconds": 212.1, "max_memory_gb": 15.95, "ce": 0.029015911743044853, "kl": 0.03298746970540378, "context": 0.04550964676309377}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.22562283277511597, "loss_scale": 4.0, "retries": 0, "seconds": 212.5, "max_memory_gb": 15.95, "ce": 0.07297552365344018, "kl": 0.051119797382852994, "context": 0.09853542596101761}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 2.0889644622802734, "loss_scale": 4.0, "retries": 0, "seconds": 213.0, "max_memory_gb": 15.95, "ce": 0.14035148159018718, "kl": 0.946400603512302, "context": 0.6135517787188292}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.2530151903629303, "loss_scale": 4.0, "retries": 0, "seconds": 213.5, "max_memory_gb": 15.95, "ce": 0.04580738244112581, "kl": 0.018625755881657824, "context": 0.055120262084528804}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 0.5261475443840027, "loss_scale": 4.0, "retries": 0, "seconds": 214.0, "max_memory_gb": 15.95, "ce": 0.17508681476465426, "kl": 0.007173488371336134, "context": 0.1786735630594194}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 4.036534786224365, "loss_scale": 4.0, "retries": 0, "seconds": 214.5, "max_memory_gb": 15.95, "ce": 0.521525238873437, "kl": 0.44368465756997466, "context": 0.7433675676584244}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.41506221890449524, "loss_scale": 4.0, "retries": 0, "seconds": 215.0, "max_memory_gb": 15.95, "ce": 0.05093007176765241, "kl": 0.03383444491191767, "context": 0.06784729519858956}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.7150752544403076, "loss_scale": 4.0, "retries": 0, "seconds": 215.5, "max_memory_gb": 15.95, "ce": 0.0924458634108305, "kl": 0.11074953672141419, "context": 0.14782063179882243}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 0.8601461052894592, "loss_scale": 4.0, "retries": 0, "seconds": 216.0, "max_memory_gb": 15.95, "ce": 0.1994800460524857, "kl": 0.06839490868151188, "context": 0.2336774985305965}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.411938339471817, "loss_scale": 4.0, "retries": 0, "seconds": 216.5, "max_memory_gb": 15.95, "ce": 0.07008817410678603, "kl": 0.014562015904630243, "context": 0.07736918060982134}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.1321067810058594, "loss_scale": 4.0, "retries": 0, "seconds": 217.0, "max_memory_gb": 15.95, "ce": 0.007181042223237455, "kl": 0.09336708698720031, "context": 0.053864587331190705}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.9512024521827698, "loss_scale": 4.0, "retries": 0, "seconds": 217.5, "max_memory_gb": 15.95, "ce": 0.008990730566438287, "kl": 0.15308739351166878, "context": 0.08553442801348865}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.4029066562652588, "loss_scale": 4.0, "retries": 0, "seconds": 218.0, "max_memory_gb": 15.95, "ce": 0.10201781219802797, "kl": 0.00917017932079034, "context": 0.10660290357191116}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 4.105701446533203, "loss_scale": 4.0, "retries": 0, "seconds": 218.5, "max_memory_gb": 15.95, "ce": 0.43677739502163604, "kl": 0.10180750751169398, "context": 0.48768115625716746}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.4273940920829773, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 15.95, "ce": 0.07696770341135561, "kl": 0.04713350837118924, "context": 0.10053445678204298}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 0.8749368786811829, "loss_scale": 4.0, "retries": 0, "seconds": 219.5, "max_memory_gb": 15.95, "ce": 0.09794704458909109, "kl": 0.10052199236815795, "context": 0.14820803701877594}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.07015752792358398, "loss_scale": 4.0, "retries": 0, "seconds": 220.0, "max_memory_gb": 15.95, "ce": 0.008445403189398348, "kl": 0.0002412717731203884, "context": 0.008566039148718119}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.6720308065414429, "loss_scale": 4.0, "retries": 0, "seconds": 220.5, "max_memory_gb": 15.95, "ce": 0.1188317236956209, "kl": 0.054630573806207394, "context": 0.1461470106150955}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.5079016089439392, "loss_scale": 4.0, "retries": 0, "seconds": 221.0, "max_memory_gb": 15.95, "ce": 0.09208218823187053, "kl": 0.01580587515491061, "context": 0.09998512547463179}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 0.921623170375824, "loss_scale": 4.0, "retries": 0, "seconds": 221.5, "max_memory_gb": 15.95, "ce": 0.09816173819126561, "kl": 0.10052633859413618, "context": 0.14842490362934768}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.5911686420440674, "loss_scale": 4.0, "retries": 0, "seconds": 222.0, "max_memory_gb": 15.95, "ce": 0.09287287780898623, "kl": 0.029092837998177856, "context": 0.10741929721552879}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.631352424621582, "loss_scale": 4.0, "retries": 0, "seconds": 222.5, "max_memory_gb": 15.95, "ce": 0.1447397144511342, "kl": 0.20266818720847368, "context": 0.24607381224632263}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.2911646068096161, "loss_scale": 4.0, "retries": 0, "seconds": 223.0, "max_memory_gb": 15.95, "ce": 0.046285239048302174, "kl": 0.05462619500758592, "context": 0.0735983366612345}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 1.816319227218628, "loss_scale": 4.0, "retries": 0, "seconds": 223.5, "max_memory_gb": 15.95, "ce": 0.18782729608938098, "kl": 0.20124624898016918, "context": 0.2884504187386483}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.42816904187202454, "loss_scale": 4.0, "retries": 0, "seconds": 223.9, "max_memory_gb": 15.95, "ce": 0.09194658824708313, "kl": 0.02990770520409569, "context": 0.10690043680369854}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.30482256412506104, "loss_scale": 4.0, "retries": 0, "seconds": 224.4, "max_memory_gb": 15.95, "ce": 0.07027813512831926, "kl": 0.02888497171079507, "context": 0.08472062065266073}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.3706229329109192, "loss_scale": 4.0, "retries": 0, "seconds": 224.9, "max_memory_gb": 15.95, "ce": 0.1031833792803809, "kl": 0.02756600425345823, "context": 0.11696638364810497}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 1.0251420736312866, "loss_scale": 4.0, "retries": 0, "seconds": 225.4, "max_memory_gb": 15.95, "ce": 0.37227674361201935, "kl": 0.02297135279241047, "context": 0.3837624250445515}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 1.9535808563232422, "loss_scale": 4.0, "retries": 0, "seconds": 225.9, "max_memory_gb": 15.95, "ce": 0.1413639197126031, "kl": 0.08827201875828905, "context": 0.18549992726184428}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.40005388855934143, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 15.95, "ce": 0.18565238756127656, "kl": 0.011260524274803174, "context": 0.1912826441694051}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 1.522573709487915, "loss_scale": 4.0, "retries": 0, "seconds": 226.9, "max_memory_gb": 15.95, "ce": 0.18840317241847515, "kl": 0.09012509370222688, "context": 0.23346571624279022}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.7617127895355225, "loss_scale": 4.0, "retries": 0, "seconds": 227.4, "max_memory_gb": 15.95, "ce": 0.0277971958857961, "kl": 0.01160620756127173, "context": 0.03360030014300719}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.7005104422569275, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 15.95, "ce": 0.020715904014650732, "kl": 0.02904644264071976, "context": 0.03523912528180517}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 1.1109939813613892, "loss_scale": 4.0, "retries": 0, "seconds": 228.4, "max_memory_gb": 15.95, "ce": 0.03229545115027577, "kl": 0.0796352667020983, "context": 0.07211308262776583}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.7327872514724731, "loss_scale": 4.0, "retries": 0, "seconds": 228.9, "max_memory_gb": 15.95, "ce": 0.14823769859503955, "kl": 0.08949282683897763, "context": 0.19298411475028843}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.4136083722114563, "loss_scale": 4.0, "retries": 0, "seconds": 229.4, "max_memory_gb": 15.95, "ce": 0.04644307401031256, "kl": 0.016052081831730902, "context": 0.054469114169478416}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 0.7123482823371887, "loss_scale": 4.0, "retries": 0, "seconds": 229.9, "max_memory_gb": 15.95, "ce": 0.148557293927297, "kl": 0.007177807710831985, "context": 0.15214619715698063}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 1.5684059858322144, "loss_scale": 4.0, "retries": 0, "seconds": 230.3, "max_memory_gb": 15.95, "ce": 0.09899847873020917, "kl": 0.17729758186032996, "context": 0.1876472687581554}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.20894362032413483, "loss_scale": 4.0, "retries": 0, "seconds": 230.8, "max_memory_gb": 15.95, "ce": 0.021922749758232385, "kl": 0.011490474091260694, "context": 0.027667987218592316}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.2151758372783661, "loss_scale": 4.0, "retries": 0, "seconds": 231.3, "max_memory_gb": 15.95, "ce": 0.03567630983889103, "kl": 0.019876057311194018, "context": 0.04561434034258127}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 0.5755382180213928, "loss_scale": 4.0, "retries": 0, "seconds": 231.8, "max_memory_gb": 15.95, "ce": 0.12383708287961781, "kl": 0.12460338641540147, "context": 0.18613877752795815}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 1.410789966583252, "loss_scale": 4.0, "retries": 0, "seconds": 232.4, "max_memory_gb": 15.95, "ce": 0.16124535707058385, "kl": 0.0926499079214409, "context": 0.20757030497770756}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.2757810652256012, "loss_scale": 4.0, "retries": 0, "seconds": 232.9, "max_memory_gb": 15.95, "ce": 0.02675434050615877, "kl": 0.028486869487096556, "context": 0.0409977751551196}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 0.31347474455833435, "loss_scale": 4.0, "retries": 0, "seconds": 233.4, "max_memory_gb": 15.95, "ce": 0.17979506484698504, "kl": 0.140035591495689, "context": 0.24981285980902612}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.6282051801681519, "loss_scale": 4.0, "retries": 0, "seconds": 233.9, "max_memory_gb": 15.95, "ce": 0.05454307672334835, "kl": 0.005444434140372323, "context": 0.05726529416278936}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.3076580762863159, "loss_scale": 4.0, "retries": 0, "seconds": 234.4, "max_memory_gb": 15.95, "ce": 0.037735132209490985, "kl": 0.004331611409725156, "context": 0.03990093804895878}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.11791368573904037, "loss_scale": 4.0, "retries": 0, "seconds": 234.9, "max_memory_gb": 15.95, "ce": 0.008694869233295321, "kl": 0.0017039500235114247, "context": 0.009546844230499119}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 1.5022215843200684, "loss_scale": 4.0, "retries": 0, "seconds": 235.4, "max_memory_gb": 15.95, "ce": 0.021718745701946318, "kl": 0.6202662431569479, "context": 0.3318518686864991}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 1.4359663724899292, "loss_scale": 4.0, "retries": 0, "seconds": 235.9, "max_memory_gb": 15.95, "ce": 0.0052691309683723375, "kl": 0.20899383186770137, "context": 0.10976605030009523}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.32974642515182495, "loss_scale": 4.0, "retries": 0, "seconds": 236.4, "max_memory_gb": 15.95, "ce": 0.04521253646817058, "kl": 0.02218627603724599, "context": 0.05630567448679358}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.23412352800369263, "loss_scale": 4.0, "retries": 0, "seconds": 236.9, "max_memory_gb": 15.95, "ce": 0.027273339277599007, "kl": 0.013733832838624949, "context": 0.03414025518577546}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 1.859748363494873, "loss_scale": 4.0, "retries": 0, "seconds": 237.3, "max_memory_gb": 15.95, "ce": 0.1542997278738767, "kl": 0.03455934813246131, "context": 0.17157939833123237}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.713592290878296, "loss_scale": 4.0, "retries": 0, "seconds": 237.8, "max_memory_gb": 15.95, "ce": 0.004181898781098425, "kl": 0.3485275248749531, "context": 0.17844565922860056}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.0152082443237305, "loss_scale": 4.0, "retries": 0, "seconds": 238.3, "max_memory_gb": 15.95, "ce": 0.1576281636953354, "kl": 0.11853097740822705, "context": 0.21689365536440164}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.1260860711336136, "loss_scale": 4.0, "retries": 0, "seconds": 238.8, "max_memory_gb": 15.95, "ce": 0.008356103440746665, "kl": 0.0043419705689302646, "context": 0.010527088888920844}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.35484787821769714, "loss_scale": 4.0, "retries": 0, "seconds": 239.3, "max_memory_gb": 15.95, "ce": 0.02959036751417443, "kl": 0.01565458952973131, "context": 0.03741766378516331}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.29757916927337646, "loss_scale": 4.0, "retries": 0, "seconds": 239.8, "max_memory_gb": 15.95, "ce": 0.08439421188086271, "kl": 0.012774968796293251, "context": 0.09078169567510486}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 1.0019348859786987, "loss_scale": 4.0, "retries": 0, "seconds": 240.3, "max_memory_gb": 15.95, "ce": 0.18988177480059676, "kl": 0.010660644971721922, "context": 0.19521209778031334}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.5490062832832336, "loss_scale": 4.0, "retries": 0, "seconds": 240.8, "max_memory_gb": 15.95, "ce": 0.03190734327654354, "kl": 0.007192913682956714, "context": 0.0355037989502307}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 3.454287052154541, "loss_scale": 4.0, "retries": 0, "seconds": 241.3, "max_memory_gb": 15.95, "ce": 0.011493732570670545, "kl": 0.5407031418289989, "context": 0.2818453005165793}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.4020825922489166, "loss_scale": 4.0, "retries": 0, "seconds": 241.8, "max_memory_gb": 15.95, "ce": 0.03253051167121157, "kl": 0.0748421007936031, "context": 0.06995156189077534}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 3.109004497528076, "loss_scale": 4.0, "retries": 0, "seconds": 242.3, "max_memory_gb": 15.95, "ce": 0.3759098512819037, "kl": 0.3091605266672559, "context": 0.5304901166819036}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.4336355924606323, "loss_scale": 4.0, "retries": 0, "seconds": 242.8, "max_memory_gb": 15.95, "ce": 0.18025694810785353, "kl": 0.0081729804223869, "context": 0.18434343650005758}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.3358149826526642, "loss_scale": 4.0, "retries": 0, "seconds": 243.3, "max_memory_gb": 15.95, "ce": 0.007004446699284017, "kl": 0.027112870244309306, "context": 0.020560881646815687}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.6880489587783813, "loss_scale": 4.0, "retries": 0, "seconds": 243.8, "max_memory_gb": 15.95, "ce": 0.11468637466896325, "kl": 0.01169431651942432, "context": 0.1205335350241512}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 1.590643048286438, "loss_scale": 4.0, "retries": 0, "seconds": 244.3, "max_memory_gb": 15.95, "ce": 0.1484224750311114, "kl": 0.48661700543016195, "context": 0.39173098653554916}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 1.3371410369873047, "loss_scale": 4.0, "retries": 0, "seconds": 244.8, "max_memory_gb": 15.95, "ce": 0.48990759963635355, "kl": 0.10971907491330057, "context": 0.5447671550791711}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.9558014273643494, "loss_scale": 4.0, "retries": 0, "seconds": 245.3, "max_memory_gb": 15.95, "ce": 0.11483310256153345, "kl": 0.063130269467365, "context": 0.14639823511242867}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 0.6786777377128601, "loss_scale": 4.0, "retries": 0, "seconds": 245.7, "max_memory_gb": 15.95, "ce": 0.15154063527006656, "kl": 0.016929309931583703, "context": 0.160005287558306}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 2.2537949085235596, "loss_scale": 4.0, "retries": 0, "seconds": 246.2, "max_memory_gb": 15.95, "ce": 0.10420572324073873, "kl": 0.08969152541249059, "context": 0.14905148593243212}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.3290829062461853, "loss_scale": 4.0, "retries": 0, "seconds": 246.7, "max_memory_gb": 15.95, "ce": 0.10883697948884219, "kl": 0.02037405699957162, "context": 0.11902401130646467}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 2.148249387741089, "loss_scale": 4.0, "retries": 0, "seconds": 247.2, "max_memory_gb": 15.95, "ce": 0.2616321048117243, "kl": 0.06061989319277927, "context": 0.29194205114617944}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.29457220435142517, "loss_scale": 4.0, "retries": 0, "seconds": 247.7, "max_memory_gb": 15.95, "ce": 0.04277778076357208, "kl": 0.022978057340878877, "context": 0.054266807914245874}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.2810398042201996, "loss_scale": 4.0, "retries": 0, "seconds": 248.2, "max_memory_gb": 15.95, "ce": 0.09187650261446834, "kl": 0.11042887740768492, "context": 0.14709094387944788}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.4612812101840973, "loss_scale": 4.0, "retries": 0, "seconds": 248.7, "max_memory_gb": 15.95, "ce": 0.021371223439928144, "kl": 0.006946910823899088, "context": 0.024844679108355194}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 0.6174985766410828, "loss_scale": 4.0, "retries": 0, "seconds": 249.2, "max_memory_gb": 15.95, "ce": 0.08417998440563679, "kl": 0.029405560111626983, "context": 0.09888276713900268}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 2.133005380630493, "loss_scale": 4.0, "retries": 0, "seconds": 249.7, "max_memory_gb": 15.95, "ce": 0.010546009056270123, "kl": 0.5453383053536527, "context": 0.2832151675829664}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 1.372592806816101, "loss_scale": 4.0, "retries": 0, "seconds": 250.2, "max_memory_gb": 15.95, "ce": 0.08792217820882797, "kl": 0.009632554894778877, "context": 0.09273845748975873}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.7640207409858704, "loss_scale": 4.0, "retries": 0, "seconds": 250.7, "max_memory_gb": 15.95, "ce": 0.03948132193181664, "kl": 0.025855947518721223, "context": 0.05240929429419339}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 0.6614050269126892, "loss_scale": 4.0, "retries": 0, "seconds": 251.2, "max_memory_gb": 15.95, "ce": 0.12831330206245184, "kl": 0.008045780967222527, "context": 0.13233618962112814}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.5060808658599854, "loss_scale": 4.0, "retries": 0, "seconds": 251.6, "max_memory_gb": 15.95, "ce": 0.12724227225407958, "kl": 0.025946870387997478, "context": 0.14021571283228695}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 0.7196337580680847, "loss_scale": 4.0, "retries": 0, "seconds": 252.1, "max_memory_gb": 15.95, "ce": 0.11226816981798038, "kl": 0.021969029654428596, "context": 0.12325268151471391}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 1.2196285724639893, "loss_scale": 4.0, "retries": 0, "seconds": 252.6, "max_memory_gb": 15.95, "ce": 0.26802872779080644, "kl": 0.011575098498724401, "context": 0.27381628239527345}
diff --git a/runs/single_llava/context/L22_s13/adapter/README.md b/runs/single_llava/context/L22_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L22_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L22_s13/adapter/adapter_config.json b/runs/single_llava/context/L22_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2564c875cf7998de1faa5889cca59ddaaa0fbc86
--- /dev/null
+++ b/runs/single_llava/context/L22_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.22.self_attn.v_proj",
+    "model.language_model.layers.22.mlp.down_proj",
+    "model.language_model.layers.22.self_attn.k_proj",
+    "model.language_model.layers.22.mlp.gate_proj",
+    "model.language_model.layers.22.self_attn.q_proj",
+    "model.language_model.layers.22.mlp.up_proj",
+    "model.language_model.layers.22.self_attn.o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L22_s13/config.json b/runs/single_llava/context/L22_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..88434e5d3a3bae484e45ba5361238287bd7d5bc7
--- /dev/null
+++ b/runs/single_llava/context/L22_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 22,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L22_s13/steps.jsonl b/runs/single_llava/context/L22_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..233ae983f98ac325896e46e93b2dd97c4dcdcb49
--- /dev/null
+++ b/runs/single_llava/context/L22_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.1447151005268097, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 15.37, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.23595233261585236, "loss_scale": 4.0, "retries": 0, "seconds": 1.9, "max_memory_gb": 15.38, "ce": 2.679742763051763, "kl": 0.010484062353498302, "context": 2.6849847289267927}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.49216145277023315, "loss_scale": 4.0, "retries": 0, "seconds": 2.3, "max_memory_gb": 15.38, "ce": 5.3925782442092896, "kl": 0.1790521570837882, "context": 5.482104420661926}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.3927610218524933, "loss_scale": 4.0, "retries": 0, "seconds": 2.9, "max_memory_gb": 15.38, "ce": 3.8242339394055307, "kl": 0.05094990033467184, "context": 3.84970880439505}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.030837465077638626, "loss_scale": 4.0, "retries": 0, "seconds": 3.4, "max_memory_gb": 15.38, "ce": 0.08597294939681888, "kl": 0.036797641310840845, "context": 0.10437177028506994}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.1863621473312378, "loss_scale": 4.0, "retries": 0, "seconds": 3.8, "max_memory_gb": 15.38, "ce": 1.4670848567038774, "kl": 0.1835886447224766, "context": 1.55887920409441}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.4207121431827545, "loss_scale": 4.0, "retries": 0, "seconds": 4.3, "max_memory_gb": 15.38, "ce": 3.440500933211297, "kl": 0.10667812876636162, "context": 3.493840005248785}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.3914019465446472, "loss_scale": 4.0, "retries": 0, "seconds": 4.8, "max_memory_gb": 15.38, "ce": 3.696218006312847, "kl": 0.16034812107682228, "context": 3.7763920426368713}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.3667440116405487, "loss_scale": 4.0, "retries": 0, "seconds": 5.3, "max_memory_gb": 15.38, "ce": 2.8415215387940407, "kl": 0.16296003386378288, "context": 2.9230015575885773}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.20532751083374023, "loss_scale": 4.0, "retries": 0, "seconds": 5.8, "max_memory_gb": 15.38, "ce": 1.632177211344242, "kl": 0.06052192521747202, "context": 1.6624381765723228}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.373079389333725, "loss_scale": 4.0, "retries": 0, "seconds": 6.3, "max_memory_gb": 15.38, "ce": 2.629451770335436, "kl": 0.009391197447257582, "context": 2.634147404693067}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.2102476954460144, "loss_scale": 4.0, "retries": 0, "seconds": 6.8, "max_memory_gb": 15.38, "ce": 1.5370866290759295, "kl": 0.01781383276102133, "context": 1.54599355510436}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.40859103202819824, "loss_scale": 4.0, "retries": 0, "seconds": 7.3, "max_memory_gb": 15.38, "ce": 2.67365266662091, "kl": 0.007509968243539333, "context": 2.677407573442906}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.6693932414054871, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 15.38, "ce": 3.893991768360138, "kl": 0.11816903192084283, "context": 3.9530762135982513}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.43497195839881897, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 15.38, "ce": 2.590552619600203, "kl": 0.2682365085056517, "context": 2.724670823547058}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.24731437861919403, "loss_scale": 4.0, "retries": 0, "seconds": 8.7, "max_memory_gb": 15.38, "ce": 1.3747978461906314, "kl": 0.009397531801369041, "context": 1.3794966370332986}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.26990458369255066, "loss_scale": 4.0, "retries": 0, "seconds": 9.2, "max_memory_gb": 15.38, "ce": 1.5944116380997002, "kl": 0.07866036775521934, "context": 1.6337417783215642}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.6799309253692627, "loss_scale": 4.0, "retries": 0, "seconds": 9.7, "max_memory_gb": 15.39, "ce": 3.5692549608647823, "kl": 0.09410130139440298, "context": 3.6163055934011936}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.0892112255096436, "loss_scale": 4.0, "retries": 0, "seconds": 10.2, "max_memory_gb": 15.39, "ce": 5.336391091346741, "kl": 0.3330159450415522, "context": 5.502899169921875}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.009622259065508842, "loss_scale": 4.0, "retries": 0, "seconds": 10.7, "max_memory_gb": 15.39, "ce": 0.012918371823616326, "kl": 0.0021795077991555445, "context": 0.014008125988766551}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.6394709348678589, "loss_scale": 4.0, "retries": 0, "seconds": 11.2, "max_memory_gb": 15.39, "ce": 2.7752427915111184, "kl": 0.029605940231704153, "context": 2.790045689092949}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.6371290683746338, "loss_scale": 4.0, "retries": 0, "seconds": 11.7, "max_memory_gb": 15.39, "ce": 2.9375192038714886, "kl": 0.054401000728830695, "context": 2.964719647541642}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.6353893280029297, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 15.39, "ce": 2.649510978721082, "kl": 0.5097468895546626, "context": 2.904384480789304}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.1800673007965088, "loss_scale": 4.0, "retries": 0, "seconds": 12.6, "max_memory_gb": 15.39, "ce": 5.013323187828064, "kl": 0.014151616371236742, "context": 5.020399034023285}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.0385704040527344, "loss_scale": 4.0, "retries": 0, "seconds": 13.1, "max_memory_gb": 15.39, "ce": 4.233376115560532, "kl": 0.03543674759566784, "context": 4.251094542443752}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.2222214937210083, "loss_scale": 4.0, "retries": 0, "seconds": 13.6, "max_memory_gb": 15.39, "ce": 5.22807240486145, "kl": 0.06483183626551181, "context": 5.260488390922546}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.0877403020858765, "loss_scale": 4.0, "retries": 0, "seconds": 14.1, "max_memory_gb": 15.39, "ce": 4.188569098711014, "kl": 0.0843204811681062, "context": 4.230729341506958}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.7407625913619995, "loss_scale": 4.0, "retries": 0, "seconds": 14.6, "max_memory_gb": 15.39, "ce": 2.798060894012451, "kl": 0.2342296540737152, "context": 2.915175646543503}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 0.80387943983078, "loss_scale": 4.0, "retries": 0, "seconds": 15.1, "max_memory_gb": 15.39, "ce": 2.604506124742329, "kl": 0.5218250418256503, "context": 2.8654186478815973}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 0.8619220852851868, "loss_scale": 4.0, "retries": 0, "seconds": 15.6, "max_memory_gb": 15.39, "ce": 2.7495791567489505, "kl": 0.03736492316238582, "context": 2.768261577002704}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 0.8811210989952087, "loss_scale": 4.0, "retries": 0, "seconds": 16.1, "max_memory_gb": 15.39, "ce": 2.8475463315844536, "kl": 0.02767716330708936, "context": 2.861384868621826}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.717814564704895, "loss_scale": 4.0, "retries": 0, "seconds": 16.6, "max_memory_gb": 15.39, "ce": 5.310238242149353, "kl": 0.014943106274586171, "context": 5.317709684371948}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 0.7562623620033264, "loss_scale": 4.0, "retries": 0, "seconds": 17.1, "max_memory_gb": 15.39, "ce": 2.458728243596852, "kl": 0.17929666046984494, "context": 2.5483766365796328}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 0.9863699674606323, "loss_scale": 4.0, "retries": 0, "seconds": 17.6, "max_memory_gb": 15.39, "ce": 2.6115812635980546, "kl": 0.626547490246594, "context": 2.924854918383062}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.0050848722457886, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 15.39, "ce": 2.7167016100138426, "kl": 0.02935692727623973, "context": 2.7313800351694226}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 0.99996417760849, "loss_scale": 4.0, "retries": 0, "seconds": 18.5, "max_memory_gb": 15.39, "ce": 2.546422939747572, "kl": 0.006929373717866838, "context": 2.549887679517269}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.084216833114624, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 15.39, "ce": 2.4781553335487843, "kl": 0.3401722807902843, "context": 2.6482414938509464}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 1.8119847774505615, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 15.39, "ce": 4.65876829624176, "kl": 0.4937805974041112, "context": 4.905658662319183}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.6710986495018005, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 15.39, "ce": 1.4309589555487037, "kl": 0.034979192074388266, "context": 1.448448516894132}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.1132361888885498, "loss_scale": 4.0, "retries": 0, "seconds": 20.5, "max_memory_gb": 15.39, "ce": 2.7036420106887817, "kl": 0.11882116191554815, "context": 2.7630526572465897}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.5968663096427917, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 15.39, "ce": 1.456144779920578, "kl": 0.17289754888042808, "context": 1.5425935462117195}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.2660013437271118, "loss_scale": 4.0, "retries": 0, "seconds": 21.4, "max_memory_gb": 15.39, "ce": 2.336194299161434, "kl": 0.007080634357407689, "context": 2.339734574779868}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 1.1748462915420532, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 15.39, "ce": 2.4484261944890022, "kl": 0.01171386805435759, "context": 2.4542830949649215}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 0.7124558687210083, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 15.39, "ce": 1.3637008839286864, "kl": 0.274963989155367, "context": 1.501182908192277}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 0.7823801040649414, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 15.39, "ce": 1.3938252702355385, "kl": 0.056081425747834146, "context": 1.4218659363687038}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.2571191787719727, "loss_scale": 4.0, "retries": 0, "seconds": 23.3, "max_memory_gb": 15.39, "ce": 2.2692699879407883, "kl": 0.07355826572165824, "context": 2.3060490489006042}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 1.159786581993103, "loss_scale": 4.0, "retries": 0, "seconds": 23.8, "max_memory_gb": 15.39, "ce": 2.0217790617607534, "kl": 0.03471098514273763, "context": 2.039134527090937}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 2.2366390228271484, "loss_scale": 4.0, "retries": 0, "seconds": 24.3, "max_memory_gb": 15.39, "ce": 3.7699204310774803, "kl": 0.01718903680011863, "context": 3.7785149812698364}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 2.0076682567596436, "loss_scale": 4.0, "retries": 0, "seconds": 24.8, "max_memory_gb": 15.39, "ce": 3.56139699742198, "kl": 0.09523266338510439, "context": 3.6090133283287287}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 2.1819632053375244, "loss_scale": 4.0, "retries": 0, "seconds": 25.3, "max_memory_gb": 15.39, "ce": 3.6174549348652363, "kl": 0.05385017444496043, "context": 3.6443800944834948}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 0.7921296954154968, "loss_scale": 4.0, "retries": 0, "seconds": 25.8, "max_memory_gb": 15.39, "ce": 1.3817239487543702, "kl": 0.1308171411510557, "context": 1.4471324924379587}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 1.6126371622085571, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 15.39, "ce": 2.3075468204915524, "kl": 0.11745475075440481, "context": 2.3662742525339127}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 2.3270909786224365, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 15.39, "ce": 3.7147588431835175, "kl": 0.18475857842713594, "context": 3.8071381896734238}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 2.303886651992798, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 15.39, "ce": 3.789268746972084, "kl": 0.04176152328727767, "context": 3.810149408876896}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 1.5964356660842896, "loss_scale": 4.0, "retries": 0, "seconds": 27.7, "max_memory_gb": 15.39, "ce": 2.452249457128346, "kl": 0.018357325679971837, "context": 2.4614280839450657}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 3.1592650413513184, "loss_scale": 4.0, "retries": 0, "seconds": 28.2, "max_memory_gb": 15.39, "ce": 4.758417010307312, "kl": 0.3155522494344041, "context": 4.916193246841431}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 0.9034658670425415, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 15.39, "ce": 1.2183745224028826, "kl": 0.10169642791152, "context": 1.2692227624356747}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 1.9320193529129028, "loss_scale": 4.0, "retries": 0, "seconds": 29.2, "max_memory_gb": 15.39, "ce": 2.3419697171775624, "kl": 0.003873235811624909, "context": 2.3439063255209476}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.0402084589004517, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 15.39, "ce": 1.4012653697282076, "kl": 0.0732467737980187, "context": 1.4378887098282576}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 1.7053176164627075, "loss_scale": 4.0, "retries": 0, "seconds": 30.1, "max_memory_gb": 15.39, "ce": 2.3816625657491386, "kl": 0.013745878066401929, "context": 2.3885355251841247}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 2.7263123989105225, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 15.39, "ce": 3.437977157533169, "kl": 0.06291324432822876, "context": 3.4694337248802185}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.0441316366195679, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 15.39, "ce": 1.3039333997294307, "kl": 0.03974574740277603, "context": 1.3238062914460897}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 2.97643780708313, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 15.39, "ce": 3.600076080299914, "kl": 0.08618271962768631, "context": 3.643167429137975}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 3.1831276416778564, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 15.4, "ce": 3.5599467400461435, "kl": 0.03600419545546174, "context": 3.5779488757252693}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 2.9209671020507812, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 15.4, "ce": 3.627407357096672, "kl": 0.14102360018296167, "context": 3.6979191452264786}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 0.7618008255958557, "loss_scale": 4.0, "retries": 0, "seconds": 33.1, "max_memory_gb": 15.4, "ce": 0.9162593034561723, "kl": 0.06217379753070418, "context": 0.9473461792804301}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 3.225292205810547, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 15.4, "ce": 3.3060623854398727, "kl": 0.017581686261110008, "context": 3.314853146672249}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.141864538192749, "loss_scale": 4.0, "retries": 0, "seconds": 34.1, "max_memory_gb": 15.4, "ce": 1.1639200528152287, "kl": 0.04439434362575412, "context": 1.1861172646749765}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 2.1930220127105713, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 15.4, "ce": 2.170982665847987, "kl": 0.013170758582418784, "context": 2.1775680095888674}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 2.1990463733673096, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 15.4, "ce": 2.212914186529815, "kl": 0.0040971774578792974, "context": 2.214962838217616}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 3.5730788707733154, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 15.4, "ce": 3.251340886577964, "kl": 0.010898625070694834, "context": 3.256790202111006}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 3.8725690841674805, "loss_scale": 4.0, "retries": 0, "seconds": 36.0, "max_memory_gb": 15.4, "ce": 3.8294644355773926, "kl": 0.15395300276577473, "context": 3.9064409136772156}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.08396172523498535, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 15.4, "ce": 0.2284877742640674, "kl": 0.036843399284407496, "context": 0.24690946703776717}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.358461380004883, "loss_scale": 4.0, "retries": 0, "seconds": 37.0, "max_memory_gb": 15.4, "ce": 2.1895365938544273, "kl": 0.028625304243178107, "context": 2.20384929375723}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.597172498703003, "loss_scale": 4.0, "retries": 0, "seconds": 37.5, "max_memory_gb": 15.4, "ce": 1.5719699934124947, "kl": 0.20013469643890858, "context": 1.6720373407006264}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 2.5449330806732178, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 15.4, "ce": 2.0308461589738727, "kl": 0.4763842076063156, "context": 2.269038274884224}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 4.198533058166504, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 15.4, "ce": 3.3680194914340973, "kl": 0.040461988653987646, "context": 3.388250559568405}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 2.4668264389038086, "loss_scale": 4.0, "retries": 0, "seconds": 38.9, "max_memory_gb": 15.4, "ce": 1.692825736477971, "kl": 0.06211739269201644, "context": 1.7238844353705645}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.4046968221664429, "loss_scale": 4.0, "retries": 0, "seconds": 39.4, "max_memory_gb": 15.4, "ce": 0.908107272349298, "kl": 0.07670348874853516, "context": 0.9464590051211417}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 2.6791913509368896, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 15.4, "ce": 1.9747898344649002, "kl": 0.008469387888908386, "context": 1.9790245924377814}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 3.988036870956421, "loss_scale": 4.0, "retries": 0, "seconds": 40.3, "max_memory_gb": 15.4, "ce": 2.8809148222208023, "kl": 0.048912020603893325, "context": 2.9053708910942078}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 4.78130578994751, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 15.4, "ce": 2.844404101371765, "kl": 0.06112129183020443, "context": 2.874964766204357}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 3.7909297943115234, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 15.4, "ce": 2.4784702006727457, "kl": 0.2085691411048174, "context": 2.582754798233509}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.6600440740585327, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 15.4, "ce": 1.0595357593847439, "kl": 0.07758164367987774, "context": 1.0983265889808536}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.07344183325767517, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 15.4, "ce": 0.07166020595468581, "kl": 0.036651452850492205, "context": 0.08998593408614397}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 5.0249247550964355, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 15.4, "ce": 2.6713771522045135, "kl": 0.20005703071365133, "context": 2.7714056707918644}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 4.808764934539795, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 15.4, "ce": 2.665681380778551, "kl": 0.06839275767561048, "context": 2.6998777836561203}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 3.834118366241455, "loss_scale": 4.0, "retries": 0, "seconds": 43.7, "max_memory_gb": 15.4, "ce": 1.7814393760636449, "kl": 0.02104671909910394, "context": 1.7919627423398197}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.035829175263643265, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 15.4, "ce": 0.11196915083564818, "kl": 0.006183358003909234, "context": 0.1150608288589865}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 3.1246531009674072, "loss_scale": 4.0, "retries": 0, "seconds": 44.7, "max_memory_gb": 15.4, "ce": 1.470609369687736, "kl": 0.01889412710443139, "context": 1.4800564204342663}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.12769736349582672, "loss_scale": 4.0, "retries": 0, "seconds": 45.2, "max_memory_gb": 15.4, "ce": 0.3345063030719757, "kl": 0.09397222800180316, "context": 0.3814924191683531}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.2975471019744873, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 15.4, "ce": 0.8525380238424987, "kl": 0.08007768029347062, "context": 0.8925768344197422}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 4.085591793060303, "loss_scale": 4.0, "retries": 0, "seconds": 46.2, "max_memory_gb": 15.4, "ce": 1.4397837342694402, "kl": 0.06777234474429861, "context": 1.473669925937429}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 6.014838695526123, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 15.4, "ce": 2.594129115343094, "kl": 0.10451018763706088, "context": 2.6463841795921326}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 3.867532968521118, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 15.4, "ce": 1.6992890518158674, "kl": 0.06296284310519695, "context": 1.7307704649865627}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 6.624657154083252, "loss_scale": 4.0, "retries": 0, "seconds": 47.6, "max_memory_gb": 15.4, "ce": 2.453876458108425, "kl": 0.033014987071510404, "context": 2.470383934676647}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.4718737602233887, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 15.4, "ce": 1.0040350295603275, "kl": 0.021134165581315756, "context": 1.0146021284162998}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 4.2077460289001465, "loss_scale": 4.0, "retries": 0, "seconds": 48.6, "max_memory_gb": 15.4, "ce": 1.437376738525927, "kl": 0.04973419762427511, "context": 1.4622438517399132}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.2115192711353302, "loss_scale": 4.0, "retries": 0, "seconds": 49.1, "max_memory_gb": 15.4, "ce": 0.2314399869646877, "kl": 0.06830897077452391, "context": 0.26559447520412505}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 5.982785224914551, "loss_scale": 4.0, "retries": 0, "seconds": 49.5, "max_memory_gb": 15.4, "ce": 2.1613095365464687, "kl": 0.10554088347998913, "context": 2.214079961180687}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 6.478857517242432, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 15.4, "ce": 2.103746771812439, "kl": 0.38759200647473335, "context": 2.297542780637741}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 6.688803672790527, "loss_scale": 4.0, "retries": 0, "seconds": 50.5, "max_memory_gb": 15.4, "ce": 2.295185774564743, "kl": 0.1479680947959423, "context": 2.369169794023037}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 5.232373237609863, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 15.4, "ce": 1.256632765289396, "kl": 0.09949368494562805, "context": 1.3063796011265367}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.9578042030334473, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 15.4, "ce": 0.7149120121612214, "kl": 0.038977303833235055, "context": 0.7344006386119872}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 6.942251205444336, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 15.4, "ce": 1.8375591300427914, "kl": 0.2202371166786179, "context": 1.9476776421070099}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 7.560049057006836, "loss_scale": 4.0, "retries": 0, "seconds": 52.4, "max_memory_gb": 15.4, "ce": 1.670512355864048, "kl": 0.060525392182171345, "context": 1.7007750794291496}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.8390209674835205, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 15.4, "ce": 0.6951434034854174, "kl": 0.01725506060756743, "context": 0.7037709560245275}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 3.0092597007751465, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 15.4, "ce": 0.7068318664096296, "kl": 0.026056181057356298, "context": 0.7198599551338702}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 8.179048538208008, "loss_scale": 4.0, "retries": 0, "seconds": 53.9, "max_memory_gb": 15.4, "ce": 1.6472920924425125, "kl": 0.40654195798560977, "context": 1.850563108921051}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 5.570401668548584, "loss_scale": 4.0, "retries": 0, "seconds": 54.4, "max_memory_gb": 15.4, "ce": 1.2275392226874828, "kl": 0.0429942824121099, "context": 1.2490364164113998}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 4.924809455871582, "loss_scale": 4.0, "retries": 0, "seconds": 54.9, "max_memory_gb": 15.4, "ce": 0.890105961356312, "kl": 0.07621968452622241, "context": 0.9282157807610929}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 5.44606351852417, "loss_scale": 4.0, "retries": 0, "seconds": 55.4, "max_memory_gb": 15.4, "ce": 1.037235061172396, "kl": 0.016188882931601256, "context": 1.0453294988255948}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 6.5084381103515625, "loss_scale": 4.0, "retries": 0, "seconds": 55.8, "max_memory_gb": 15.4, "ce": 1.0009391140192747, "kl": 0.04385090805590153, "context": 1.0228645745664835}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 8.863205909729004, "loss_scale": 4.0, "retries": 0, "seconds": 56.3, "max_memory_gb": 15.4, "ce": 1.481965497136116, "kl": 0.716853360994719, "context": 1.8403922319412231}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 5.795590877532959, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 15.4, "ce": 0.9910114035010338, "kl": 0.2681125346571207, "context": 1.1250676438212395}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.055209532380104065, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 15.4, "ce": 0.11782786715775728, "kl": 0.0061081684834789485, "context": 0.12088195187970996}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 5.038554668426514, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 15.4, "ce": 0.8381923220586032, "kl": 0.012885488336905837, "context": 0.8446350666927174}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 3.010685682296753, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 15.4, "ce": 0.5382845154963434, "kl": 0.048565161298029125, "context": 0.5625670934095979}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.6643643379211426, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 15.4, "ce": 0.38370199501514435, "kl": 0.0619402167503722, "context": 0.41467210836708546}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 5.065732479095459, "loss_scale": 4.0, "retries": 0, "seconds": 59.3, "max_memory_gb": 15.4, "ce": 0.7200466911308467, "kl": 0.0667081200517714, "context": 0.7534007551148534}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 3.0159082412719727, "loss_scale": 4.0, "retries": 0, "seconds": 59.8, "max_memory_gb": 15.4, "ce": 0.618950767442584, "kl": 0.08677597317728214, "context": 0.6623387569561601}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 4.2332282066345215, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 15.4, "ce": 0.5130375297740102, "kl": 0.020037512702401727, "context": 0.5230562859214842}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.7275147438049316, "loss_scale": 4.0, "retries": 0, "seconds": 60.8, "max_memory_gb": 15.4, "ce": 0.7120546344667673, "kl": 0.059761858225101605, "context": 0.7419355753809214}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 1.718770146369934, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 15.4, "ce": 0.33553558122366667, "kl": 0.03273879818152636, "context": 0.3519049845635891}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.298888683319092, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 15.4, "ce": 0.2895185109227896, "kl": 0.02045945217832923, "context": 0.29974823724478483}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 4.818630218505859, "loss_scale": 4.0, "retries": 0, "seconds": 62.2, "max_memory_gb": 15.4, "ce": 0.6905460879206657, "kl": 0.10419372137403116, "context": 0.7426429390907288}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 2.0493366718292236, "loss_scale": 4.0, "retries": 0, "seconds": 62.7, "max_memory_gb": 15.4, "ce": 0.17945058806799352, "kl": 0.0045346216538746376, "context": 0.18171789939515293}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 3.379328489303589, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 15.4, "ce": 0.2992504984140396, "kl": 0.39454281633516075, "context": 0.4965219001751393}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 5.699127674102783, "loss_scale": 4.0, "retries": 0, "seconds": 63.6, "max_memory_gb": 15.4, "ce": 0.3687811978161335, "kl": 0.10140452568884939, "context": 0.4194834604859352}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 4.954501628875732, "loss_scale": 4.0, "retries": 0, "seconds": 64.1, "max_memory_gb": 15.4, "ce": 0.5562704596668482, "kl": 0.4206342546676751, "context": 0.7665876084938645}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 1.6705937385559082, "loss_scale": 4.0, "retries": 0, "seconds": 64.6, "max_memory_gb": 15.4, "ce": 0.3130168356001377, "kl": 0.04912610398605466, "context": 0.3375798910856247}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 3.646599054336548, "loss_scale": 4.0, "retries": 0, "seconds": 65.1, "max_memory_gb": 15.4, "ce": 0.5426438525319099, "kl": 0.25036430318141356, "context": 0.6678259982727468}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 3.1561241149902344, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 15.4, "ce": 0.30656207352876663, "kl": 0.13769420702010393, "context": 0.37540917843580246}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.18458734452724457, "loss_scale": 4.0, "retries": 0, "seconds": 66.0, "max_memory_gb": 15.4, "ce": 0.21163092087954283, "kl": 0.015591851435601711, "context": 0.21942684100940824}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 1.9160704612731934, "loss_scale": 4.0, "retries": 0, "seconds": 66.5, "max_memory_gb": 15.4, "ce": 0.1845645997673273, "kl": 0.02796603401657194, "context": 0.19854762218892574}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.8574165105819702, "loss_scale": 4.0, "retries": 0, "seconds": 67.0, "max_memory_gb": 15.4, "ce": 0.18800923600792885, "kl": 0.044459841679781675, "context": 0.21023915708065033}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 1.3780465126037598, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 15.4, "ce": 0.16932260245084763, "kl": 0.022806261898949742, "context": 0.1807257365435362}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 1.0331687927246094, "loss_scale": 4.0, "retries": 0, "seconds": 68.0, "max_memory_gb": 15.4, "ce": 0.17235196568071842, "kl": 0.0597093200776726, "context": 0.20220662467181683}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 1.5162484645843506, "loss_scale": 4.0, "retries": 0, "seconds": 68.5, "max_memory_gb": 15.4, "ce": 0.25653909891843796, "kl": 0.1299273099284619, "context": 0.32150275632739067}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.26852506399154663, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 15.4, "ce": 0.09195927111431956, "kl": 0.023261764901690185, "context": 0.10359015362337232}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.2361062467098236, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 15.4, "ce": 0.13860740512609482, "kl": 0.01306289411149919, "context": 0.14513885136693716}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.3894960880279541, "loss_scale": 4.0, "retries": 0, "seconds": 69.9, "max_memory_gb": 15.4, "ce": 0.1471020937897265, "kl": 0.015454309585038573, "context": 0.1548292487859726}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.08893922716379166, "loss_scale": 4.0, "retries": 0, "seconds": 70.4, "max_memory_gb": 15.4, "ce": 0.13528231077361852, "kl": 0.060432148369727656, "context": 0.165498387767002}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.6423380970954895, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 15.4, "ce": 0.24649967066943645, "kl": 0.11220239044632763, "context": 0.3026008661836386}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.23227857053279877, "loss_scale": 4.0, "retries": 0, "seconds": 71.4, "max_memory_gb": 15.4, "ce": 0.037987306248396635, "kl": 0.0024663759613758884, "context": 0.03922049398534}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.30899810791015625, "loss_scale": 4.0, "retries": 0, "seconds": 71.9, "max_memory_gb": 15.4, "ce": 0.13964847847819328, "kl": 0.015298191108740866, "context": 0.14729757141321898}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.5662816166877747, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 15.4, "ce": 0.1823608357226476, "kl": 0.354306265769992, "context": 0.3595139749813825}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.17642389237880707, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 15.4, "ce": 0.02773913403507322, "kl": 0.1091988324187696, "context": 0.0823385501280427}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.2476901113986969, "loss_scale": 4.0, "retries": 0, "seconds": 73.4, "max_memory_gb": 15.4, "ce": 0.11867080419324338, "kl": 0.010646741895470768, "context": 0.12399417301639915}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.2854485511779785, "loss_scale": 4.0, "retries": 0, "seconds": 73.9, "max_memory_gb": 15.4, "ce": 0.1897527389228344, "kl": 0.02185814769472927, "context": 0.20068181306123734}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.4955754280090332, "loss_scale": 4.0, "retries": 0, "seconds": 74.4, "max_memory_gb": 15.4, "ce": 0.1550624065566808, "kl": 0.019252430603955872, "context": 0.1646886256057769}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 0.5762892961502075, "loss_scale": 4.0, "retries": 0, "seconds": 74.9, "max_memory_gb": 15.4, "ce": 0.07787212822586298, "kl": 0.4864969390910119, "context": 0.32112059462815523}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.39522331953048706, "loss_scale": 4.0, "retries": 0, "seconds": 75.4, "max_memory_gb": 15.4, "ce": 0.05810255743563175, "kl": 0.21886707725934684, "context": 0.1675360957160592}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.348848581314087, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 15.4, "ce": 0.4796782569028437, "kl": 0.32009390561142936, "context": 0.6397252171300352}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.3843023777008057, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 15.4, "ce": 0.4422088237479329, "kl": 0.1658400344895199, "context": 0.5251288455910981}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.4157678186893463, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 15.4, "ce": 0.15387167630251497, "kl": 0.16103211983499932, "context": 0.23438773211091757}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.358301043510437, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 15.4, "ce": 0.17817748058587313, "kl": 0.05104662131634541, "context": 0.20370078645646572}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.36370065808296204, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 15.4, "ce": 0.2666472252458334, "kl": 0.281016047578305, "context": 0.4071552511304617}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.15250994265079498, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 15.4, "ce": 0.05626622820273042, "kl": 0.07876271195709705, "context": 0.09564757999032736}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.41116955876350403, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 15.4, "ce": 0.1124950828962028, "kl": 0.0714953038841486, "context": 0.14824273716658354}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.29358533024787903, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 15.4, "ce": 0.21591041749343276, "kl": 0.048042841779533774, "context": 0.2399318404495716}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.11342312395572662, "loss_scale": 4.0, "retries": 0, "seconds": 79.8, "max_memory_gb": 15.4, "ce": 0.10292623843997717, "kl": 0.026288001448847353, "context": 0.11607024073600769}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.7956795692443848, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 15.4, "ce": 0.07956849737092853, "kl": 0.08207722549923346, "context": 0.1206071141641587}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.603461503982544, "loss_scale": 4.0, "retries": 0, "seconds": 80.7, "max_memory_gb": 15.4, "ce": 0.2642159927636385, "kl": 0.02094460278749466, "context": 0.2746883016079664}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.22633057832717896, "loss_scale": 4.0, "retries": 0, "seconds": 81.2, "max_memory_gb": 15.4, "ce": 0.17630736017599702, "kl": 0.0535371441674215, "context": 0.20307593420147896}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.5571944713592529, "loss_scale": 4.0, "retries": 0, "seconds": 81.7, "max_memory_gb": 15.4, "ce": 0.1761550735682249, "kl": 0.06038276641629636, "context": 0.20634645223617554}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.22900862991809845, "loss_scale": 4.0, "retries": 0, "seconds": 82.2, "max_memory_gb": 15.4, "ce": 0.20481879147700965, "kl": 0.009920764365233481, "context": 0.2097791787236929}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.30727434158325195, "loss_scale": 4.0, "retries": 0, "seconds": 82.7, "max_memory_gb": 15.4, "ce": 0.08801575424149632, "kl": 0.06881080206949264, "context": 0.12242115335538983}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.162771463394165, "loss_scale": 4.0, "retries": 0, "seconds": 83.2, "max_memory_gb": 15.4, "ce": 0.22592892893590033, "kl": 0.9373412413842743, "context": 0.6945995392743498}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.08474640548229218, "loss_scale": 4.0, "retries": 0, "seconds": 83.7, "max_memory_gb": 15.4, "ce": 0.023754166555590928, "kl": 0.005263705337711144, "context": 0.026386019540950656}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 0.6368759274482727, "loss_scale": 4.0, "retries": 0, "seconds": 84.2, "max_memory_gb": 15.4, "ce": 0.05112189520150423, "kl": 0.2568041479680687, "context": 0.179523971863091}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.5708072781562805, "loss_scale": 4.0, "retries": 0, "seconds": 84.6, "max_memory_gb": 15.4, "ce": 0.09098904649727046, "kl": 0.008035304839722812, "context": 0.09500670246779919}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 0.47855544090270996, "loss_scale": 4.0, "retries": 0, "seconds": 85.1, "max_memory_gb": 15.4, "ce": 0.1236862987279892, "kl": 0.09475963079603389, "context": 0.17106611980125308}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.5535677075386047, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 15.4, "ce": 0.09747983235865831, "kl": 0.05363446852425113, "context": 0.12429706938564777}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.32572752237319946, "loss_scale": 4.0, "retries": 0, "seconds": 86.0, "max_memory_gb": 15.4, "ce": 0.19852460222318769, "kl": 0.167708910536021, "context": 0.2823790553957224}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.2580639719963074, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 15.4, "ce": 0.03315679752267897, "kl": 0.010196946415817365, "context": 0.0382552701048553}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.7483511567115784, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 15.4, "ce": 0.19672443182207644, "kl": 0.11465963724185713, "context": 0.25405424647033215}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.22431282699108124, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 15.4, "ce": 0.19613877311348915, "kl": 0.07351316109998152, "context": 0.23289535706862807}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.15458564460277557, "loss_scale": 4.0, "retries": 0, "seconds": 87.9, "max_memory_gb": 15.4, "ce": 0.32846722286194563, "kl": 0.004909759038127959, "context": 0.3309221086092293}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 0.877203643321991, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 15.4, "ce": 0.2612545215524733, "kl": 0.05418049747822806, "context": 0.28834476228803396}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.5294577479362488, "loss_scale": 4.0, "retries": 0, "seconds": 88.9, "max_memory_gb": 15.4, "ce": 0.0702370353974402, "kl": 0.048254566034302115, "context": 0.094364317599684}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 0.3635766804218292, "loss_scale": 4.0, "retries": 0, "seconds": 89.4, "max_memory_gb": 15.4, "ce": 0.15039601316675544, "kl": 0.005704484661691822, "context": 0.153248252812773}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.07684515416622162, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 15.4, "ce": 0.015006119501776993, "kl": 0.006246513999940362, "context": 0.01812937657814473}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.38032618165016174, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 15.4, "ce": 0.03858875355217606, "kl": 0.09352645432227291, "context": 0.08535198087338358}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.09588927775621414, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 15.4, "ce": 0.02982711116783321, "kl": 0.008538712019799277, "context": 0.034096467308700085}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.24535398185253143, "loss_scale": 4.0, "retries": 0, "seconds": 91.3, "max_memory_gb": 15.4, "ce": 0.059847204480320215, "kl": 0.0055166882229968905, "context": 0.0626055495813489}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.8741748929023743, "loss_scale": 4.0, "retries": 0, "seconds": 91.8, "max_memory_gb": 15.4, "ce": 0.09446442034095526, "kl": 0.31203121336875483, "context": 0.25048003625124693}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.12728716433048248, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 15.4, "ce": 0.1494214286794886, "kl": 0.04424493592159706, "context": 0.17154390510404482}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.17342154681682587, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 15.4, "ce": 0.04358178935945034, "kl": 0.00625941640464589, "context": 0.046711497474461794}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 0.984795868396759, "loss_scale": 4.0, "retries": 0, "seconds": 93.2, "max_memory_gb": 15.4, "ce": 0.016968662035651505, "kl": 0.3648004594142549, "context": 0.19936889596283436}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.12432344257831573, "loss_scale": 4.0, "retries": 0, "seconds": 93.7, "max_memory_gb": 15.4, "ce": 0.2563188746571541, "kl": 0.04778330214321613, "context": 0.28021052945405245}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.16137872636318207, "loss_scale": 4.0, "retries": 0, "seconds": 94.2, "max_memory_gb": 15.4, "ce": 0.18610521592199802, "kl": 0.006350347575789783, "context": 0.18928039353340864}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.3210119307041168, "loss_scale": 4.0, "retries": 0, "seconds": 94.7, "max_memory_gb": 15.4, "ce": 0.04122723359614611, "kl": 0.035696909588295966, "context": 0.05907568824477494}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 0.26914873719215393, "loss_scale": 4.0, "retries": 0, "seconds": 95.1, "max_memory_gb": 15.4, "ce": 0.28156303986907005, "kl": 0.015359360724687576, "context": 0.2892427174374461}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.2510368824005127, "loss_scale": 4.0, "retries": 0, "seconds": 95.6, "max_memory_gb": 15.4, "ce": 0.10168093419633806, "kl": 0.015908084926195443, "context": 0.10963497590273619}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.4761783182621002, "loss_scale": 4.0, "retries": 0, "seconds": 96.1, "max_memory_gb": 15.4, "ce": 0.04286992968991399, "kl": 0.038087397668277845, "context": 0.06191362673416734}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.19678615033626556, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 15.4, "ce": 0.1395555881317705, "kl": 0.024476966318616178, "context": 0.151794076198712}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.19138577580451965, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 15.4, "ce": 0.18409627117216587, "kl": 0.04320831553195603, "context": 0.20570043521001935}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.7677895426750183, "loss_scale": 4.0, "retries": 0, "seconds": 97.6, "max_memory_gb": 15.4, "ce": 0.08143285196274519, "kl": 0.2108237474458292, "context": 0.18684472516179085}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.20813024044036865, "loss_scale": 4.0, "retries": 0, "seconds": 98.1, "max_memory_gb": 15.4, "ce": 0.2542047929018736, "kl": 0.010660767904482782, "context": 0.2595351720228791}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.14688701927661896, "loss_scale": 4.0, "retries": 0, "seconds": 98.6, "max_memory_gb": 15.4, "ce": 0.12128601386211812, "kl": 0.011326175255817361, "context": 0.12694910448044538}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.04234497249126434, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 15.4, "ce": 0.009958699345588684, "kl": 0.007285498882993124, "context": 0.013601448968984187}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.42623355984687805, "loss_scale": 4.0, "retries": 0, "seconds": 99.5, "max_memory_gb": 15.4, "ce": 0.17548016179353, "kl": 0.08140905760228634, "context": 0.21618469152599573}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 0.7654614448547363, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 15.4, "ce": 0.128956719301641, "kl": 0.02735718060284853, "context": 0.1426353061106056}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.6739610433578491, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 15.4, "ce": 0.07991940435022116, "kl": 0.3486730006407015, "context": 0.25425590260419995}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.8064468502998352, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 15.4, "ce": 0.09451346041169018, "kl": 0.06175360461929813, "context": 0.12539026199374348}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 0.49389228224754333, "loss_scale": 4.0, "retries": 0, "seconds": 101.5, "max_memory_gb": 15.4, "ce": 0.5756328902207315, "kl": 0.41880829058936797, "context": 0.7850370393134654}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.11052399128675461, "loss_scale": 4.0, "retries": 0, "seconds": 102.0, "max_memory_gb": 15.4, "ce": 0.017030722461640835, "kl": 0.006181516007927712, "context": 0.020121481036767364}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.19838494062423706, "loss_scale": 4.0, "retries": 0, "seconds": 102.4, "max_memory_gb": 15.4, "ce": 0.22055009379982948, "kl": 0.014464370557107031, "context": 0.2277822787873447}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.3476921617984772, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 15.4, "ce": 0.21589403762482107, "kl": 0.06236601618002169, "context": 0.24707704433239996}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.5665578842163086, "loss_scale": 4.0, "retries": 0, "seconds": 103.4, "max_memory_gb": 15.4, "ce": 0.21300468384288251, "kl": 0.09343462157994509, "context": 0.259721998590976}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.6910268068313599, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 15.4, "ce": 0.19616166781634092, "kl": 0.5994713329710066, "context": 0.4958973303437233}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.1809917688369751, "loss_scale": 4.0, "retries": 0, "seconds": 104.4, "max_memory_gb": 15.4, "ce": 0.2741568465717137, "kl": 0.04661969411972677, "context": 0.29746669298037887}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 0.9192262887954712, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 15.4, "ce": 0.26397998013999313, "kl": 0.2795204231515527, "context": 0.40374018624424934}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 0.43450403213500977, "loss_scale": 4.0, "retries": 0, "seconds": 105.4, "max_memory_gb": 15.4, "ce": 0.36297371820546687, "kl": 0.004924591456074268, "context": 0.36543601332232356}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 0.8195327520370483, "loss_scale": 4.0, "retries": 0, "seconds": 105.9, "max_memory_gb": 15.4, "ce": 0.2846005936153233, "kl": 0.12006953157833777, "context": 0.34463535947725177}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.13421973586082458, "loss_scale": 4.0, "retries": 0, "seconds": 106.4, "max_memory_gb": 15.4, "ce": 0.07058707065880299, "kl": 0.03576096665346995, "context": 0.08846755418926477}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.48625698685646057, "loss_scale": 4.0, "retries": 0, "seconds": 106.9, "max_memory_gb": 15.4, "ce": 0.1832309712190181, "kl": 0.3888171836733818, "context": 0.3776395618915558}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.4695272147655487, "loss_scale": 4.0, "retries": 0, "seconds": 107.4, "max_memory_gb": 15.4, "ce": 0.13739710673689842, "kl": 0.027603177586570382, "context": 0.151198694948107}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.6967727541923523, "loss_scale": 4.0, "retries": 0, "seconds": 107.9, "max_memory_gb": 15.4, "ce": 0.09648661268875003, "kl": 0.02346466068411246, "context": 0.10821894300170243}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 0.9111424088478088, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 15.4, "ce": 0.13134237146005034, "kl": 0.0690091701108031, "context": 0.16584696248173714}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.433351755142212, "loss_scale": 4.0, "retries": 0, "seconds": 108.9, "max_memory_gb": 15.4, "ce": 0.28991277422755957, "kl": 0.30378158536041155, "context": 0.44180355314165354}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.18190982937812805, "loss_scale": 4.0, "retries": 0, "seconds": 109.4, "max_memory_gb": 15.4, "ce": 0.23191123898141086, "kl": 0.03799597092438489, "context": 0.25090922322124243}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.4141000807285309, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 15.4, "ce": 0.09808978345245123, "kl": 0.08020902742282487, "context": 0.1381942993029952}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.21483540534973145, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 15.4, "ce": 0.09022754314355552, "kl": 0.03832955192774534, "context": 0.10939232166856527}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.1972530484199524, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 15.4, "ce": 0.13721284782513976, "kl": 0.02202730596764013, "context": 0.1482265004888177}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.09968860447406769, "loss_scale": 4.0, "retries": 0, "seconds": 111.4, "max_memory_gb": 15.4, "ce": 0.016135154059156775, "kl": 0.00696858816081658, "context": 0.019619447644799948}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.26195502281188965, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 15.4, "ce": 0.11635346955154091, "kl": 0.07920200563967228, "context": 0.15595447272062302}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.24574506282806396, "loss_scale": 4.0, "retries": 0, "seconds": 112.4, "max_memory_gb": 15.4, "ce": 0.2614530832506716, "kl": 0.1283818888477981, "context": 0.32564403116703033}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.3953770697116852, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 15.4, "ce": 0.3116671411553398, "kl": 0.08810267923399806, "context": 0.35571849159896374}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.20064958930015564, "loss_scale": 4.0, "retries": 0, "seconds": 113.3, "max_memory_gb": 15.4, "ce": 0.21395796886645257, "kl": 0.02094876515911892, "context": 0.2244323566555977}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.21001999080181122, "loss_scale": 4.0, "retries": 0, "seconds": 113.8, "max_memory_gb": 15.4, "ce": 0.029294480569660664, "kl": 0.007619708398124203, "context": 0.03310433495789766}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.3453563451766968, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 15.4, "ce": 0.18646672368049622, "kl": 0.11097634142424795, "context": 0.24195490218698978}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.21321536600589752, "loss_scale": 4.0, "retries": 0, "seconds": 114.8, "max_memory_gb": 15.4, "ce": 0.19964645756408572, "kl": 0.020902166230371222, "context": 0.21009753458201885}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.21590012311935425, "loss_scale": 4.0, "retries": 0, "seconds": 115.2, "max_memory_gb": 15.4, "ce": 0.26946935523301363, "kl": 0.03425653593149036, "context": 0.2865976211614907}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.322222501039505, "loss_scale": 4.0, "retries": 0, "seconds": 115.7, "max_memory_gb": 15.4, "ce": 0.09336174163036048, "kl": 0.023166262748418376, "context": 0.10494487290270627}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 1.0216851234436035, "loss_scale": 4.0, "retries": 0, "seconds": 116.2, "max_memory_gb": 15.4, "ce": 0.1827275299001485, "kl": 0.0709896431362722, "context": 0.2182223463896662}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.40252959728240967, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 15.4, "ce": 0.05857055023079738, "kl": 0.12162789522699313, "context": 0.11938449833542109}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.29446038603782654, "loss_scale": 4.0, "retries": 0, "seconds": 117.2, "max_memory_gb": 15.4, "ce": 0.22634033299982548, "kl": 0.03183681395603344, "context": 0.24225874012336135}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.5062927007675171, "loss_scale": 4.0, "retries": 0, "seconds": 117.6, "max_memory_gb": 15.4, "ce": 0.1329890349879861, "kl": 0.046879111556336284, "context": 0.1564285890199244}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.23092542588710785, "loss_scale": 4.0, "retries": 0, "seconds": 118.1, "max_memory_gb": 15.4, "ce": 0.2216418795287609, "kl": 0.01836848951643333, "context": 0.2308261226862669}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.3364754617214203, "loss_scale": 4.0, "retries": 0, "seconds": 118.6, "max_memory_gb": 15.4, "ce": 0.15044748364016414, "kl": 0.03093368577538058, "context": 0.16591432294808328}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.31265029311180115, "loss_scale": 4.0, "retries": 0, "seconds": 119.0, "max_memory_gb": 15.4, "ce": 0.0270583126693964, "kl": 0.082834281551186, "context": 0.06847545295022428}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 0.29831981658935547, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 15.4, "ce": 0.25511002191342413, "kl": 0.06052029113197932, "context": 0.28537016606424004}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.2120722234249115, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 15.4, "ce": 0.3783230644185096, "kl": 0.011664289224427193, "context": 0.38415520265698433}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.23253373801708221, "loss_scale": 4.0, "retries": 0, "seconds": 120.5, "max_memory_gb": 15.4, "ce": 0.029369467054493725, "kl": 0.029943911024020053, "context": 0.044341422501020133}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.29778167605400085, "loss_scale": 4.0, "retries": 0, "seconds": 120.9, "max_memory_gb": 15.4, "ce": 0.11287245142739266, "kl": 0.04265529243275523, "context": 0.13420010032132268}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.5124765038490295, "loss_scale": 4.0, "retries": 0, "seconds": 121.4, "max_memory_gb": 15.4, "ce": 0.09254807978868484, "kl": 0.0384015227900818, "context": 0.11174884252250195}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.4190372824668884, "loss_scale": 4.0, "retries": 0, "seconds": 121.9, "max_memory_gb": 15.4, "ce": 0.08452014392241836, "kl": 0.02761392344837077, "context": 0.09832710539922118}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.4941425323486328, "loss_scale": 4.0, "retries": 0, "seconds": 122.4, "max_memory_gb": 15.4, "ce": 0.06076079490594566, "kl": 0.014037742075743154, "context": 0.06777966674417257}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.1528947651386261, "loss_scale": 4.0, "retries": 0, "seconds": 123.0, "max_memory_gb": 15.4, "ce": 0.15979468368459493, "kl": 0.02605865195801016, "context": 0.17282401339616627}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.41755446791648865, "loss_scale": 4.0, "retries": 0, "seconds": 123.5, "max_memory_gb": 15.4, "ce": 0.04672495252452791, "kl": 0.1411237780121155, "context": 0.11728684045374393}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.10117572546005249, "loss_scale": 4.0, "retries": 0, "seconds": 124.0, "max_memory_gb": 15.4, "ce": 0.10243385448120534, "kl": 0.007891748151450884, "context": 0.10637972503900528}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.7540014386177063, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 15.4, "ce": 0.1867159188259393, "kl": 0.21252343361265957, "context": 0.29297763481736183}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.2965354025363922, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 15.4, "ce": 0.19285033643245697, "kl": 0.2089856518432498, "context": 0.2973431646823883}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.14336663484573364, "loss_scale": 4.0, "retries": 0, "seconds": 125.4, "max_memory_gb": 15.4, "ce": 0.09744661999866366, "kl": 0.02640949648048263, "context": 0.11065137106925249}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.11715585738420486, "loss_scale": 4.0, "retries": 0, "seconds": 125.9, "max_memory_gb": 15.4, "ce": 0.08122014545369893, "kl": 0.014179761681589298, "context": 0.08831002493388951}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 0.78140789270401, "loss_scale": 4.0, "retries": 0, "seconds": 126.4, "max_memory_gb": 15.4, "ce": 0.06976732052862644, "kl": 0.09463576169218868, "context": 0.1170852012000978}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.07840125262737274, "loss_scale": 4.0, "retries": 0, "seconds": 126.9, "max_memory_gb": 15.4, "ce": 0.03139449190348387, "kl": 0.03719965703203343, "context": 0.04999431851319969}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.35763415694236755, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 15.4, "ce": 0.13907392043620348, "kl": 0.11707302019931376, "context": 0.1976104388013482}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.8409661650657654, "loss_scale": 4.0, "retries": 0, "seconds": 127.9, "max_memory_gb": 15.4, "ce": 0.16578286327421665, "kl": 0.11956527573056519, "context": 0.22556549962610006}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.324727326631546, "loss_scale": 4.0, "retries": 0, "seconds": 128.4, "max_memory_gb": 15.4, "ce": 0.248572516720742, "kl": 0.08650067099370062, "context": 0.2918228632770479}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.6165391206741333, "loss_scale": 4.0, "retries": 0, "seconds": 128.9, "max_memory_gb": 15.4, "ce": 0.2420174377039075, "kl": 0.04897586070001125, "context": 0.26650536246597767}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 0.4051951467990875, "loss_scale": 4.0, "retries": 0, "seconds": 129.4, "max_memory_gb": 15.4, "ce": 0.1842012475244701, "kl": 0.07452176496735774, "context": 0.22146213753148913}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.3400169312953949, "loss_scale": 4.0, "retries": 0, "seconds": 129.9, "max_memory_gb": 15.4, "ce": 0.09772585495375097, "kl": 0.07541329367086291, "context": 0.13543250365182757}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 1.8072034120559692, "loss_scale": 4.0, "retries": 0, "seconds": 130.4, "max_memory_gb": 15.4, "ce": 0.3252056580968201, "kl": 0.10561280307592824, "context": 0.3780120642622933}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.09945656359195709, "loss_scale": 4.0, "retries": 0, "seconds": 130.9, "max_memory_gb": 15.4, "ce": 0.15893979440443218, "kl": 0.006902921013534069, "context": 0.16239125281572342}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.14938606321811676, "loss_scale": 4.0, "retries": 0, "seconds": 131.4, "max_memory_gb": 15.4, "ce": 0.26143530663102865, "kl": 0.057261725392891094, "context": 0.29006616678088903}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 1.921977162361145, "loss_scale": 4.0, "retries": 0, "seconds": 131.9, "max_memory_gb": 15.4, "ce": 0.3583393730223179, "kl": 0.4626715521044389, "context": 0.589675160124898}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.22582517564296722, "loss_scale": 4.0, "retries": 0, "seconds": 132.4, "max_memory_gb": 15.4, "ce": 0.14321821322664618, "kl": 0.008239357885031495, "context": 0.14733789581805468}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.33559441566467285, "loss_scale": 4.0, "retries": 0, "seconds": 132.9, "max_memory_gb": 15.4, "ce": 0.07378784380853176, "kl": 0.038801397080533206, "context": 0.09318854054436088}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 0.6879345178604126, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 15.4, "ce": 0.03638438193593174, "kl": 0.318319126003189, "context": 0.19554394739679992}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.727984607219696, "loss_scale": 4.0, "retries": 0, "seconds": 133.8, "max_memory_gb": 15.4, "ce": 0.03434765187557787, "kl": 0.20953146350802854, "context": 0.1391133787110448}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.41785964369773865, "loss_scale": 4.0, "retries": 0, "seconds": 134.3, "max_memory_gb": 15.4, "ce": 0.15658339485526085, "kl": 0.05331774678779766, "context": 0.18324226420372725}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 0.48219820857048035, "loss_scale": 4.0, "retries": 0, "seconds": 134.8, "max_memory_gb": 15.4, "ce": 0.4848431061836891, "kl": 0.3716959678567946, "context": 0.6706910759676248}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.1107642874121666, "loss_scale": 4.0, "retries": 0, "seconds": 135.3, "max_memory_gb": 15.4, "ce": 0.009742444148287177, "kl": 0.00628224533284083, "context": 0.012883566785603762}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.45327937602996826, "loss_scale": 4.0, "retries": 0, "seconds": 135.8, "max_memory_gb": 15.4, "ce": 0.019652945455163717, "kl": 0.0929546115512494, "context": 0.06613025162369013}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 0.7066075801849365, "loss_scale": 4.0, "retries": 0, "seconds": 136.3, "max_memory_gb": 15.4, "ce": 0.15719368180725724, "kl": 0.16373105388629483, "context": 0.23905920755350962}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 0.8669945001602173, "loss_scale": 4.0, "retries": 0, "seconds": 136.8, "max_memory_gb": 15.4, "ce": 0.08955122344195843, "kl": 0.10944398050196469, "context": 0.14427321730181575}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.4152814745903015, "loss_scale": 4.0, "retries": 0, "seconds": 137.3, "max_memory_gb": 15.4, "ce": 0.03980624536052346, "kl": 0.013442290888633579, "context": 0.04652739269658923}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 1.6611826419830322, "loss_scale": 4.0, "retries": 0, "seconds": 137.8, "max_memory_gb": 15.4, "ce": 0.25427349912934005, "kl": 0.0706175196391996, "context": 0.289582259953022}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.1511317491531372, "loss_scale": 4.0, "retries": 0, "seconds": 138.3, "max_memory_gb": 15.4, "ce": 0.11993907298892736, "kl": 0.008348232484422624, "context": 0.12411318731028587}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.6060538291931152, "loss_scale": 4.0, "retries": 0, "seconds": 138.8, "max_memory_gb": 15.4, "ce": 0.08402979280799627, "kl": 0.145856170260231, "context": 0.156957870349288}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 0.6730821132659912, "loss_scale": 4.0, "retries": 0, "seconds": 139.3, "max_memory_gb": 15.4, "ce": 0.10724604036659002, "kl": 0.01836755420663394, "context": 0.11642981483601034}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.3585633635520935, "loss_scale": 4.0, "retries": 0, "seconds": 139.8, "max_memory_gb": 15.4, "ce": 0.1317824083380401, "kl": 0.022949548612814397, "context": 0.14325718022882938}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.17932181060314178, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 15.4, "ce": 0.01512786466628313, "kl": 0.015479904352105223, "context": 0.02286781696602702}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.29084140062332153, "loss_scale": 4.0, "retries": 0, "seconds": 140.8, "max_memory_gb": 15.4, "ce": 0.15826452570036054, "kl": 0.043764305766671896, "context": 0.18014668254181743}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.13001415133476257, "loss_scale": 4.0, "retries": 0, "seconds": 141.4, "max_memory_gb": 15.4, "ce": 0.009716456173919141, "kl": 0.011693822161760181, "context": 0.015563367051072419}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.22765012085437775, "loss_scale": 4.0, "retries": 0, "seconds": 141.8, "max_memory_gb": 15.4, "ce": 0.0717780354898423, "kl": 0.0607365655596368, "context": 0.10214631934650242}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.2740846872329712, "loss_scale": 4.0, "retries": 0, "seconds": 142.4, "max_memory_gb": 15.4, "ce": 0.16976342676207423, "kl": 0.05423626842093654, "context": 0.1968815573491156}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.069212406873703, "loss_scale": 4.0, "retries": 0, "seconds": 142.8, "max_memory_gb": 15.4, "ce": 0.00791269534965977, "kl": 0.003164849946188042, "context": 0.009495120262727141}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 0.977459728717804, "loss_scale": 4.0, "retries": 0, "seconds": 143.3, "max_memory_gb": 15.4, "ce": 0.08305205288343132, "kl": 0.5867156647145748, "context": 0.3764098798856139}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.5252783298492432, "loss_scale": 4.0, "retries": 0, "seconds": 143.8, "max_memory_gb": 15.4, "ce": 0.09163188585080206, "kl": 0.09884238831000403, "context": 0.141053079161793}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 0.7566336393356323, "loss_scale": 4.0, "retries": 0, "seconds": 144.3, "max_memory_gb": 15.4, "ce": 0.10942955873906612, "kl": 0.04037949279882014, "context": 0.1296193040907383}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 0.6671411991119385, "loss_scale": 4.0, "retries": 0, "seconds": 144.8, "max_memory_gb": 15.4, "ce": 0.2863528726156801, "kl": 0.07174177243723534, "context": 0.3222237713634968}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 0.8026087880134583, "loss_scale": 4.0, "retries": 0, "seconds": 145.3, "max_memory_gb": 15.4, "ce": 0.21459542319644243, "kl": 0.39080546621698886, "context": 0.4099981505423784}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.06722210347652435, "loss_scale": 4.0, "retries": 0, "seconds": 145.8, "max_memory_gb": 15.4, "ce": 0.09504646877758205, "kl": 0.022219284146558493, "context": 0.10615611332468688}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 1.7184110879898071, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 15.4, "ce": 0.47957750549539924, "kl": 0.131419240438845, "context": 0.5452871285378933}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.47871133685112, "loss_scale": 4.0, "retries": 0, "seconds": 146.8, "max_memory_gb": 15.4, "ce": 0.015807254414539784, "kl": 0.1020821596175665, "context": 0.06684833508916199}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.19707873463630676, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 15.4, "ce": 0.27916050096973777, "kl": 0.03873218326771166, "context": 0.29852658696472645}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 0.26927927136421204, "loss_scale": 4.0, "retries": 0, "seconds": 147.8, "max_memory_gb": 15.4, "ce": 0.1537135832477361, "kl": 0.10024732717283769, "context": 0.20383725280407816}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.5518786907196045, "loss_scale": 4.0, "retries": 0, "seconds": 148.3, "max_memory_gb": 15.4, "ce": 0.046379889943636954, "kl": 0.012455668445909396, "context": 0.052607723395340145}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.49227374792099, "loss_scale": 4.0, "retries": 0, "seconds": 148.8, "max_memory_gb": 15.4, "ce": 0.0414376676781103, "kl": 0.024385637116211, "context": 0.053630487993359566}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.23748396337032318, "loss_scale": 4.0, "retries": 0, "seconds": 149.3, "max_memory_gb": 15.4, "ce": 0.028560074511915445, "kl": 0.006875654267787468, "context": 0.03199790185317397}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 1.253938913345337, "loss_scale": 4.0, "retries": 0, "seconds": 149.8, "max_memory_gb": 15.4, "ce": 0.12216449552215636, "kl": 0.1387622866895981, "context": 0.19154563162010163}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 0.6336358189582825, "loss_scale": 4.0, "retries": 0, "seconds": 150.3, "max_memory_gb": 15.4, "ce": 0.021286328323185444, "kl": 0.08701441599987447, "context": 0.06479353737086058}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.13485819101333618, "loss_scale": 4.0, "retries": 0, "seconds": 150.8, "max_memory_gb": 15.4, "ce": 0.07426471612416208, "kl": 0.025956790545023978, "context": 0.08724311040714383}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 0.6335402131080627, "loss_scale": 4.0, "retries": 0, "seconds": 151.3, "max_memory_gb": 15.4, "ce": 0.023612492950633168, "kl": 0.5264900492766174, "context": 0.28685752442106605}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 0.2042175829410553, "loss_scale": 4.0, "retries": 0, "seconds": 151.7, "max_memory_gb": 15.4, "ce": 0.19754846103023738, "kl": 0.01999527995940298, "context": 0.20754609908908606}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.6128832697868347, "loss_scale": 4.0, "retries": 0, "seconds": 152.3, "max_memory_gb": 15.4, "ce": 0.037450954085215926, "kl": 0.20843057380989194, "context": 0.14166623773053288}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.12101477384567261, "loss_scale": 4.0, "retries": 0, "seconds": 152.8, "max_memory_gb": 15.4, "ce": 0.13135487213730812, "kl": 0.018172004389271024, "context": 0.14044087124057114}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.24621059000492096, "loss_scale": 4.0, "retries": 0, "seconds": 153.3, "max_memory_gb": 15.4, "ce": 0.03198200790211558, "kl": 0.010738161130575463, "context": 0.03735108824912459}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.4363785684108734, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 15.4, "ce": 0.13845069834496826, "kl": 0.045883886370575055, "context": 0.16139264381490648}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.25351589918136597, "loss_scale": 4.0, "retries": 0, "seconds": 154.3, "max_memory_gb": 15.4, "ce": 0.11985634919255972, "kl": 0.08948507270542905, "context": 0.16459887754172087}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.2709140181541443, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 15.4, "ce": 0.07250707515049726, "kl": 0.04681831900961697, "context": 0.09591623255982995}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 0.7042652368545532, "loss_scale": 4.0, "retries": 0, "seconds": 155.3, "max_memory_gb": 15.4, "ce": 0.18997042393311858, "kl": 0.6333175846375525, "context": 0.5066291969269514}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.4843156635761261, "loss_scale": 4.0, "retries": 0, "seconds": 155.8, "max_memory_gb": 15.4, "ce": 0.1974628483876586, "kl": 0.021314871031790972, "context": 0.208120284602046}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.5170180797576904, "loss_scale": 4.0, "retries": 0, "seconds": 156.2, "max_memory_gb": 15.4, "ce": 0.043198018334805965, "kl": 0.3747348531323951, "context": 0.23056544223800302}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 0.9244309663772583, "loss_scale": 4.0, "retries": 0, "seconds": 156.7, "max_memory_gb": 15.4, "ce": 0.11328635737299919, "kl": 0.07504702667938545, "context": 0.15080987103283405}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.37582528591156, "loss_scale": 4.0, "retries": 0, "seconds": 157.3, "max_memory_gb": 15.4, "ce": 0.1760068633593619, "kl": 0.11125648135930533, "context": 0.23163511417806149}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.26495736837387085, "loss_scale": 4.0, "retries": 0, "seconds": 157.7, "max_memory_gb": 15.4, "ce": 0.11260267486795783, "kl": 0.07731710723601282, "context": 0.15126122860237956}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.3082098066806793, "loss_scale": 4.0, "retries": 0, "seconds": 158.2, "max_memory_gb": 15.4, "ce": 0.023385029286146164, "kl": 0.03442830656422302, "context": 0.040599183179438114}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.49163755774497986, "loss_scale": 4.0, "retries": 0, "seconds": 158.7, "max_memory_gb": 15.4, "ce": 0.021819339133799076, "kl": 0.07035196793731302, "context": 0.05699532199651003}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.18368616700172424, "loss_scale": 4.0, "retries": 0, "seconds": 159.2, "max_memory_gb": 15.4, "ce": 0.021691816160455346, "kl": 0.006973602459765971, "context": 0.02517861733213067}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.19970734417438507, "loss_scale": 4.0, "retries": 0, "seconds": 159.7, "max_memory_gb": 15.4, "ce": 0.2269379555946216, "kl": 0.014495927542157006, "context": 0.23418592184316367}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.24753636121749878, "loss_scale": 4.0, "retries": 0, "seconds": 160.2, "max_memory_gb": 15.4, "ce": 0.06254950282163918, "kl": 0.03872123884502798, "context": 0.08191012451425195}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 1.4226266145706177, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 15.4, "ce": 0.15855681808898225, "kl": 0.06697034309763694, "context": 0.1920419892994687}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.4027283191680908, "loss_scale": 4.0, "retries": 0, "seconds": 161.2, "max_memory_gb": 15.4, "ce": 0.1170510402880609, "kl": 0.1451618921928457, "context": 0.18963198689743876}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.23243851959705353, "loss_scale": 4.0, "retries": 0, "seconds": 161.7, "max_memory_gb": 15.4, "ce": 0.09487862512469292, "kl": 0.15741195253212936, "context": 0.17358460114337504}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.5234545469284058, "loss_scale": 4.0, "retries": 0, "seconds": 162.2, "max_memory_gb": 15.4, "ce": 0.1696660955203697, "kl": 0.011334687820635736, "context": 0.17533344170078635}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.5019903779029846, "loss_scale": 4.0, "retries": 0, "seconds": 162.7, "max_memory_gb": 15.4, "ce": 0.262909721583128, "kl": 0.07847622153349221, "context": 0.3021478410810232}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.4425544738769531, "loss_scale": 4.0, "retries": 0, "seconds": 163.2, "max_memory_gb": 15.4, "ce": 0.19555861502885818, "kl": 0.0402010157413315, "context": 0.215659121517092}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 0.24381963908672333, "loss_scale": 4.0, "retries": 0, "seconds": 163.7, "max_memory_gb": 15.4, "ce": 0.2909347643144429, "kl": 0.025978889083489776, "context": 0.30392421456053853}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 0.3469201624393463, "loss_scale": 4.0, "retries": 0, "seconds": 164.2, "max_memory_gb": 15.4, "ce": 0.17352258739992976, "kl": 0.09408417687518522, "context": 0.22056467202492058}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 2.4089221954345703, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 15.4, "ce": 0.32708238856866956, "kl": 0.20117956609465182, "context": 0.42767217895016074}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.48808786273002625, "loss_scale": 4.0, "retries": 0, "seconds": 165.2, "max_memory_gb": 15.4, "ce": 0.03544633206911385, "kl": 0.04956428718287498, "context": 0.06022847513668239}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 1.126731276512146, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 15.4, "ce": 0.21822935808449984, "kl": 0.11266023362986743, "context": 0.2745594750158489}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 0.8169593214988708, "loss_scale": 4.0, "retries": 0, "seconds": 166.2, "max_memory_gb": 15.4, "ce": 0.20017244620248675, "kl": 0.1463514305651188, "context": 0.2733481624163687}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.23195263743400574, "loss_scale": 4.0, "retries": 0, "seconds": 166.7, "max_memory_gb": 15.4, "ce": 0.1450178457889706, "kl": 0.011980333365499973, "context": 0.15100801549851894}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.3481121361255646, "loss_scale": 4.0, "retries": 0, "seconds": 167.2, "max_memory_gb": 15.4, "ce": 0.19642392406240106, "kl": 0.06037515282514505, "context": 0.22661150409840047}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 1.0889956951141357, "loss_scale": 4.0, "retries": 0, "seconds": 167.8, "max_memory_gb": 15.4, "ce": 0.22081494005396962, "kl": 0.029223980847746134, "context": 0.2354269241914153}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 0.8164137005805969, "loss_scale": 4.0, "retries": 0, "seconds": 168.3, "max_memory_gb": 15.4, "ce": 0.24199744022917002, "kl": 0.8320641964673996, "context": 0.658029530197382}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.3117169439792633, "loss_scale": 4.0, "retries": 0, "seconds": 168.8, "max_memory_gb": 15.4, "ce": 0.02266967436298728, "kl": 0.04348941275384277, "context": 0.04441438103094697}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.356818825006485, "loss_scale": 4.0, "retries": 0, "seconds": 169.3, "max_memory_gb": 15.4, "ce": 0.038235400803387165, "kl": 0.019602943386416882, "context": 0.0480368728749454}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.15274293720722198, "loss_scale": 4.0, "retries": 0, "seconds": 169.8, "max_memory_gb": 15.4, "ce": 0.14848983334377408, "kl": 0.16198989673284814, "context": 0.22948478162288666}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.3575687110424042, "loss_scale": 4.0, "retries": 0, "seconds": 170.3, "max_memory_gb": 15.4, "ce": 0.21679599478375167, "kl": 0.006819571106461808, "context": 0.22020578105002642}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 0.9737385511398315, "loss_scale": 4.0, "retries": 0, "seconds": 170.7, "max_memory_gb": 15.4, "ce": 0.20340174157172441, "kl": 0.029063781024888158, "context": 0.21793363522738218}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.0607379674911499, "loss_scale": 4.0, "retries": 0, "seconds": 171.2, "max_memory_gb": 15.4, "ce": 0.012034356128424406, "kl": 0.004659570404328406, "context": 0.014364141854457557}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.07917540520429611, "loss_scale": 4.0, "retries": 0, "seconds": 171.8, "max_memory_gb": 15.4, "ce": 0.10950859333388507, "kl": 0.017739267073920928, "context": 0.11837822699453682}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.23230095207691193, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 15.4, "ce": 0.16296467557549477, "kl": 0.04865530878305435, "context": 0.1872923281043768}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 0.9760363698005676, "loss_scale": 4.0, "retries": 0, "seconds": 172.8, "max_memory_gb": 15.4, "ce": 0.03825792297720909, "kl": 0.27591491091880016, "context": 0.1762153811287135}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.182715654373169, "loss_scale": 4.0, "retries": 0, "seconds": 173.3, "max_memory_gb": 15.4, "ce": 0.2626752551877871, "kl": 0.28323519555851817, "context": 0.4042928423732519}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.20652824640274048, "loss_scale": 4.0, "retries": 0, "seconds": 173.8, "max_memory_gb": 15.4, "ce": 0.13941044360399246, "kl": 0.05409206071635708, "context": 0.16645648144185543}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.7220286130905151, "loss_scale": 4.0, "retries": 0, "seconds": 174.3, "max_memory_gb": 15.4, "ce": 0.21802026592195034, "kl": 0.03202691415208392, "context": 0.23403372848406434}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.5946110486984253, "loss_scale": 4.0, "retries": 0, "seconds": 174.8, "max_memory_gb": 15.4, "ce": 0.15773606818402186, "kl": 0.023300946166273206, "context": 0.1693865389097482}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.13095736503601074, "loss_scale": 4.0, "retries": 0, "seconds": 175.2, "max_memory_gb": 15.4, "ce": 0.04714291647542268, "kl": 0.026371259591542184, "context": 0.06032854539807886}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.269622802734375, "loss_scale": 4.0, "retries": 0, "seconds": 175.7, "max_memory_gb": 15.4, "ce": 0.08019575825892389, "kl": 0.027095028577605262, "context": 0.09374327282421291}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 0.24314570426940918, "loss_scale": 4.0, "retries": 0, "seconds": 176.2, "max_memory_gb": 15.4, "ce": 0.23061798932030797, "kl": 0.013466329779475927, "context": 0.23735115630552173}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.18384303152561188, "loss_scale": 4.0, "retries": 0, "seconds": 176.7, "max_memory_gb": 15.4, "ce": 0.02448546257801354, "kl": 0.020394593011587858, "context": 0.0346827597822994}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.23463614284992218, "loss_scale": 4.0, "retries": 0, "seconds": 177.2, "max_memory_gb": 15.4, "ce": 0.07817243365570903, "kl": 0.017153199485619552, "context": 0.08674903376959264}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.076684832572937, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 15.4, "ce": 0.08079360885312781, "kl": 0.41279390706768027, "context": 0.28719056490808725}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 1.671782374382019, "loss_scale": 4.0, "retries": 0, "seconds": 178.2, "max_memory_gb": 15.4, "ce": 0.2938468448119238, "kl": 0.0455718501470983, "context": 0.31663276976905763}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.24719667434692383, "loss_scale": 4.0, "retries": 0, "seconds": 178.7, "max_memory_gb": 15.4, "ce": 0.08161125640617684, "kl": 0.07157388236373663, "context": 0.11739819776266813}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.2833677530288696, "loss_scale": 4.0, "retries": 0, "seconds": 179.2, "max_memory_gb": 15.4, "ce": 0.01573999592801556, "kl": 0.4071353483304847, "context": 0.2193076663534157}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.3007417619228363, "loss_scale": 4.0, "retries": 0, "seconds": 179.7, "max_memory_gb": 15.4, "ce": 0.16517858859151602, "kl": 0.018823016522219405, "context": 0.17459010030142963}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.5341799259185791, "loss_scale": 4.0, "retries": 0, "seconds": 180.2, "max_memory_gb": 15.4, "ce": 0.07573373429477215, "kl": 0.29796870599966496, "context": 0.2247180799022317}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.2533748745918274, "loss_scale": 4.0, "retries": 0, "seconds": 180.7, "max_memory_gb": 15.4, "ce": 0.14788877731189132, "kl": 0.11150774490670301, "context": 0.2036426574923098}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.4678506851196289, "loss_scale": 4.0, "retries": 0, "seconds": 181.1, "max_memory_gb": 15.4, "ce": 0.03267364460043609, "kl": 0.03875807824624644, "context": 0.05205268366262317}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.1859121024608612, "loss_scale": 4.0, "retries": 0, "seconds": 181.6, "max_memory_gb": 15.4, "ce": 0.06476155365817249, "kl": 0.012073386053089052, "context": 0.07079824944958091}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 0.6185338497161865, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 15.4, "ce": 0.394179729744792, "kl": 0.15422594384290278, "context": 0.47129270527511835}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.3355064392089844, "loss_scale": 4.0, "retries": 0, "seconds": 182.6, "max_memory_gb": 15.4, "ce": 0.12192957533989102, "kl": 0.02001230935275089, "context": 0.13193573197349906}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.20417140424251556, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 15.4, "ce": 0.04922553198412061, "kl": 0.027259949725703336, "context": 0.06285550817847252}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.14294590055942535, "loss_scale": 4.0, "retries": 0, "seconds": 183.6, "max_memory_gb": 15.4, "ce": 0.0971808317117393, "kl": 0.044027402240317315, "context": 0.11919453437440097}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.4238734841346741, "loss_scale": 4.0, "retries": 0, "seconds": 184.1, "max_memory_gb": 15.4, "ce": 0.16702630592044443, "kl": 0.10680395862800651, "context": 0.22042828873964027}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.8698848485946655, "loss_scale": 4.0, "retries": 0, "seconds": 184.6, "max_memory_gb": 15.4, "ce": 0.19543385854922235, "kl": 0.17372337383858394, "context": 0.28229554649442434}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.3597739636898041, "loss_scale": 4.0, "retries": 0, "seconds": 185.1, "max_memory_gb": 15.4, "ce": 0.1254778227303177, "kl": 0.039282865036511794, "context": 0.14511924795806408}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.1742037683725357, "loss_scale": 4.0, "retries": 0, "seconds": 185.6, "max_memory_gb": 15.4, "ce": 0.07365916180424392, "kl": 0.09662958746775985, "context": 0.12197395064868033}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.48426926136016846, "loss_scale": 4.0, "retries": 0, "seconds": 186.1, "max_memory_gb": 15.4, "ce": 0.08967112563550472, "kl": 0.03978721908060834, "context": 0.10956473369151354}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.1198723241686821, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 15.4, "ce": 0.10345874563790858, "kl": 0.04228300112299621, "context": 0.12460024608299136}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.3972000181674957, "loss_scale": 4.0, "retries": 0, "seconds": 187.1, "max_memory_gb": 15.4, "ce": 0.039873044937849045, "kl": 0.09172200196189806, "context": 0.08573404606431723}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.24162204563617706, "loss_scale": 4.0, "retries": 0, "seconds": 187.6, "max_memory_gb": 15.4, "ce": 0.2778704399242997, "kl": 0.08601892518345267, "context": 0.3208799031563103}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 1.2133476734161377, "loss_scale": 4.0, "retries": 0, "seconds": 188.0, "max_memory_gb": 15.4, "ce": 0.30722080456325784, "kl": 0.2375675551011227, "context": 0.4260045816190541}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.1902768462896347, "loss_scale": 4.0, "retries": 0, "seconds": 188.5, "max_memory_gb": 15.4, "ce": 0.09071423334535211, "kl": 0.06431178651109803, "context": 0.12287012708839029}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.13747283816337585, "loss_scale": 4.0, "retries": 0, "seconds": 189.0, "max_memory_gb": 15.4, "ce": 0.012878493056632578, "kl": 0.009977162146242335, "context": 0.017867074348032475}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.16127954423427582, "loss_scale": 4.0, "retries": 0, "seconds": 189.4, "max_memory_gb": 15.4, "ce": 0.20124971540644765, "kl": 0.006268048688070849, "context": 0.20438373927026987}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.1648857444524765, "loss_scale": 4.0, "retries": 0, "seconds": 189.9, "max_memory_gb": 15.4, "ce": 0.07359148189425468, "kl": 0.02095129579629429, "context": 0.08406713278964162}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.17032350599765778, "loss_scale": 4.0, "retries": 0, "seconds": 190.3, "max_memory_gb": 15.4, "ce": 0.08092411269899458, "kl": 0.014272971340687945, "context": 0.08806059625931084}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.12397414445877075, "loss_scale": 4.0, "retries": 0, "seconds": 190.8, "max_memory_gb": 15.4, "ce": 0.04818964464357123, "kl": 0.004353112948592752, "context": 0.050366202485747635}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.36214637756347656, "loss_scale": 4.0, "retries": 0, "seconds": 191.2, "max_memory_gb": 15.4, "ce": 0.027932589640840888, "kl": 0.027597930529736914, "context": 0.04173155454918742}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 1.1599230766296387, "loss_scale": 4.0, "retries": 0, "seconds": 191.7, "max_memory_gb": 15.4, "ce": 0.10459582274779677, "kl": 0.06354432506486773, "context": 0.1363679850474}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 0.6283210515975952, "loss_scale": 4.0, "retries": 0, "seconds": 192.2, "max_memory_gb": 15.4, "ce": 0.25855540856719017, "kl": 0.15640740108210593, "context": 0.3367591090500355}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.08465103060007095, "loss_scale": 4.0, "retries": 0, "seconds": 192.7, "max_memory_gb": 15.4, "ce": 0.0076844049617648125, "kl": 0.004620886698830873, "context": 0.009994848223868757}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 0.9306550025939941, "loss_scale": 4.0, "retries": 0, "seconds": 193.2, "max_memory_gb": 15.4, "ce": 0.022119165398180485, "kl": 0.7324291469994932, "context": 0.3883337266743183}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.3312014639377594, "loss_scale": 4.0, "retries": 0, "seconds": 193.7, "max_memory_gb": 15.4, "ce": 0.10771094402298331, "kl": 0.021775619592517614, "context": 0.11859875312075019}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.25491863489151, "loss_scale": 4.0, "retries": 0, "seconds": 194.2, "max_memory_gb": 15.4, "ce": 0.062073148670606315, "kl": 0.01731601287610829, "context": 0.07073115441016853}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 0.7665557861328125, "loss_scale": 4.0, "retries": 0, "seconds": 194.6, "max_memory_gb": 15.4, "ce": 0.018204326974228024, "kl": 0.1792466833794606, "context": 0.1078276684274897}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.2640487253665924, "loss_scale": 4.0, "retries": 0, "seconds": 195.1, "max_memory_gb": 15.4, "ce": 0.14745001518167555, "kl": 0.07698397601780016, "context": 0.1859420012915507}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.43020734190940857, "loss_scale": 4.0, "retries": 0, "seconds": 195.6, "max_memory_gb": 15.4, "ce": 0.10819490347057581, "kl": 0.007713314378634095, "context": 0.11205156240612268}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.28749462962150574, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 15.4, "ce": 0.1495147819223348, "kl": 0.001159353019829723, "context": 0.15009446302428842}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 1.3447535037994385, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 15.4, "ce": 0.08901425986550748, "kl": 0.0811504320736276, "context": 0.1295894756913185}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.4390156269073486, "loss_scale": 4.0, "retries": 0, "seconds": 197.1, "max_memory_gb": 15.4, "ce": 0.06835501873865724, "kl": 0.635310256970115, "context": 0.38601014111191034}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.23952873051166534, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 15.4, "ce": 0.14604138396680355, "kl": 0.07769973878748715, "context": 0.18489125277847052}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.057798366993665695, "loss_scale": 4.0, "retries": 0, "seconds": 198.1, "max_memory_gb": 15.4, "ce": 0.010644966270774603, "kl": 0.011093520384747535, "context": 0.016191726666875184}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 0.2332935929298401, "loss_scale": 4.0, "retries": 0, "seconds": 198.6, "max_memory_gb": 15.4, "ce": 0.17161060019861907, "kl": 0.06266170546587091, "context": 0.20294145331718028}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.3322718143463135, "loss_scale": 4.0, "retries": 0, "seconds": 199.0, "max_memory_gb": 15.4, "ce": 0.06360949610825628, "kl": 0.036148556624539196, "context": 0.08168377401307225}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.3163456320762634, "loss_scale": 4.0, "retries": 0, "seconds": 199.5, "max_memory_gb": 15.4, "ce": 0.191971420426853, "kl": 0.02964458876522258, "context": 0.20679370942525566}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 0.7179010510444641, "loss_scale": 4.0, "retries": 0, "seconds": 200.0, "max_memory_gb": 15.4, "ce": 0.7323058848269284, "kl": 0.3406556178815663, "context": 0.9026336895767599}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 0.4258269965648651, "loss_scale": 4.0, "retries": 0, "seconds": 200.5, "max_memory_gb": 15.4, "ce": 0.2346894070506096, "kl": 0.24395895114867017, "context": 0.35666888346895576}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.5984184741973877, "loss_scale": 4.0, "retries": 0, "seconds": 201.0, "max_memory_gb": 15.4, "ce": 0.20121080288663507, "kl": 0.06112707086504088, "context": 0.23177433689124882}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.6291112899780273, "loss_scale": 4.0, "retries": 0, "seconds": 201.5, "max_memory_gb": 15.4, "ce": 0.0455096079967916, "kl": 0.08785852276923833, "context": 0.08943886961787939}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 3.22845458984375, "loss_scale": 4.0, "retries": 0, "seconds": 202.0, "max_memory_gb": 15.4, "ce": 0.22403637832030654, "kl": 0.1995364883041475, "context": 0.32380462251603603}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.16108165681362152, "loss_scale": 4.0, "retries": 0, "seconds": 202.5, "max_memory_gb": 15.4, "ce": 0.15543171856552362, "kl": 0.05724336189450696, "context": 0.1840534028597176}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.3374859690666199, "loss_scale": 4.0, "retries": 0, "seconds": 203.0, "max_memory_gb": 15.4, "ce": 0.09340757271274924, "kl": 0.05982006684644148, "context": 0.12331760628148913}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 0.3345222771167755, "loss_scale": 4.0, "retries": 0, "seconds": 203.5, "max_memory_gb": 15.4, "ce": 0.20513678598217666, "kl": 0.007132786653528456, "context": 0.2087031793780625}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 0.7181216478347778, "loss_scale": 4.0, "retries": 0, "seconds": 203.9, "max_memory_gb": 15.4, "ce": 0.29634808585979044, "kl": 0.0645136454841122, "context": 0.32860491680912673}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.1564292162656784, "loss_scale": 4.0, "retries": 0, "seconds": 204.4, "max_memory_gb": 15.4, "ce": 0.01706345472484827, "kl": 0.004367611547422712, "context": 0.019247260875999928}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 0.24680563807487488, "loss_scale": 4.0, "retries": 0, "seconds": 204.9, "max_memory_gb": 15.4, "ce": 0.3551950855180621, "kl": 0.06527736259158701, "context": 0.38783376663923264}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.11712228506803513, "loss_scale": 4.0, "retries": 0, "seconds": 205.4, "max_memory_gb": 15.4, "ce": 0.04084077547304332, "kl": 0.03344657490379177, "context": 0.05756406509317458}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.1271049827337265, "loss_scale": 4.0, "retries": 0, "seconds": 205.8, "max_memory_gb": 15.4, "ce": 0.08341971109621227, "kl": 0.050534213456558064, "context": 0.10868681664578617}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 1.122700572013855, "loss_scale": 4.0, "retries": 0, "seconds": 206.3, "max_memory_gb": 15.4, "ce": 0.1349693704105448, "kl": 0.818486754084006, "context": 0.5442127352580428}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.19127005338668823, "loss_scale": 4.0, "retries": 0, "seconds": 206.8, "max_memory_gb": 15.4, "ce": 0.07053585117682815, "kl": 0.02225639823154779, "context": 0.08166405069641769}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 0.8145810961723328, "loss_scale": 4.0, "retries": 0, "seconds": 207.3, "max_memory_gb": 15.4, "ce": 0.20850533281918615, "kl": 0.011552602511073928, "context": 0.21428163995733485}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 1.6084110736846924, "loss_scale": 4.0, "retries": 0, "seconds": 207.8, "max_memory_gb": 15.4, "ce": 0.35774213320109993, "kl": 0.4516475321725011, "context": 0.583565890789032}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.23203997313976288, "loss_scale": 4.0, "retries": 0, "seconds": 208.3, "max_memory_gb": 15.4, "ce": 0.05627426621504128, "kl": 0.026751285477075726, "context": 0.06964990927372128}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.3083719313144684, "loss_scale": 4.0, "retries": 0, "seconds": 208.8, "max_memory_gb": 15.4, "ce": 0.10443593293894082, "kl": 0.11090004369907547, "context": 0.15988596039824188}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 0.7082985043525696, "loss_scale": 4.0, "retries": 0, "seconds": 209.3, "max_memory_gb": 15.4, "ce": 0.24460422061383724, "kl": 0.09768789121881127, "context": 0.2934481631964445}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.1262146532535553, "loss_scale": 4.0, "retries": 0, "seconds": 209.8, "max_memory_gb": 15.4, "ce": 0.06722830326179974, "kl": 0.01246316787910473, "context": 0.0734598855196964}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.1767607927322388, "loss_scale": 4.0, "retries": 0, "seconds": 210.3, "max_memory_gb": 15.4, "ce": 0.010285983677022159, "kl": 0.18368745894622407, "context": 0.10212971339933574}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.6545191407203674, "loss_scale": 4.0, "retries": 0, "seconds": 210.8, "max_memory_gb": 15.4, "ce": 0.020012946682982147, "kl": 0.12266576418187469, "context": 0.08134582941420376}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.2652541399002075, "loss_scale": 4.0, "retries": 0, "seconds": 211.3, "max_memory_gb": 15.4, "ce": 0.12089876411482692, "kl": 0.008944787288783118, "context": 0.12537116138264537}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 3.0694918632507324, "loss_scale": 4.0, "retries": 0, "seconds": 211.8, "max_memory_gb": 15.4, "ce": 0.2590917490888387, "kl": 0.2651172624609899, "context": 0.39165036543272436}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.5371381640434265, "loss_scale": 4.0, "retries": 0, "seconds": 212.3, "max_memory_gb": 15.4, "ce": 0.08969698479631916, "kl": 0.09201469423715025, "context": 0.13570433296263218}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 0.68465656042099, "loss_scale": 4.0, "retries": 0, "seconds": 212.8, "max_memory_gb": 15.4, "ce": 0.12555293971672654, "kl": 0.10156378388637677, "context": 0.17633483349345624}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.06014884635806084, "loss_scale": 4.0, "retries": 0, "seconds": 213.3, "max_memory_gb": 15.4, "ce": 0.012722826097160578, "kl": 0.0004540025292953942, "context": 0.012949827127158642}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.3750098943710327, "loss_scale": 4.0, "retries": 0, "seconds": 213.8, "max_memory_gb": 15.4, "ce": 0.11267374805174768, "kl": 0.0504027024107927, "context": 0.13787509931717068}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.3032509982585907, "loss_scale": 4.0, "retries": 0, "seconds": 214.3, "max_memory_gb": 15.4, "ce": 0.10387086449190974, "kl": 0.015512483922066167, "context": 0.11162710702046752}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 0.8507779836654663, "loss_scale": 4.0, "retries": 0, "seconds": 214.8, "max_memory_gb": 15.4, "ce": 0.13837254024110734, "kl": 0.11283709930285113, "context": 0.19479109707754105}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.5119898915290833, "loss_scale": 4.0, "retries": 0, "seconds": 215.3, "max_memory_gb": 15.4, "ce": 0.10101266868878156, "kl": 0.035158613114617765, "context": 0.11859197588637471}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.299354076385498, "loss_scale": 4.0, "retries": 0, "seconds": 215.8, "max_memory_gb": 15.4, "ce": 0.16305583529174328, "kl": 0.1994555335259065, "context": 0.2627836037427187}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.15646272897720337, "loss_scale": 4.0, "retries": 0, "seconds": 216.3, "max_memory_gb": 15.4, "ce": 0.05917502543888986, "kl": 0.057958975245128386, "context": 0.08815451501868665}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 0.9842172861099243, "loss_scale": 4.0, "retries": 0, "seconds": 216.8, "max_memory_gb": 15.4, "ce": 0.20895994850434363, "kl": 0.1879649623297155, "context": 0.3029424250125885}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.33442753553390503, "loss_scale": 4.0, "retries": 0, "seconds": 217.2, "max_memory_gb": 15.4, "ce": 0.10383913025725633, "kl": 0.03226154751610011, "context": 0.11996990302577615}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.2033429592847824, "loss_scale": 4.0, "retries": 0, "seconds": 217.7, "max_memory_gb": 15.4, "ce": 0.08736958727240562, "kl": 0.03652642388624372, "context": 0.10563279967755079}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.2383938729763031, "loss_scale": 4.0, "retries": 0, "seconds": 218.2, "max_memory_gb": 15.4, "ce": 0.13314288295805454, "kl": 0.03196162567473948, "context": 0.1491236980073154}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 0.4377128481864929, "loss_scale": 4.0, "retries": 0, "seconds": 218.6, "max_memory_gb": 15.4, "ce": 0.36589385685510933, "kl": 0.021345350749470526, "context": 0.3765665245591663}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 1.902221441268921, "loss_scale": 4.0, "retries": 0, "seconds": 219.1, "max_memory_gb": 15.4, "ce": 0.17773703765124083, "kl": 0.07820359324978199, "context": 0.21683883154764771}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.42982450127601624, "loss_scale": 4.0, "retries": 0, "seconds": 219.6, "max_memory_gb": 15.4, "ce": 0.23860060051083565, "kl": 0.018256037565151928, "context": 0.24772861553356051}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 1.1328610181808472, "loss_scale": 4.0, "retries": 0, "seconds": 220.1, "max_memory_gb": 15.4, "ce": 0.22722601145505905, "kl": 0.11723278276622295, "context": 0.2858423963189125}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.5958309173583984, "loss_scale": 4.0, "retries": 0, "seconds": 220.6, "max_memory_gb": 15.4, "ce": 0.03682544582989067, "kl": 0.018619167494762223, "context": 0.04613502841675654}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.5220016241073608, "loss_scale": 4.0, "retries": 0, "seconds": 221.1, "max_memory_gb": 15.4, "ce": 0.021361216495279223, "kl": 0.02318479821406072, "context": 0.032953615533187985}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.5352242588996887, "loss_scale": 4.0, "retries": 0, "seconds": 221.6, "max_memory_gb": 15.4, "ce": 0.031106495764106512, "kl": 0.061301621040911414, "context": 0.061757307150401175}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.1738173961639404, "loss_scale": 4.0, "retries": 0, "seconds": 222.1, "max_memory_gb": 15.4, "ce": 0.1687532674986869, "kl": 0.07223045939463191, "context": 0.20486850221641362}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.3480472266674042, "loss_scale": 4.0, "retries": 0, "seconds": 222.5, "max_memory_gb": 15.4, "ce": 0.05145600717514753, "kl": 0.025673901080153883, "context": 0.0642929570749402}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 0.6153863072395325, "loss_scale": 4.0, "retries": 0, "seconds": 223.0, "max_memory_gb": 15.4, "ce": 0.16470472572837025, "kl": 0.006102384679252282, "context": 0.16775591962505132}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 1.4789081811904907, "loss_scale": 4.0, "retries": 0, "seconds": 223.5, "max_memory_gb": 15.4, "ce": 0.1420501726679504, "kl": 0.15519236703403294, "context": 0.21964635606855154}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.18793416023254395, "loss_scale": 4.0, "retries": 0, "seconds": 224.0, "max_memory_gb": 15.4, "ce": 0.03264549281448126, "kl": 0.013105469406582415, "context": 0.03919822734314948}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.18299347162246704, "loss_scale": 4.0, "retries": 0, "seconds": 224.5, "max_memory_gb": 15.4, "ce": 0.059209213592112064, "kl": 0.025219091272447258, "context": 0.07181875954847783}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 0.44389334321022034, "loss_scale": 4.0, "retries": 0, "seconds": 225.0, "max_memory_gb": 15.4, "ce": 0.14313536300323904, "kl": 0.123640942532802, "context": 0.2049558327998966}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 0.9136226773262024, "loss_scale": 4.0, "retries": 0, "seconds": 225.5, "max_memory_gb": 15.4, "ce": 0.17226654407568276, "kl": 0.08515207809978165, "context": 0.21484258002601564}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.23500651121139526, "loss_scale": 4.0, "retries": 0, "seconds": 226.0, "max_memory_gb": 15.4, "ce": 0.03636097046546638, "kl": 0.02847604698035866, "context": 0.05059899343177676}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 0.11440568417310715, "loss_scale": 4.0, "retries": 0, "seconds": 226.5, "max_memory_gb": 15.4, "ce": 0.17583030404057354, "kl": 0.13789124169852585, "context": 0.2447759173810482}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.13883863389492035, "loss_scale": 4.0, "retries": 0, "seconds": 226.9, "max_memory_gb": 15.4, "ce": 0.040494189772289246, "kl": 0.005561201134696603, "context": 0.04327479033963755}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.14820684492588043, "loss_scale": 4.0, "retries": 0, "seconds": 227.4, "max_memory_gb": 15.4, "ce": 0.04366948641836643, "kl": 0.004262178874341771, "context": 0.04580057645216584}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.1449793428182602, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 15.4, "ce": 0.016651117359288037, "kl": 0.003541748446878046, "context": 0.01842199149541557}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 0.753603994846344, "loss_scale": 4.0, "retries": 0, "seconds": 228.4, "max_memory_gb": 15.4, "ce": 0.03497000056086108, "kl": 0.5624628332952852, "context": 0.3162014302215539}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 0.773505449295044, "loss_scale": 4.0, "retries": 0, "seconds": 228.9, "max_memory_gb": 15.4, "ce": 0.013615355303045362, "kl": 0.20720379247359233, "context": 0.11721725191455334}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.1987641155719757, "loss_scale": 4.0, "retries": 0, "seconds": 229.4, "max_memory_gb": 15.4, "ce": 0.053035560878925025, "kl": 0.02079066017176956, "context": 0.06343089090660214}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.13087588548660278, "loss_scale": 4.0, "retries": 0, "seconds": 229.8, "max_memory_gb": 15.4, "ce": 0.03861968650016934, "kl": 0.012023097311612219, "context": 0.04463123541790992}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 1.360841989517212, "loss_scale": 4.0, "retries": 0, "seconds": 230.3, "max_memory_gb": 15.4, "ce": 0.12898317584767938, "kl": 0.035053790372330695, "context": 0.14651007298380136}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 0.8559434413909912, "loss_scale": 4.0, "retries": 0, "seconds": 230.8, "max_memory_gb": 15.4, "ce": 0.006598107982426882, "kl": 0.4489944546294282, "context": 0.23109532869420946}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 0.8328222632408142, "loss_scale": 4.0, "retries": 0, "seconds": 231.3, "max_memory_gb": 15.4, "ce": 0.1553752692998387, "kl": 0.14134721978916787, "context": 0.22604888048954308}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.1766882985830307, "loss_scale": 4.0, "retries": 0, "seconds": 231.8, "max_memory_gb": 15.4, "ce": 0.014738455647602677, "kl": 0.006408663874026388, "context": 0.017942787846550345}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.5953813791275024, "loss_scale": 4.0, "retries": 0, "seconds": 232.3, "max_memory_gb": 15.4, "ce": 0.053472962230443954, "kl": 0.025747848354512826, "context": 0.06634688458871096}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.1779264509677887, "loss_scale": 4.0, "retries": 0, "seconds": 232.8, "max_memory_gb": 15.4, "ce": 0.09693447314202785, "kl": 0.015742978037451394, "context": 0.10480596427805722}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 0.3449428975582123, "loss_scale": 4.0, "retries": 0, "seconds": 233.2, "max_memory_gb": 15.4, "ce": 0.13886056374758482, "kl": 0.014002853975398466, "context": 0.14586198760662228}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.5618606209754944, "loss_scale": 4.0, "retries": 0, "seconds": 233.7, "max_memory_gb": 15.4, "ce": 0.04170947533566505, "kl": 0.007818208250682801, "context": 0.04561858042143285}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 2.305684804916382, "loss_scale": 4.0, "retries": 0, "seconds": 234.2, "max_memory_gb": 15.4, "ce": 0.01964755414519459, "kl": 0.6063974794233218, "context": 0.3228462862316519}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.2693469822406769, "loss_scale": 4.0, "retries": 0, "seconds": 234.8, "max_memory_gb": 15.4, "ce": 0.040530977421440184, "kl": 0.07222821081268194, "context": 0.07664508424932137}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 2.5081803798675537, "loss_scale": 4.0, "retries": 0, "seconds": 235.3, "max_memory_gb": 15.4, "ce": 0.38009707257151604, "kl": 0.26055086200358346, "context": 0.5103725125081837}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.261691153049469, "loss_scale": 4.0, "retries": 0, "seconds": 235.8, "max_memory_gb": 15.4, "ce": 0.2068392166402191, "kl": 0.007738077896647155, "context": 0.21070824866183102}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.25283709168434143, "loss_scale": 4.0, "retries": 0, "seconds": 236.3, "max_memory_gb": 15.4, "ce": 0.014111079042777419, "kl": 0.026229803275782615, "context": 0.027225980535149574}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.44770461320877075, "loss_scale": 4.0, "retries": 0, "seconds": 236.8, "max_memory_gb": 15.4, "ce": 0.1311174954753369, "kl": 0.014726607827469707, "context": 0.1384807974100113}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 0.7260427474975586, "loss_scale": 4.0, "retries": 0, "seconds": 237.2, "max_memory_gb": 15.4, "ce": 0.16809853236190975, "kl": 0.5239234352484345, "context": 0.4300602488219738}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 0.6317477822303772, "loss_scale": 4.0, "retries": 0, "seconds": 237.7, "max_memory_gb": 15.4, "ce": 0.4157754338812083, "kl": 0.0997625823947601, "context": 0.4656567252241075}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.4803822338581085, "loss_scale": 4.0, "retries": 0, "seconds": 238.2, "max_memory_gb": 15.4, "ce": 0.14209343306720257, "kl": 0.056643026997335255, "context": 0.17041494231671095}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 0.4721987247467041, "loss_scale": 4.0, "retries": 0, "seconds": 238.7, "max_memory_gb": 15.4, "ce": 0.153867126617115, "kl": 0.02933690301142633, "context": 0.1685355727095157}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 1.4517276287078857, "loss_scale": 4.0, "retries": 0, "seconds": 239.2, "max_memory_gb": 15.4, "ce": 0.07484985934570432, "kl": 0.05715778382727876, "context": 0.10342875146307051}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.30234459042549133, "loss_scale": 4.0, "retries": 0, "seconds": 239.7, "max_memory_gb": 15.4, "ce": 0.13889249530620873, "kl": 0.02078395092394203, "context": 0.14928446803241968}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 1.6728990077972412, "loss_scale": 4.0, "retries": 0, "seconds": 240.1, "max_memory_gb": 15.4, "ce": 0.24594060110393912, "kl": 0.0642997257236857, "context": 0.27809046604670584}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.20805735886096954, "loss_scale": 4.0, "retries": 0, "seconds": 240.6, "max_memory_gb": 15.4, "ce": 0.056178608618211, "kl": 0.022739614447345957, "context": 0.06754841608926654}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.15253351628780365, "loss_scale": 4.0, "retries": 0, "seconds": 241.1, "max_memory_gb": 15.4, "ce": 0.09669207420665771, "kl": 0.09055502816045191, "context": 0.14196958707179874}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.19552147388458252, "loss_scale": 4.0, "retries": 0, "seconds": 241.6, "max_memory_gb": 15.4, "ce": 0.02141600748291239, "kl": 0.007987861765286652, "context": 0.025409938709344715}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 0.7533615827560425, "loss_scale": 4.0, "retries": 0, "seconds": 242.0, "max_memory_gb": 15.4, "ce": 0.13013318832963705, "kl": 0.03323507867753506, "context": 0.14675072766840458}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.1994599103927612, "loss_scale": 4.0, "retries": 0, "seconds": 242.6, "max_memory_gb": 15.4, "ce": 0.01626601885072887, "kl": 0.55252151361492, "context": 0.2925267810933292}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.6104184985160828, "loss_scale": 4.0, "retries": 0, "seconds": 243.0, "max_memory_gb": 15.4, "ce": 0.0727907307446003, "kl": 0.011157926812302321, "context": 0.07836969522759318}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.8373903632164001, "loss_scale": 4.0, "retries": 0, "seconds": 243.5, "max_memory_gb": 15.4, "ce": 0.06083692307583988, "kl": 0.03726007463410497, "context": 0.07946695992723107}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 0.47408926486968994, "loss_scale": 4.0, "retries": 0, "seconds": 244.0, "max_memory_gb": 15.4, "ce": 0.13057234836742282, "kl": 0.010690437775338069, "context": 0.1359175693942234}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.3865896761417389, "loss_scale": 4.0, "retries": 0, "seconds": 244.5, "max_memory_gb": 15.4, "ce": 0.1469314838759601, "kl": 0.028998066831263714, "context": 0.1614305135444738}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 0.5020090341567993, "loss_scale": 4.0, "retries": 0, "seconds": 245.0, "max_memory_gb": 15.4, "ce": 0.11421112582320347, "kl": 0.04656281133975426, "context": 0.13749252900015563}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 0.5625962615013123, "loss_scale": 4.0, "retries": 0, "seconds": 245.4, "max_memory_gb": 15.4, "ce": 0.26704425481148064, "kl": 0.012200989527627826, "context": 0.2731447520200163}
diff --git a/runs/single_llava/context/L26_s13/adapter/README.md b/runs/single_llava/context/L26_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L26_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L26_s13/adapter/adapter_config.json b/runs/single_llava/context/L26_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9b3289f83576fa1d1c405e5ee7183c823bca5e6e
--- /dev/null
+++ b/runs/single_llava/context/L26_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.26.self_attn.k_proj",
+    "model.language_model.layers.26.mlp.up_proj",
+    "model.language_model.layers.26.self_attn.v_proj",
+    "model.language_model.layers.26.mlp.down_proj",
+    "model.language_model.layers.26.self_attn.o_proj",
+    "model.language_model.layers.26.self_attn.q_proj",
+    "model.language_model.layers.26.mlp.gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L26_s13/config.json b/runs/single_llava/context/L26_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e593e0911f0478da689f773739955934973b9624
--- /dev/null
+++ b/runs/single_llava/context/L26_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 26,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L26_s13/steps.jsonl b/runs/single_llava/context/L26_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a8c5767ca542ede468235d33fbcc18fb4459eac2
--- /dev/null
+++ b/runs/single_llava/context/L26_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.12332332134246826, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 14.83, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.23394504189491272, "loss_scale": 4.0, "retries": 0, "seconds": 1.8, "max_memory_gb": 14.84, "ce": 2.680602861335501, "kl": 0.010403205720649567, "context": 2.685804510023445}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.46363529562950134, "loss_scale": 4.0, "retries": 0, "seconds": 2.3, "max_memory_gb": 14.84, "ce": 5.392461657524109, "kl": 0.17895766515721334, "context": 5.481940448284149}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.40025773644447327, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 14.84, "ce": 3.824256087653339, "kl": 0.05090116740029771, "context": 3.849706603679806}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.01764947548508644, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 14.84, "ce": 0.08586372574791312, "kl": 0.03702217270620167, "context": 0.10437481012195349}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.17658129334449768, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 14.84, "ce": 1.4668547715991735, "kl": 0.18297943647485226, "context": 1.558344542980194}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.43249979615211487, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 14.84, "ce": 3.4409061982296407, "kl": 0.10665230075392174, "context": 3.494232313707471}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.39864498376846313, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 14.84, "ce": 3.696650341153145, "kl": 0.1606046911329031, "context": 3.776952639222145}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.3157649040222168, "loss_scale": 4.0, "retries": 0, "seconds": 4.9, "max_memory_gb": 14.84, "ce": 2.843168891966343, "kl": 0.16267576592508703, "context": 2.924506828188896}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.1794046014547348, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 14.84, "ce": 1.6342066302895546, "kl": 0.05966064194217324, "context": 1.6640369519591331}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.40503063797950745, "loss_scale": 4.0, "retries": 0, "seconds": 5.8, "max_memory_gb": 14.84, "ce": 2.632302198559046, "kl": 0.009422432754945476, "context": 2.6370134418830276}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.19512619078159332, "loss_scale": 4.0, "retries": 0, "seconds": 6.3, "max_memory_gb": 14.84, "ce": 1.539179774466902, "kl": 0.017835500300861895, "context": 1.5480975566897541}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.4340137243270874, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 14.84, "ce": 2.674172123428434, "kl": 0.007480421103537083, "context": 2.6779123097658157}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.6157382130622864, "loss_scale": 4.0, "retries": 0, "seconds": 7.1, "max_memory_gb": 14.84, "ce": 3.8987858593463898, "kl": 0.11832928168587387, "context": 3.9579505771398544}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.42225202918052673, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 14.84, "ce": 2.5924622746533714, "kl": 0.2680342357198242, "context": 2.7264793384820223}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.27750417590141296, "loss_scale": 4.0, "retries": 0, "seconds": 7.9, "max_memory_gb": 14.84, "ce": 1.3757048749830574, "kl": 0.009377270936965942, "context": 1.3803935202304274}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.24346186220645905, "loss_scale": 4.0, "retries": 0, "seconds": 8.3, "max_memory_gb": 14.84, "ce": 1.5972506538964808, "kl": 0.07828654057811946, "context": 1.6363938804715872}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.6293990015983582, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 14.85, "ce": 3.573540784418583, "kl": 0.09478936064988375, "context": 3.620935432612896}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 0.9845517873764038, "loss_scale": 4.0, "retries": 0, "seconds": 9.2, "max_memory_gb": 14.85, "ce": 5.341086149215698, "kl": 0.3352963994257152, "context": 5.508734226226807}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.007009113673120737, "loss_scale": 4.0, "retries": 0, "seconds": 9.6, "max_memory_gb": 14.85, "ce": 0.012899803230538964, "kl": 0.002172574619180523, "context": 0.013986090547405183}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.54896080493927, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 14.85, "ce": 2.780932492343709, "kl": 0.0300379304389935, "context": 2.7959514567628503}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.5431503653526306, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 14.85, "ce": 2.9453892447054386, "kl": 0.05446836445480585, "context": 2.9726233780384064}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.6272291541099548, "loss_scale": 4.0, "retries": 0, "seconds": 10.8, "max_memory_gb": 14.85, "ce": 2.6542481114156544, "kl": 0.5118376077152789, "context": 2.9101669129449874}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.1043798923492432, "loss_scale": 4.0, "retries": 0, "seconds": 11.2, "max_memory_gb": 14.85, "ce": 5.0149195194244385, "kl": 0.014376887353137136, "context": 5.02210795879364}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.0581823587417603, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 14.85, "ce": 4.237137533724308, "kl": 0.03662805026397109, "context": 4.255451560020447}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.1629273891448975, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 14.85, "ce": 5.244677126407623, "kl": 0.06526146421674639, "context": 5.277307868003845}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 0.9700148105621338, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 14.85, "ce": 4.194089841097593, "kl": 0.0843917760066688, "context": 4.236285757273436}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.723382294178009, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 14.85, "ce": 2.803161174058914, "kl": 0.23419933393597603, "context": 2.9202608168125153}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 0.8826901912689209, "loss_scale": 4.0, "retries": 0, "seconds": 13.4, "max_memory_gb": 14.85, "ce": 2.6083038155920804, "kl": 0.5225339743774384, "context": 2.869570815935731}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 0.7900645136833191, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 14.85, "ce": 2.7591950120404363, "kl": 0.03738192666787654, "context": 2.777886001393199}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 0.8960583806037903, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 14.85, "ce": 2.8491636887192726, "kl": 0.027298885281197727, "context": 2.862813077867031}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.560283899307251, "loss_scale": 4.0, "retries": 0, "seconds": 14.6, "max_memory_gb": 14.85, "ce": 5.3206857442855835, "kl": 0.014860642666462809, "context": 5.328116059303284}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 0.7748112678527832, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 14.85, "ce": 2.459372991230339, "kl": 0.18064139992929995, "context": 2.549693699926138}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 0.9506566524505615, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 14.85, "ce": 2.6174774765968323, "kl": 0.6330789821222425, "context": 2.9340169616043568}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 0.885862410068512, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 14.85, "ce": 2.7288693552836776, "kl": 0.03015909410896711, "context": 2.743948900606483}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 0.9111475348472595, "loss_scale": 4.0, "retries": 0, "seconds": 16.3, "max_memory_gb": 14.85, "ce": 2.5536642968654633, "kl": 0.006807832163758576, "context": 2.557068157941103}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 0.9551144242286682, "loss_scale": 4.0, "retries": 0, "seconds": 16.7, "max_memory_gb": 14.85, "ce": 2.4873169735074043, "kl": 0.34603066300041974, "context": 2.6603322438895702}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 1.7626088857650757, "loss_scale": 4.0, "retries": 0, "seconds": 17.1, "max_memory_gb": 14.85, "ce": 4.670456886291504, "kl": 0.4967500987841049, "context": 4.918831944465637}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.627423882484436, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 14.85, "ce": 1.4397085583768785, "kl": 0.03445811034180224, "context": 1.4569376413710415}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.1143193244934082, "loss_scale": 4.0, "retries": 0, "seconds": 17.9, "max_memory_gb": 14.85, "ce": 2.7094337418675423, "kl": 0.11973337503150105, "context": 2.7693003863096237}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.7481273412704468, "loss_scale": 4.0, "retries": 0, "seconds": 18.3, "max_memory_gb": 14.85, "ce": 1.4464619979262352, "kl": 0.17430900037288666, "context": 1.5336164459586143}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.176999807357788, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 14.85, "ce": 2.3540874887257814, "kl": 0.007034298905637115, "context": 2.3576046088710427}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 1.202771782875061, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 14.85, "ce": 2.455803452525288, "kl": 0.011508387979120016, "context": 2.4615577165968716}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 0.696028470993042, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 14.85, "ce": 1.360478007234633, "kl": 0.2749768063222291, "context": 1.4979663956910372}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 0.6794421076774597, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 14.85, "ce": 1.4072913862764835, "kl": 0.056268573622219265, "context": 1.435425728559494}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.2265585660934448, "loss_scale": 4.0, "retries": 0, "seconds": 20.3, "max_memory_gb": 14.85, "ce": 2.2775717601180077, "kl": 0.07369764457689598, "context": 2.3144205436110497}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 1.314183235168457, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 14.85, "ce": 2.018146324204281, "kl": 0.035103139554848894, "context": 2.0356979321222752}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 1.9737740755081177, "loss_scale": 4.0, "retries": 0, "seconds": 21.1, "max_memory_gb": 14.85, "ce": 3.801949344575405, "kl": 0.01751102889102185, "context": 3.810704916715622}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 1.8891268968582153, "loss_scale": 4.0, "retries": 0, "seconds": 21.5, "max_memory_gb": 14.85, "ce": 3.5625188453122973, "kl": 0.09569271071814, "context": 3.6103652147576213}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 1.9547393321990967, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 14.85, "ce": 3.620739513542503, "kl": 0.05323900090297684, "context": 3.6473589222878218}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 0.7509483098983765, "loss_scale": 4.0, "retries": 0, "seconds": 22.3, "max_memory_gb": 14.85, "ce": 1.3768870853818953, "kl": 0.13138394756242633, "context": 1.4425790188834071}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 1.8144752979278564, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 14.85, "ce": 2.3047036044299603, "kl": 0.119779851695057, "context": 2.364593528211117}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 2.062592029571533, "loss_scale": 4.0, "retries": 0, "seconds": 23.1, "max_memory_gb": 14.85, "ce": 3.708097629249096, "kl": 0.18961288151331246, "context": 3.8029041439294815}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 2.2519114017486572, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 14.85, "ce": 3.7896782383322716, "kl": 0.041425964096561074, "context": 3.8103912994265556}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 1.3234529495239258, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 14.85, "ce": 2.4692063038237393, "kl": 0.01834331599093275, "context": 2.4783779340796173}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 2.803292751312256, "loss_scale": 4.0, "retries": 0, "seconds": 24.3, "max_memory_gb": 14.85, "ce": 4.754014253616333, "kl": 0.32369176286738366, "context": 4.915860056877136}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 0.7685257792472839, "loss_scale": 4.0, "retries": 0, "seconds": 24.7, "max_memory_gb": 14.85, "ce": 1.2281424943357706, "kl": 0.10339490056503564, "context": 1.279839914292097}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 1.7341387271881104, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 14.85, "ce": 2.3536342184524983, "kl": 0.003829835870419629, "context": 2.355549128493294}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 0.9302363991737366, "loss_scale": 4.0, "retries": 0, "seconds": 25.5, "max_memory_gb": 14.85, "ce": 1.4077517092227936, "kl": 0.07353766029700637, "context": 1.4445205479860306}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 1.4791061878204346, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 14.85, "ce": 2.385358619969338, "kl": 0.014902015827829018, "context": 2.3928096373565495}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 2.6534528732299805, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 14.85, "ce": 3.4609951823949814, "kl": 0.05948895908659324, "context": 3.490739680826664}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 0.9404218792915344, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 14.85, "ce": 1.29915902717039, "kl": 0.03962727967882529, "context": 1.3189726220443845}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 2.579360008239746, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 14.85, "ce": 3.594573677983135, "kl": 0.09356177232621121, "context": 3.6413546656258404}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 2.5700156688690186, "loss_scale": 4.0, "retries": 0, "seconds": 27.5, "max_memory_gb": 14.86, "ce": 3.5759642086923122, "kl": 0.03392566030379385, "context": 3.592927061021328}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 2.534580707550049, "loss_scale": 4.0, "retries": 0, "seconds": 27.9, "max_memory_gb": 14.86, "ce": 3.607948884367943, "kl": 0.14295157208107412, "context": 3.6794247031211853}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 0.7765852212905884, "loss_scale": 4.0, "retries": 0, "seconds": 28.4, "max_memory_gb": 14.86, "ce": 0.8921743790851906, "kl": 0.062407632955000736, "context": 0.9233782092342153}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 3.015902042388916, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 14.86, "ce": 3.319433569908142, "kl": 0.017171106839668937, "context": 3.3280190601944923}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.0061955451965332, "loss_scale": 4.0, "retries": 0, "seconds": 29.2, "max_memory_gb": 14.86, "ce": 1.157428574282676, "kl": 0.04500900022685528, "context": 1.1799330273643136}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 1.9212249517440796, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 14.86, "ce": 2.184123534709215, "kl": 0.012644160247873515, "context": 2.1904455958865583}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 2.20361590385437, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 14.86, "ce": 2.1664085160009563, "kl": 0.003960036934586242, "context": 2.1683885413222015}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 3.048189878463745, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 14.86, "ce": 3.2290687561035156, "kl": 0.011621454090345651, "context": 3.234879456460476}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 3.7114551067352295, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 14.86, "ce": 3.7597779631614685, "kl": 0.149268320761621, "context": 3.834412097930908}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.0525517463684082, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 14.86, "ce": 0.22808195091784, "kl": 0.036482314113527536, "context": 0.24632310960441828}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.230808734893799, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 14.86, "ce": 2.1630573393777013, "kl": 0.02963443601038307, "context": 2.17787456093356}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.4658095836639404, "loss_scale": 4.0, "retries": 0, "seconds": 32.0, "max_memory_gb": 14.86, "ce": 1.56290914863348, "kl": 0.19876295514404774, "context": 1.6622906625270844}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 2.2558510303497314, "loss_scale": 4.0, "retries": 0, "seconds": 32.4, "max_memory_gb": 14.86, "ce": 2.0097038112580776, "kl": 0.48938156105577946, "context": 2.254394616931677}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 3.323899507522583, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 14.86, "ce": 3.3584249019622803, "kl": 0.041082406882196665, "context": 3.378966122865677}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 2.358934164047241, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 14.86, "ce": 1.6450102580711246, "kl": 0.07198881948716007, "context": 1.6810046853497624}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.4220333099365234, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 14.86, "ce": 0.8577000321820378, "kl": 0.07826744178964873, "context": 0.8968337294645607}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 2.691370725631714, "loss_scale": 4.0, "retries": 0, "seconds": 34.0, "max_memory_gb": 14.86, "ce": 1.8871476592030376, "kl": 0.007871304638683796, "context": 1.8910833230474964}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 3.6851797103881836, "loss_scale": 4.0, "retries": 0, "seconds": 34.4, "max_memory_gb": 14.86, "ce": 2.793612912297249, "kl": 0.04832051001721993, "context": 2.8177731335163116}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 3.8295297622680664, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 14.86, "ce": 2.850356437265873, "kl": 0.06492467131465673, "context": 2.882818818092346}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 3.6161794662475586, "loss_scale": 4.0, "retries": 0, "seconds": 35.2, "max_memory_gb": 14.86, "ce": 2.333339523524046, "kl": 0.21672577410936356, "context": 2.44170237518847}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.4289456605911255, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 14.86, "ce": 1.0290675068972632, "kl": 0.07764607283752412, "context": 1.0678905581589788}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.05709044635295868, "loss_scale": 4.0, "retries": 0, "seconds": 36.0, "max_memory_gb": 14.86, "ce": 0.07108031259849668, "kl": 0.03659725400939351, "context": 0.08937893970869482}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 4.183473587036133, "loss_scale": 4.0, "retries": 0, "seconds": 36.4, "max_memory_gb": 14.86, "ce": 2.6323889903724194, "kl": 0.20938541466603056, "context": 2.737081665545702}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 4.202709197998047, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 14.86, "ce": 2.566464889794588, "kl": 0.06534869084134698, "context": 2.599139243364334}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.9002394676208496, "loss_scale": 4.0, "retries": 0, "seconds": 37.2, "max_memory_gb": 14.86, "ce": 1.7633503447286785, "kl": 0.02323542496014852, "context": 1.7749680445995182}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.026233991608023643, "loss_scale": 4.0, "retries": 0, "seconds": 37.6, "max_memory_gb": 14.86, "ce": 0.11262545827776194, "kl": 0.005874518989003263, "context": 0.11556271882727742}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.5274946689605713, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 14.86, "ce": 1.4438824362587184, "kl": 0.019382026861421764, "context": 1.45357346534729}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.08723508566617966, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 14.86, "ce": 0.33328690379858017, "kl": 0.0934995433781296, "context": 0.3800366874784231}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.8622651100158691, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 14.86, "ce": 0.8281981715699658, "kl": 0.0800185666885227, "context": 0.8682074276730418}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 3.8888354301452637, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 14.86, "ce": 1.3744335998781025, "kl": 0.0752783031493891, "context": 1.4120727559784427}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 5.061584949493408, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 14.86, "ce": 2.389955960214138, "kl": 0.10456970427185297, "context": 2.442240834236145}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 3.3883559703826904, "loss_scale": 4.0, "retries": 0, "seconds": 40.0, "max_memory_gb": 14.86, "ce": 1.630064913071692, "kl": 0.06016419758088887, "context": 1.660146988928318}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 5.301230430603027, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 14.86, "ce": 2.3121387660503387, "kl": 0.03734435120713897, "context": 2.330810897052288}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 1.9189989566802979, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 14.86, "ce": 0.9567707013338804, "kl": 0.021129996981471777, "context": 0.9673357121646404}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 3.6916885375976562, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 14.86, "ce": 1.3184102678205818, "kl": 0.05555128579180746, "context": 1.3461858620867133}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.12304125726222992, "loss_scale": 4.0, "retries": 0, "seconds": 41.7, "max_memory_gb": 14.86, "ce": 0.23051046580076218, "kl": 0.06866415287368, "context": 0.26484254701063037}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 5.195813179016113, "loss_scale": 4.0, "retries": 0, "seconds": 42.1, "max_memory_gb": 14.86, "ce": 2.0006625317037106, "kl": 0.10819095512852073, "context": 2.054757997393608}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 5.678652286529541, "loss_scale": 4.0, "retries": 0, "seconds": 42.5, "max_memory_gb": 14.86, "ce": 1.832347895950079, "kl": 0.3898261534050107, "context": 2.027260959148407}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 5.444009304046631, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 14.86, "ce": 2.1303477101027966, "kl": 0.13623351976275444, "context": 2.1984644904732704}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 3.7737109661102295, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 14.86, "ce": 1.259930843487382, "kl": 0.0903013312490657, "context": 1.3050814997404814}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.2478582859039307, "loss_scale": 4.0, "retries": 0, "seconds": 43.7, "max_memory_gb": 14.86, "ce": 0.634408614016138, "kl": 0.036086498061195016, "context": 0.6524518933147192}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 5.82080078125, "loss_scale": 4.0, "retries": 0, "seconds": 44.1, "max_memory_gb": 14.86, "ce": 1.5994733329862356, "kl": 0.21116757130948827, "context": 1.7050571367144585}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 5.974542617797852, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 14.86, "ce": 1.4214439317584038, "kl": 0.06425520591437817, "context": 1.4535715356469154}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.2213194370269775, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 14.86, "ce": 0.6264569973573089, "kl": 0.016453374846605584, "context": 0.6346836751326919}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.0090324878692627, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 14.86, "ce": 0.643328276462853, "kl": 0.02647913683904335, "context": 0.6565678289625794}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 6.0326666831970215, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 14.86, "ce": 1.5675932168960571, "kl": 0.4417544365278445, "context": 1.7884704619646072}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 4.351017475128174, "loss_scale": 4.0, "retries": 0, "seconds": 46.2, "max_memory_gb": 14.86, "ce": 1.0207612812519073, "kl": 0.04205386486137286, "context": 1.0417882204055786}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 3.7034151554107666, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 14.86, "ce": 0.784179279115051, "kl": 0.0745821729178715, "context": 0.8214703747071326}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 4.1738667488098145, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 14.86, "ce": 0.745515363290906, "kl": 0.016256508242804557, "context": 0.7536435956135392}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 4.222293853759766, "loss_scale": 4.0, "retries": 0, "seconds": 47.5, "max_memory_gb": 14.86, "ce": 0.8842397108674049, "kl": 0.04234947613440454, "context": 0.9054144229739904}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 5.9862565994262695, "loss_scale": 4.0, "retries": 0, "seconds": 48.0, "max_memory_gb": 14.86, "ce": 1.0782015472650528, "kl": 0.7610868697520345, "context": 1.4587449431419373}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 4.156919479370117, "loss_scale": 4.0, "retries": 0, "seconds": 48.4, "max_memory_gb": 14.86, "ce": 0.833581268787384, "kl": 0.2816760349087417, "context": 0.9744192957878113}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.0548131987452507, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 14.86, "ce": 0.11834166245535016, "kl": 0.0059113140450790524, "context": 0.12129732314497232}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 3.705371141433716, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 14.86, "ce": 0.771361027029343, "kl": 0.011439861496910453, "context": 0.7770809709327295}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 2.0023584365844727, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 14.86, "ce": 0.3980788323096931, "kl": 0.0483750224520918, "context": 0.4222663436084986}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 1.8390777111053467, "loss_scale": 4.0, "retries": 0, "seconds": 50.2, "max_memory_gb": 14.86, "ce": 0.2754686512053013, "kl": 0.0708365504979156, "context": 0.3108869297429919}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 3.198719024658203, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 14.86, "ce": 0.5801940555684268, "kl": 0.06450054771266878, "context": 0.6124443267472088}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 1.7225022315979004, "loss_scale": 4.0, "retries": 0, "seconds": 51.1, "max_memory_gb": 14.86, "ce": 0.44811947364360094, "kl": 0.08503939404909033, "context": 0.490639163646847}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 2.2606608867645264, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 14.86, "ce": 0.3867028006352484, "kl": 0.026138757879380137, "context": 0.3997721925843507}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.8770610094070435, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 14.86, "ce": 0.6793318968266249, "kl": 0.05713346748962067, "context": 0.7078986298292875}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 0.8013256788253784, "loss_scale": 4.0, "retries": 0, "seconds": 52.3, "max_memory_gb": 14.86, "ce": 0.27881786134094, "kl": 0.03202026279177517, "context": 0.29482800140976906}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 1.250646948814392, "loss_scale": 4.0, "retries": 0, "seconds": 52.7, "max_memory_gb": 14.86, "ce": 0.22843723557889462, "kl": 0.020082420436665416, "context": 0.23847845196723938}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 2.9366800785064697, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 14.86, "ce": 0.5636624097824097, "kl": 0.12056917324662209, "context": 0.6239470094442368}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 1.1622705459594727, "loss_scale": 4.0, "retries": 0, "seconds": 53.5, "max_memory_gb": 14.86, "ce": 0.12291526328772306, "kl": 0.004424967741215369, "context": 0.12512775138020515}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.6039613485336304, "loss_scale": 4.0, "retries": 0, "seconds": 53.9, "max_memory_gb": 14.86, "ce": 0.18971983762457967, "kl": 0.4203972940304084, "context": 0.399918481009081}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 2.515667200088501, "loss_scale": 4.0, "retries": 0, "seconds": 54.3, "max_memory_gb": 14.86, "ce": 0.21903464943170547, "kl": 0.09413848092663102, "context": 0.2661038860678673}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 1.3922696113586426, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 14.86, "ce": 0.3642941862344742, "kl": 0.4369865778135136, "context": 0.5827874829992652}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 0.7869998812675476, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 14.86, "ce": 0.26800480484962463, "kl": 0.04537412244826555, "context": 0.290691863745451}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 2.4103851318359375, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 14.86, "ce": 0.4714001202955842, "kl": 0.2597026199218817, "context": 0.6012514191679657}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 1.8351482152938843, "loss_scale": 4.0, "retries": 0, "seconds": 55.9, "max_memory_gb": 14.86, "ce": 0.2825442887842655, "kl": 0.13817262323573232, "context": 0.3516306020319462}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.1408354490995407, "loss_scale": 4.0, "retries": 0, "seconds": 56.3, "max_memory_gb": 14.86, "ce": 0.21106794103980064, "kl": 0.015499685803661123, "context": 0.21881778119131923}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 1.1005357503890991, "loss_scale": 4.0, "retries": 0, "seconds": 56.7, "max_memory_gb": 14.86, "ce": 0.17975988052785397, "kl": 0.021565999719314277, "context": 0.19054288230836391}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.5927760004997253, "loss_scale": 4.0, "retries": 0, "seconds": 57.1, "max_memory_gb": 14.86, "ce": 0.18431011494249105, "kl": 0.04393989173695445, "context": 0.20628006383776665}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.5184247493743896, "loss_scale": 4.0, "retries": 0, "seconds": 57.5, "max_memory_gb": 14.86, "ce": 0.13677859120070934, "kl": 0.022101685768575408, "context": 0.14782943669706583}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.755721926689148, "loss_scale": 4.0, "retries": 0, "seconds": 57.9, "max_memory_gb": 14.86, "ce": 0.16248379088938236, "kl": 0.058890402084216475, "context": 0.19192898832261562}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 1.3307710886001587, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 14.86, "ce": 0.21662602201104164, "kl": 0.12550286971963942, "context": 0.27937744557857513}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.2619144916534424, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 14.86, "ce": 0.09760015551000834, "kl": 0.023212529544252902, "context": 0.10920642130076885}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.33161288499832153, "loss_scale": 4.0, "retries": 0, "seconds": 59.3, "max_memory_gb": 14.86, "ce": 0.14757915725931525, "kl": 0.013138852547854185, "context": 0.1541485833004117}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.6362044215202332, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 14.86, "ce": 0.17051101010292768, "kl": 0.016689824813511223, "context": 0.17885592253878713}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.0791490450501442, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 14.86, "ce": 0.13665472599677742, "kl": 0.05874057256733067, "context": 0.16602500900626183}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.945878267288208, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 14.86, "ce": 0.2568017356097698, "kl": 0.11183800443541259, "context": 0.3127207402139902}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.2937999963760376, "loss_scale": 4.0, "retries": 0, "seconds": 61.1, "max_memory_gb": 14.86, "ce": 0.047929061111062765, "kl": 0.002560171160439495, "context": 0.049209145829081535}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.3814282715320587, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 14.86, "ce": 0.14434399269521236, "kl": 0.01541707431897521, "context": 0.15205253288149834}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.6951006650924683, "loss_scale": 4.0, "retries": 0, "seconds": 62.0, "max_memory_gb": 14.86, "ce": 0.1869463638868183, "kl": 0.35582877637352794, "context": 0.36486076237633824}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.24378900229930878, "loss_scale": 4.0, "retries": 0, "seconds": 62.5, "max_memory_gb": 14.86, "ce": 0.031714130891487, "kl": 0.09955849533434957, "context": 0.08149337791837752}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.25562986731529236, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 14.86, "ce": 0.12143762107007205, "kl": 0.009853013063548133, "context": 0.12636413052678108}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.4335353672504425, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 14.86, "ce": 0.21135966386646032, "kl": 0.017727497906889766, "context": 0.22022341191768646}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.423259437084198, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 14.86, "ce": 0.14493471407331526, "kl": 0.015696402682806365, "context": 0.15278291702270508}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 0.6552568674087524, "loss_scale": 4.0, "retries": 0, "seconds": 64.3, "max_memory_gb": 14.86, "ce": 0.0827289205044508, "kl": 0.5274344627105165, "context": 0.34644615231081843}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.5711323022842407, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 14.86, "ce": 0.06546642351895571, "kl": 0.2134758567553945, "context": 0.1722043501213193}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.5708591938018799, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 14.86, "ce": 0.4917166749946773, "kl": 0.3020908564940328, "context": 0.6427621184848249}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 3.13407564163208, "loss_scale": 4.0, "retries": 0, "seconds": 65.7, "max_memory_gb": 14.86, "ce": 0.4674990847706795, "kl": 0.1752371225738898, "context": 0.5551176462322474}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.5576255917549133, "loss_scale": 4.0, "retries": 0, "seconds": 66.2, "max_memory_gb": 14.86, "ce": 0.15488056000322104, "kl": 0.17244850977294846, "context": 0.24110481142997742}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.36096253991127014, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 14.86, "ce": 0.16972647001966834, "kl": 0.045180905057350174, "context": 0.19231692096218467}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.3591543138027191, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 14.86, "ce": 0.2599944192916155, "kl": 0.3153904133941978, "context": 0.417689623311162}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.19899499416351318, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 14.86, "ce": 0.06283922027796507, "kl": 0.07843212189618498, "context": 0.10205528046935797}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.6003887057304382, "loss_scale": 4.0, "retries": 0, "seconds": 68.0, "max_memory_gb": 14.86, "ce": 0.11874912586063147, "kl": 0.0831076973117888, "context": 0.16030297335237265}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.41465747356414795, "loss_scale": 4.0, "retries": 0, "seconds": 68.4, "max_memory_gb": 14.86, "ce": 0.22441355837509036, "kl": 0.0484795916127041, "context": 0.2486533517949283}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.1286214292049408, "loss_scale": 4.0, "retries": 0, "seconds": 68.8, "max_memory_gb": 14.86, "ce": 0.10811242274940014, "kl": 0.025799940980505198, "context": 0.12101239338517189}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 1.0708943605422974, "loss_scale": 4.0, "retries": 0, "seconds": 69.2, "max_memory_gb": 14.86, "ce": 0.08622635994106531, "kl": 0.08081173177015444, "context": 0.1266322247684002}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.8442350029945374, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 14.86, "ce": 0.27100640535354614, "kl": 0.02314598928205669, "context": 0.2825793996453285}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.20642422139644623, "loss_scale": 4.0, "retries": 0, "seconds": 70.1, "max_memory_gb": 14.86, "ce": 0.17916590441018343, "kl": 0.05180733757879352, "context": 0.20506957452744246}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.9033706188201904, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 14.86, "ce": 0.1959656896069646, "kl": 0.06311823619762436, "context": 0.2275248058140278}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.2868647873401642, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 14.86, "ce": 0.21134142577648163, "kl": 0.010316807893104851, "context": 0.21649983152747154}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.46008795499801636, "loss_scale": 4.0, "retries": 0, "seconds": 71.3, "max_memory_gb": 14.86, "ce": 0.09007830591872334, "kl": 0.07060168421594426, "context": 0.12537914467975497}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.2556804418563843, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 14.86, "ce": 0.22006045142188668, "kl": 1.0136792638659244, "context": 0.7269000851083547}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.11880093812942505, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 14.86, "ce": 0.02853485196828842, "kl": 0.005503054017026443, "context": 0.031286379089578986}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 0.7324103116989136, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 14.86, "ce": 0.05353431717958301, "kl": 0.27869093435583636, "context": 0.19287978345528245}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.6913730502128601, "loss_scale": 4.0, "retries": 0, "seconds": 73.0, "max_memory_gb": 14.86, "ce": 0.09889509552158415, "kl": 0.008321635774336755, "context": 0.10305591556243598}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 0.5799077153205872, "loss_scale": 4.0, "retries": 0, "seconds": 73.4, "max_memory_gb": 14.86, "ce": 0.121045928215608, "kl": 0.09553391876397654, "context": 0.16881289309822023}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.729345440864563, "loss_scale": 4.0, "retries": 0, "seconds": 73.8, "max_memory_gb": 14.86, "ce": 0.09362281486392021, "kl": 0.05686535174027085, "context": 0.122055490501225}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.31748461723327637, "loss_scale": 4.0, "retries": 0, "seconds": 74.2, "max_memory_gb": 14.86, "ce": 0.20042138313874602, "kl": 0.15536490455269814, "context": 0.27810383308678865}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.35453304648399353, "loss_scale": 4.0, "retries": 0, "seconds": 74.6, "max_memory_gb": 14.86, "ce": 0.03881831280887127, "kl": 0.010603313072351739, "context": 0.04411996994167566}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 1.101237177848816, "loss_scale": 4.0, "retries": 0, "seconds": 75.0, "max_memory_gb": 14.86, "ce": 0.2082765290979296, "kl": 0.11158099316526204, "context": 0.2640670295804739}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.22209510207176208, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 14.86, "ce": 0.19804325280711055, "kl": 0.07154012468527071, "context": 0.2338133156299591}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.18004943430423737, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 14.86, "ce": 0.3306217794306576, "kl": 0.004892182449111715, "context": 0.33306787721812725}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 1.129894495010376, "loss_scale": 4.0, "retries": 0, "seconds": 76.3, "max_memory_gb": 14.86, "ce": 0.2633523126132786, "kl": 0.045823994441889226, "context": 0.2862643012776971}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 0.6715878844261169, "loss_scale": 4.0, "retries": 0, "seconds": 76.7, "max_memory_gb": 14.86, "ce": 0.0695124575868249, "kl": 0.04348658677190542, "context": 0.0912557509727776}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 0.37043318152427673, "loss_scale": 4.0, "retries": 0, "seconds": 77.1, "max_memory_gb": 14.86, "ce": 0.1488216312136501, "kl": 0.004479163981159218, "context": 0.15106121450662613}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.10862946510314941, "loss_scale": 4.0, "retries": 0, "seconds": 77.5, "max_memory_gb": 14.86, "ce": 0.018113356083631516, "kl": 0.0059741017284977715, "context": 0.021100407000631094}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.591648280620575, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 14.86, "ce": 0.05538905563298613, "kl": 0.09289547290245537, "context": 0.10183678800240159}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.16529682278633118, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 14.86, "ce": 0.03469602600671351, "kl": 0.01200862859695917, "context": 0.04070034041069448}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.2817175090312958, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 14.86, "ce": 0.06094674160704017, "kl": 0.005331005988409743, "context": 0.06361224502325058}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 1.186899185180664, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 14.86, "ce": 0.10313368891365826, "kl": 0.3280118828115519, "context": 0.2671396341174841}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.10752471536397934, "loss_scale": 4.0, "retries": 0, "seconds": 79.7, "max_memory_gb": 14.86, "ce": 0.1510383440181613, "kl": 0.043286747895763256, "context": 0.17268172453623265}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.19564490020275116, "loss_scale": 4.0, "retries": 0, "seconds": 80.1, "max_memory_gb": 14.86, "ce": 0.04363915231078863, "kl": 0.008330881129950285, "context": 0.04780459264293313}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.238793134689331, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 14.86, "ce": 0.021054536337032914, "kl": 0.33415929443435743, "context": 0.18813417991623282}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.1049714982509613, "loss_scale": 4.0, "retries": 0, "seconds": 80.9, "max_memory_gb": 14.86, "ce": 0.2571866975631565, "kl": 0.04270326904952526, "context": 0.2785383276641369}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.1881745606660843, "loss_scale": 4.0, "retries": 0, "seconds": 81.3, "max_memory_gb": 14.86, "ce": 0.18749780394136906, "kl": 0.0054563046141993254, "context": 0.19022596115246415}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.4025910794734955, "loss_scale": 4.0, "retries": 0, "seconds": 81.7, "max_memory_gb": 14.86, "ce": 0.042267514392733574, "kl": 0.03393454267643392, "context": 0.059234785847365856}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 0.30472904443740845, "loss_scale": 4.0, "retries": 0, "seconds": 82.1, "max_memory_gb": 14.86, "ce": 0.28478442039340734, "kl": 0.017120667966082692, "context": 0.29334474354982376}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.4585702419281006, "loss_scale": 4.0, "retries": 0, "seconds": 82.5, "max_memory_gb": 14.86, "ce": 0.1079822015017271, "kl": 0.0188881506328471, "context": 0.11742627387866378}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.49464601278305054, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 14.86, "ce": 0.037712224293500185, "kl": 0.023188991806819104, "context": 0.049306720960885286}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.2764608860015869, "loss_scale": 4.0, "retries": 0, "seconds": 83.4, "max_memory_gb": 14.86, "ce": 0.14336220128461719, "kl": 0.025022085654200055, "context": 0.15587324742227793}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.2183314859867096, "loss_scale": 4.0, "retries": 0, "seconds": 83.8, "max_memory_gb": 14.86, "ce": 0.18878861935809255, "kl": 0.04396665055537596, "context": 0.21077195089310408}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 1.0293376445770264, "loss_scale": 4.0, "retries": 0, "seconds": 84.2, "max_memory_gb": 14.86, "ce": 0.10216608922928572, "kl": 0.1977929868589854, "context": 0.2010625833645463}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.21417026221752167, "loss_scale": 4.0, "retries": 0, "seconds": 84.6, "max_memory_gb": 14.86, "ce": 0.2519572963938117, "kl": 0.010232401895336807, "context": 0.2570734918117523}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.14570333063602448, "loss_scale": 4.0, "retries": 0, "seconds": 85.0, "max_memory_gb": 14.86, "ce": 0.12469121068716049, "kl": 0.011066910097724758, "context": 0.13022466539405286}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.061788562685251236, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 14.86, "ce": 0.011687512043863535, "kl": 0.007157158222980797, "context": 0.01526609156280756}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.46096569299697876, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 14.86, "ce": 0.17702705296687782, "kl": 0.07522254437208176, "context": 0.21463832911103964}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 1.096858263015747, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 14.86, "ce": 0.1387247338425368, "kl": 0.026205419562757015, "context": 0.1518274408299476}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.9387247562408447, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 14.86, "ce": 0.08990682486910373, "kl": 0.3515495629835641, "context": 0.26568161125760525}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 1.245975375175476, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 14.86, "ce": 0.08749110111966729, "kl": 0.05820536737155635, "context": 0.11659378698095679}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 0.5054436326026917, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 14.86, "ce": 0.5746670356020331, "kl": 0.40801910776644945, "context": 0.7786766174249351}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.18282362818717957, "loss_scale": 4.0, "retries": 0, "seconds": 87.9, "max_memory_gb": 14.86, "ce": 0.024875573813915253, "kl": 0.006988097258727066, "context": 0.02836962230503559}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.2135862112045288, "loss_scale": 4.0, "retries": 0, "seconds": 88.3, "max_memory_gb": 14.86, "ce": 0.22102563455700874, "kl": 0.013216127175837755, "context": 0.2276336986105889}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.42351073026657104, "loss_scale": 4.0, "retries": 0, "seconds": 88.7, "max_memory_gb": 14.86, "ce": 0.208617759286426, "kl": 0.057205638477171306, "context": 0.2372205756837502}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.6363063454627991, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 14.86, "ce": 0.1993057569488883, "kl": 0.09107982442947105, "context": 0.24484567064791918}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.7743487358093262, "loss_scale": 4.0, "retries": 0, "seconds": 89.6, "max_memory_gb": 14.86, "ce": 0.18965907394886017, "kl": 0.647108927834779, "context": 0.5132135227322578}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.2193409949541092, "loss_scale": 4.0, "retries": 0, "seconds": 90.0, "max_memory_gb": 14.86, "ce": 0.2775605358183384, "kl": 0.04692998775863089, "context": 0.3010255414992571}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 1.058018684387207, "loss_scale": 4.0, "retries": 0, "seconds": 90.4, "max_memory_gb": 14.86, "ce": 0.2658189092762768, "kl": 0.2639643047004938, "context": 0.39780106535181403}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 0.25195249915122986, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 14.86, "ce": 0.35660693142563105, "kl": 0.004979242468834855, "context": 0.3590965492185205}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.0876518487930298, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 14.86, "ce": 0.29891391610726714, "kl": 0.1282932106696535, "context": 0.3630605237558484}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.20286622643470764, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 14.86, "ce": 0.07687100116163492, "kl": 0.03780293795716716, "context": 0.09577247244305909}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.42588385939598083, "loss_scale": 4.0, "retries": 0, "seconds": 92.0, "max_memory_gb": 14.86, "ce": 0.18510143063031137, "kl": 0.36641900846734643, "context": 0.36831092927604914}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.6842132210731506, "loss_scale": 4.0, "retries": 0, "seconds": 92.4, "max_memory_gb": 14.86, "ce": 0.13833999261260033, "kl": 0.028936372546013445, "context": 0.15280817728489637}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.998315155506134, "loss_scale": 4.0, "retries": 0, "seconds": 92.9, "max_memory_gb": 14.86, "ce": 0.09474569978192449, "kl": 0.024590682587586343, "context": 0.10704103810712695}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 1.0902968645095825, "loss_scale": 4.0, "retries": 0, "seconds": 93.3, "max_memory_gb": 14.86, "ce": 0.13030695845372975, "kl": 0.06806523707928136, "context": 0.1643395759165287}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 3.2382285594940186, "loss_scale": 4.0, "retries": 0, "seconds": 93.7, "max_memory_gb": 14.86, "ce": 0.30020557064563036, "kl": 0.3478372210156522, "context": 0.47412416618317366}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.16033345460891724, "loss_scale": 4.0, "retries": 0, "seconds": 94.1, "max_memory_gb": 14.86, "ce": 0.2320493645966053, "kl": 0.03702140226960182, "context": 0.2505600587464869}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.5937029123306274, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 14.86, "ce": 0.10351042170077562, "kl": 0.09186960139777511, "context": 0.14944522129371762}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.2745678424835205, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 14.86, "ce": 0.09467177791520953, "kl": 0.039731607772409916, "context": 0.11453758180141449}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.1965664178133011, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 14.86, "ce": 0.14205529633909464, "kl": 0.021577789913862944, "context": 0.15284419525414705}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.19719216227531433, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 14.86, "ce": 0.0212572606978938, "kl": 0.006564503099070862, "context": 0.02453951258212328}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.2146940678358078, "loss_scale": 4.0, "retries": 0, "seconds": 96.1, "max_memory_gb": 14.86, "ce": 0.12038447719532996, "kl": 0.06769807077944279, "context": 0.15423351153731346}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.21410906314849854, "loss_scale": 4.0, "retries": 0, "seconds": 96.5, "max_memory_gb": 14.86, "ce": 0.2645007437095046, "kl": 0.13008612929843366, "context": 0.32954380568116903}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.38687649369239807, "loss_scale": 4.0, "retries": 0, "seconds": 96.9, "max_memory_gb": 14.86, "ce": 0.30949532613158226, "kl": 0.08537801634520292, "context": 0.35218433290719986}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.2313321977853775, "loss_scale": 4.0, "retries": 0, "seconds": 97.3, "max_memory_gb": 14.86, "ce": 0.21155879204161465, "kl": 0.020579572883434594, "context": 0.22184858145192266}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.325704425573349, "loss_scale": 4.0, "retries": 0, "seconds": 97.7, "max_memory_gb": 14.86, "ce": 0.03880567196756601, "kl": 0.008134644769597799, "context": 0.042872993275523186}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 2.5329480171203613, "loss_scale": 4.0, "retries": 0, "seconds": 98.2, "max_memory_gb": 14.86, "ce": 0.256304619833827, "kl": 0.16016946021773038, "context": 0.3363893423229456}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.19338317215442657, "loss_scale": 4.0, "retries": 0, "seconds": 98.6, "max_memory_gb": 14.86, "ce": 0.20838415995240211, "kl": 0.020201497653033584, "context": 0.21848491672426462}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.176014706492424, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 14.86, "ce": 0.27120125107467175, "kl": 0.034258132393006235, "context": 0.28833032259717584}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.3167998492717743, "loss_scale": 4.0, "retries": 0, "seconds": 99.4, "max_memory_gb": 14.86, "ce": 0.08903197711333632, "kl": 0.01917805167613551, "context": 0.09862100216560066}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.2618331909179688, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 14.86, "ce": 0.21941143879666924, "kl": 0.09464385805404163, "context": 0.2667333707213402}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.4591163098812103, "loss_scale": 4.0, "retries": 0, "seconds": 100.2, "max_memory_gb": 14.86, "ce": 0.05609769542934373, "kl": 0.1099466474406654, "context": 0.11107101972447708}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.31939229369163513, "loss_scale": 4.0, "retries": 0, "seconds": 100.6, "max_memory_gb": 14.86, "ce": 0.23283437732607126, "kl": 0.03158099245047197, "context": 0.24862486589699984}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.6404913663864136, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 14.86, "ce": 0.1341239521279931, "kl": 0.04147507157176733, "context": 0.15486148931086063}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.19530896842479706, "loss_scale": 4.0, "retries": 0, "seconds": 101.4, "max_memory_gb": 14.86, "ce": 0.22791069373488426, "kl": 0.018427782924845815, "context": 0.23712458461523056}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.3542455732822418, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 14.86, "ce": 0.15199769102036953, "kl": 0.027054100297391415, "context": 0.16552473697811365}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.5464598536491394, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 14.86, "ce": 0.031192224472761154, "kl": 0.09696643744246103, "context": 0.07967544137500226}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 0.2944677770137787, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 14.86, "ce": 0.25200571678578854, "kl": 0.0603111458622152, "context": 0.2821612828411162}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.21644185483455658, "loss_scale": 4.0, "retries": 0, "seconds": 103.1, "max_memory_gb": 14.86, "ce": 0.38368938746862113, "kl": 0.011724293930456042, "context": 0.38955154176801443}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.3071794807910919, "loss_scale": 4.0, "retries": 0, "seconds": 103.5, "max_memory_gb": 14.86, "ce": 0.03467000718228519, "kl": 0.02998643727914896, "context": 0.04966322774998844}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.2917614281177521, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 14.86, "ce": 0.11188369046431035, "kl": 0.03293571784161031, "context": 0.12835154961794615}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.5216648578643799, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 14.86, "ce": 0.08377847820520401, "kl": 0.038485749857500196, "context": 0.10302135441452265}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 0.8370845317840576, "loss_scale": 4.0, "retries": 0, "seconds": 104.7, "max_memory_gb": 14.86, "ce": 0.110457394272089, "kl": 0.030470321129541844, "context": 0.1256925519555807}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.6253888010978699, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 14.86, "ce": 0.06826136400923133, "kl": 0.013240485801361501, "context": 0.07488160557113588}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.1360202133655548, "loss_scale": 4.0, "retries": 0, "seconds": 105.5, "max_memory_gb": 14.86, "ce": 0.1651678872294724, "kl": 0.026582788297673687, "context": 0.17845928063616157}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.4644372761249542, "loss_scale": 4.0, "retries": 0, "seconds": 105.9, "max_memory_gb": 14.86, "ce": 0.047903915867209435, "kl": 0.12400250788778067, "context": 0.10990517027676105}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.11590687930583954, "loss_scale": 4.0, "retries": 0, "seconds": 106.4, "max_memory_gb": 14.86, "ce": 0.1035174885764718, "kl": 0.00794394247350283, "context": 0.10748946014791727}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 0.8343992233276367, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 14.86, "ce": 0.1941981592681259, "kl": 0.22566410759463906, "context": 0.30703020840883255}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.29994985461235046, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 14.86, "ce": 0.1992851048707962, "kl": 0.20028533693403006, "context": 0.2994277663528919}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.19646677374839783, "loss_scale": 4.0, "retries": 0, "seconds": 107.6, "max_memory_gb": 14.86, "ce": 0.10386171843856573, "kl": 0.027946059713940485, "context": 0.11783475242555141}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.1084694042801857, "loss_scale": 4.0, "retries": 0, "seconds": 108.0, "max_memory_gb": 14.86, "ce": 0.08024668041616678, "kl": 0.014235902330256067, "context": 0.08736463147215545}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 1.2135905027389526, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 14.86, "ce": 0.0962026312481612, "kl": 0.07623650738969445, "context": 0.13432088890112936}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.0647348016500473, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 14.86, "ce": 0.03498375089839101, "kl": 0.03658804811129812, "context": 0.05327777308411896}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.45862215757369995, "loss_scale": 4.0, "retries": 0, "seconds": 109.2, "max_memory_gb": 14.86, "ce": 0.14677530992776155, "kl": 0.12075048690894619, "context": 0.2071505580097437}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 1.2348921298980713, "loss_scale": 4.0, "retries": 0, "seconds": 109.6, "max_memory_gb": 14.86, "ce": 0.18748971074819565, "kl": 0.12427964457310736, "context": 0.249629532918334}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.33413708209991455, "loss_scale": 4.0, "retries": 0, "seconds": 110.1, "max_memory_gb": 14.86, "ce": 0.25541586615145206, "kl": 0.08368248934857547, "context": 0.2972571114078164}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 0.7134042382240295, "loss_scale": 4.0, "retries": 0, "seconds": 110.5, "max_memory_gb": 14.86, "ce": 0.2508504902943969, "kl": 0.03795266931410879, "context": 0.26982682291418314}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 0.28546419739723206, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 14.86, "ce": 0.19345696177333593, "kl": 0.0761213602963835, "context": 0.23151763412170112}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.4056529700756073, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 14.86, "ce": 0.10046996851451695, "kl": 0.0674184710951522, "context": 0.13417920493520796}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 2.391453504562378, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 14.86, "ce": 0.3381060790270567, "kl": 0.11085666192229837, "context": 0.39353441284038126}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.11581655591726303, "loss_scale": 4.0, "retries": 0, "seconds": 112.1, "max_memory_gb": 14.86, "ce": 0.16164438566192985, "kl": 0.006536224856972694, "context": 0.16491249995306134}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.13077495992183685, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 14.86, "ce": 0.26925481855869293, "kl": 0.0550231920205988, "context": 0.296766416169703}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 2.8105363845825195, "loss_scale": 4.0, "retries": 0, "seconds": 112.9, "max_memory_gb": 14.86, "ce": 0.36678584665060043, "kl": 0.5153303064580541, "context": 0.6244509965181351}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.2583855986595154, "loss_scale": 4.0, "retries": 0, "seconds": 113.3, "max_memory_gb": 14.86, "ce": 0.13937836792320013, "kl": 0.0083167723241786, "context": 0.14353675465099514}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 0.5199296474456787, "loss_scale": 4.0, "retries": 0, "seconds": 113.8, "max_memory_gb": 14.86, "ce": 0.08612474519759417, "kl": 0.03934510750696063, "context": 0.10579730104655027}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 0.7809803485870361, "loss_scale": 4.0, "retries": 0, "seconds": 114.1, "max_memory_gb": 14.86, "ce": 0.04975628457032144, "kl": 0.3369224569760263, "context": 0.21821751166135073}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.9206436276435852, "loss_scale": 4.0, "retries": 0, "seconds": 114.6, "max_memory_gb": 14.86, "ce": 0.03486377908848226, "kl": 0.19561461731791496, "context": 0.1326710858847946}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.8058685660362244, "loss_scale": 4.0, "retries": 0, "seconds": 115.0, "max_memory_gb": 14.86, "ce": 0.17724505066871643, "kl": 0.05506696872180328, "context": 0.2047785371541977}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 0.5065680146217346, "loss_scale": 4.0, "retries": 0, "seconds": 115.4, "max_memory_gb": 14.86, "ce": 0.4671780717908405, "kl": 0.35412463638931513, "context": 0.6442404022673145}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.127194344997406, "loss_scale": 4.0, "retries": 0, "seconds": 115.8, "max_memory_gb": 14.86, "ce": 0.009269592235796154, "kl": 0.006609170406591147, "context": 0.012574177817441523}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.5773783326148987, "loss_scale": 4.0, "retries": 0, "seconds": 116.2, "max_memory_gb": 14.86, "ce": 0.020123712718486786, "kl": 0.0860528977937065, "context": 0.06315016024746001}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 1.0141730308532715, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 14.86, "ce": 0.1561269170488231, "kl": 0.15387263813499885, "context": 0.23306323197903112}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 1.524248719215393, "loss_scale": 4.0, "retries": 0, "seconds": 117.0, "max_memory_gb": 14.86, "ce": 0.11197242327034473, "kl": 0.1071288513703621, "context": 0.16553685069084167}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.5420641899108887, "loss_scale": 4.0, "retries": 0, "seconds": 117.4, "max_memory_gb": 14.86, "ce": 0.04180307895876467, "kl": 0.01459427407826297, "context": 0.04910021601244807}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 1.876204490661621, "loss_scale": 4.0, "retries": 0, "seconds": 117.8, "max_memory_gb": 14.86, "ce": 0.23613836895674467, "kl": 0.06230688065988943, "context": 0.2672918112948537}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.12898680567741394, "loss_scale": 4.0, "retries": 0, "seconds": 118.3, "max_memory_gb": 14.86, "ce": 0.12697256123647094, "kl": 0.008163817481545266, "context": 0.1310544703155756}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.8184725642204285, "loss_scale": 4.0, "retries": 0, "seconds": 118.7, "max_memory_gb": 14.86, "ce": 0.07633817661553621, "kl": 0.1411019833321916, "context": 0.1468891684198752}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 1.2841476202011108, "loss_scale": 4.0, "retries": 0, "seconds": 119.1, "max_memory_gb": 14.86, "ce": 0.1377293241675943, "kl": 0.013930653373790847, "context": 0.1446946517098695}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.580477237701416, "loss_scale": 4.0, "retries": 0, "seconds": 119.6, "max_memory_gb": 14.86, "ce": 0.13393036741763353, "kl": 0.027775451861089095, "context": 0.147818093188107}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.23483696579933167, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 14.86, "ce": 0.018691253731958568, "kl": 0.011037749478418846, "context": 0.024210128700360656}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.31843137741088867, "loss_scale": 4.0, "retries": 0, "seconds": 120.4, "max_memory_gb": 14.86, "ce": 0.1729319137521088, "kl": 0.04342712473589927, "context": 0.1946454755961895}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.16174167394638062, "loss_scale": 4.0, "retries": 0, "seconds": 120.8, "max_memory_gb": 14.86, "ce": 0.013008664303924888, "kl": 0.007350150641286746, "context": 0.016683739377185702}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.3531430959701538, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 14.86, "ce": 0.08395931636914611, "kl": 0.061932156269904226, "context": 0.11492539243772626}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.2657270133495331, "loss_scale": 4.0, "retries": 0, "seconds": 121.7, "max_memory_gb": 14.86, "ce": 0.18179157772101462, "kl": 0.055828960925282445, "context": 0.2097060582600534}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.10560384392738342, "loss_scale": 4.0, "retries": 0, "seconds": 122.1, "max_memory_gb": 14.86, "ce": 0.009643143916036934, "kl": 0.002812351602187846, "context": 0.01104931952431798}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.1627742052078247, "loss_scale": 4.0, "retries": 0, "seconds": 122.5, "max_memory_gb": 14.86, "ce": 0.0866740990895778, "kl": 0.6548304408788681, "context": 0.4140893258154392}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.6675337553024292, "loss_scale": 4.0, "retries": 0, "seconds": 122.9, "max_memory_gb": 14.86, "ce": 0.09789186483249068, "kl": 0.10580114572076127, "context": 0.15079243760555983}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 1.0115137100219727, "loss_scale": 4.0, "retries": 0, "seconds": 123.3, "max_memory_gb": 14.86, "ce": 0.10649190284311771, "kl": 0.04260224336758256, "context": 0.12779302522540092}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 0.743915319442749, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 14.86, "ce": 0.2980225826613605, "kl": 0.07528702064882964, "context": 0.33566610282287}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 0.9662434458732605, "loss_scale": 4.0, "retries": 0, "seconds": 124.1, "max_memory_gb": 14.86, "ce": 0.21007516491226852, "kl": 0.4030697647249326, "context": 0.4116100464016199}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.07710547745227814, "loss_scale": 4.0, "retries": 0, "seconds": 124.5, "max_memory_gb": 14.86, "ce": 0.09798037307336926, "kl": 0.02162291726563126, "context": 0.10879183025099337}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 2.3704323768615723, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 14.86, "ce": 0.44178438978269696, "kl": 0.15416804020060226, "context": 0.5188684174790978}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.623009204864502, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 14.86, "ce": 0.016664859605953097, "kl": 0.09227941288554575, "context": 0.06280456599779427}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.18476682901382446, "loss_scale": 4.0, "retries": 0, "seconds": 125.7, "max_memory_gb": 14.86, "ce": 0.28091323329135776, "kl": 0.03850674405111931, "context": 0.30016660830006003}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 0.19226059317588806, "loss_scale": 4.0, "retries": 0, "seconds": 126.1, "max_memory_gb": 14.86, "ce": 0.16356081562116742, "kl": 0.10502327347057872, "context": 0.21607245225459337}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.8363515734672546, "loss_scale": 4.0, "retries": 0, "seconds": 126.5, "max_memory_gb": 14.86, "ce": 0.05510826094541699, "kl": 0.014712231932207942, "context": 0.06246437900699675}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 0.593230664730072, "loss_scale": 4.0, "retries": 0, "seconds": 127.0, "max_memory_gb": 14.86, "ce": 0.038608518545515835, "kl": 0.02019986984669231, "context": 0.04870845330879092}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.3663138747215271, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 14.86, "ce": 0.033689338248223066, "kl": 0.007238805002998561, "context": 0.037308739963918924}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 1.3003547191619873, "loss_scale": 4.0, "retries": 0, "seconds": 127.8, "max_memory_gb": 14.86, "ce": 0.12132329260930419, "kl": 0.11999775457661599, "context": 0.1813221771735698}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 0.9093329906463623, "loss_scale": 4.0, "retries": 0, "seconds": 128.2, "max_memory_gb": 14.86, "ce": 0.027408627327531576, "kl": 0.10441669775173068, "context": 0.0796169750392437}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.14210666716098785, "loss_scale": 4.0, "retries": 0, "seconds": 128.6, "max_memory_gb": 14.86, "ce": 0.07640302227810025, "kl": 0.027394147822633386, "context": 0.09010009607300162}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 0.7660863995552063, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 14.86, "ce": 0.02119613206014037, "kl": 0.5791365986224264, "context": 0.31076442264020443}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 0.1918288767337799, "loss_scale": 4.0, "retries": 0, "seconds": 129.4, "max_memory_gb": 14.86, "ce": 0.20922453387174755, "kl": 0.020810921560041606, "context": 0.21962999482639134}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.8690798878669739, "loss_scale": 4.0, "retries": 0, "seconds": 129.8, "max_memory_gb": 14.86, "ce": 0.03978225844912231, "kl": 0.1883802937809378, "context": 0.1339724026620388}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.0982004851102829, "loss_scale": 4.0, "retries": 0, "seconds": 130.2, "max_memory_gb": 14.86, "ce": 0.13174268184229732, "kl": 0.01677748483780306, "context": 0.14013142557814717}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.2517254650592804, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 14.86, "ce": 0.027733999770134687, "kl": 0.01242765854112804, "context": 0.03394782938994467}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.5416967868804932, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 14.86, "ce": 0.136465986026451, "kl": 0.04716116009512916, "context": 0.1600465690717101}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.3014695346355438, "loss_scale": 4.0, "retries": 0, "seconds": 131.5, "max_memory_gb": 14.86, "ce": 0.11745810555294156, "kl": 0.0938212055771146, "context": 0.16436870861798525}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.1956847459077835, "loss_scale": 4.0, "retries": 0, "seconds": 132.0, "max_memory_gb": 14.86, "ce": 0.07192351331468672, "kl": 0.03233188379090279, "context": 0.08808945375494659}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 0.8584772944450378, "loss_scale": 4.0, "retries": 0, "seconds": 132.4, "max_memory_gb": 14.86, "ce": 0.20675947656854987, "kl": 0.5648444606922567, "context": 0.4891817048192024}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.9412978887557983, "loss_scale": 4.0, "retries": 0, "seconds": 132.8, "max_memory_gb": 14.86, "ce": 0.2278272146359086, "kl": 0.025906568858772516, "context": 0.24078049883246422}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.6772356629371643, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 14.86, "ce": 0.044191256165504456, "kl": 0.43965865566860884, "context": 0.26402057986706495}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 1.034554123878479, "loss_scale": 4.0, "retries": 0, "seconds": 133.7, "max_memory_gb": 14.86, "ce": 0.10458512220066041, "kl": 0.09147586242761463, "context": 0.15032305382192135}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.7689467668533325, "loss_scale": 4.0, "retries": 0, "seconds": 134.2, "max_memory_gb": 14.86, "ce": 0.14967316365800798, "kl": 0.08041632833192125, "context": 0.18988132616505027}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.20748847723007202, "loss_scale": 4.0, "retries": 0, "seconds": 134.6, "max_memory_gb": 14.86, "ce": 0.12272993475198746, "kl": 0.07827928918413818, "context": 0.1618695859797299}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.4649166762828827, "loss_scale": 4.0, "retries": 0, "seconds": 135.0, "max_memory_gb": 14.86, "ce": 0.027174995746463537, "kl": 0.046682857573614456, "context": 0.0505164242349565}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 0.4851880371570587, "loss_scale": 4.0, "retries": 0, "seconds": 135.4, "max_memory_gb": 14.86, "ce": 0.027076730621047318, "kl": 0.05687337392009795, "context": 0.05551341734826565}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.3919518291950226, "loss_scale": 4.0, "retries": 0, "seconds": 135.8, "max_memory_gb": 14.86, "ce": 0.026052881497889757, "kl": 0.013333953298570123, "context": 0.03271985822357237}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.18299122154712677, "loss_scale": 4.0, "retries": 0, "seconds": 136.2, "max_memory_gb": 14.86, "ce": 0.23238631896674633, "kl": 0.013955914590042084, "context": 0.23936428118031472}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.26989445090293884, "loss_scale": 4.0, "retries": 0, "seconds": 136.6, "max_memory_gb": 14.86, "ce": 0.06055601825937629, "kl": 0.03137932872050442, "context": 0.07624568301253021}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 2.130631446838379, "loss_scale": 4.0, "retries": 0, "seconds": 137.0, "max_memory_gb": 14.86, "ce": 0.17787694209255278, "kl": 0.07640986524347682, "context": 0.2160818824195303}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.4419201612472534, "loss_scale": 4.0, "retries": 0, "seconds": 137.4, "max_memory_gb": 14.86, "ce": 0.11594724626047537, "kl": 0.12573341083407286, "context": 0.1788139515556395}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.2730843424797058, "loss_scale": 4.0, "retries": 0, "seconds": 137.8, "max_memory_gb": 14.86, "ce": 0.09183104056864977, "kl": 0.16255597191047855, "context": 0.1731090194080025}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.4558837413787842, "loss_scale": 4.0, "retries": 0, "seconds": 138.2, "max_memory_gb": 14.86, "ce": 0.16563212778419256, "kl": 0.008419463032623753, "context": 0.16984186391346157}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.810486376285553, "loss_scale": 4.0, "retries": 0, "seconds": 138.7, "max_memory_gb": 14.86, "ce": 0.27067858073860407, "kl": 0.07906886388082057, "context": 0.31021302193403244}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.706824541091919, "loss_scale": 4.0, "retries": 0, "seconds": 139.1, "max_memory_gb": 14.86, "ce": 0.17070236010476947, "kl": 0.03978661555447616, "context": 0.19059567106887698}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 0.20410668849945068, "loss_scale": 4.0, "retries": 0, "seconds": 139.5, "max_memory_gb": 14.86, "ce": 0.28360194992274046, "kl": 0.026572751346975565, "context": 0.2968883216381073}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 0.3791394531726837, "loss_scale": 4.0, "retries": 0, "seconds": 139.9, "max_memory_gb": 14.86, "ce": 0.17570461565628648, "kl": 0.09558179171290249, "context": 0.22349551133811474}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 3.3765358924865723, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 14.86, "ce": 0.353225355502218, "kl": 0.22059783013537526, "context": 0.46352426847442985}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 0.760293185710907, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 14.86, "ce": 0.041823819279670715, "kl": 0.054624150769086555, "context": 0.06913589441683143}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 1.683744192123413, "loss_scale": 4.0, "retries": 0, "seconds": 141.1, "max_memory_gb": 14.86, "ce": 0.2365923640318215, "kl": 0.1067967985291034, "context": 0.2899907575920224}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.0854448080062866, "loss_scale": 4.0, "retries": 0, "seconds": 141.5, "max_memory_gb": 14.86, "ce": 0.2010835069231689, "kl": 0.15660236682742834, "context": 0.2793846884742379}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.21130234003067017, "loss_scale": 4.0, "retries": 0, "seconds": 141.9, "max_memory_gb": 14.86, "ce": 0.13746149162761867, "kl": 0.013072382484097034, "context": 0.14399768155999482}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.4525274634361267, "loss_scale": 4.0, "retries": 0, "seconds": 142.4, "max_memory_gb": 14.86, "ce": 0.2140320788603276, "kl": 0.06670839540311135, "context": 0.24738627206534147}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 1.58884596824646, "loss_scale": 4.0, "retries": 0, "seconds": 142.8, "max_memory_gb": 14.86, "ce": 0.26078079966828227, "kl": 0.039310181979089975, "context": 0.2804358871653676}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 0.8341027498245239, "loss_scale": 4.0, "retries": 0, "seconds": 143.2, "max_memory_gb": 14.86, "ce": 0.23753205791581422, "kl": 0.8740469079930335, "context": 0.6745555065572262}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.4407322108745575, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 14.86, "ce": 0.024841996608301997, "kl": 0.048356889048591256, "context": 0.04902044124901295}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.45688071846961975, "loss_scale": 4.0, "retries": 0, "seconds": 144.0, "max_memory_gb": 14.86, "ce": 0.04139075963757932, "kl": 0.0200625802972354, "context": 0.051422049291431904}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.13713620603084564, "loss_scale": 4.0, "retries": 0, "seconds": 144.4, "max_memory_gb": 14.86, "ce": 0.14845032978337258, "kl": 0.1702855652256403, "context": 0.23359311989042908}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.38271623849868774, "loss_scale": 4.0, "retries": 0, "seconds": 144.8, "max_memory_gb": 14.86, "ce": 0.23533679405227304, "kl": 0.009733567712828517, "context": 0.24020358407869935}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 1.0237549543380737, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 14.86, "ce": 0.1894085155799985, "kl": 0.024639185052365065, "context": 0.20172810461372137}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.08506347984075546, "loss_scale": 4.0, "retries": 0, "seconds": 145.6, "max_memory_gb": 14.86, "ce": 0.013358454452827573, "kl": 0.0045702325805905275, "context": 0.015643570572137833}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.09618079662322998, "loss_scale": 4.0, "retries": 0, "seconds": 146.0, "max_memory_gb": 14.86, "ce": 0.11097777239046991, "kl": 0.017020132872858085, "context": 0.11948783695697784}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.16065803170204163, "loss_scale": 4.0, "retries": 0, "seconds": 146.5, "max_memory_gb": 14.86, "ce": 0.1608947552740574, "kl": 0.0483613763935864, "context": 0.18507544696331024}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.1448358297348022, "loss_scale": 4.0, "retries": 0, "seconds": 146.9, "max_memory_gb": 14.86, "ce": 0.03955758782103658, "kl": 0.31067658972460777, "context": 0.19489588355645537}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.3322672843933105, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 14.86, "ce": 0.2889536852017045, "kl": 0.2858799663372338, "context": 0.4318936616182327}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.19606472551822662, "loss_scale": 4.0, "retries": 0, "seconds": 147.7, "max_memory_gb": 14.86, "ce": 0.13864850043319166, "kl": 0.05569082938018255, "context": 0.16649391688406467}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 0.9558770656585693, "loss_scale": 4.0, "retries": 0, "seconds": 148.1, "max_memory_gb": 14.86, "ce": 0.21728089870885015, "kl": 0.027856608721776865, "context": 0.23120919661596417}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 1.2202746868133545, "loss_scale": 4.0, "retries": 0, "seconds": 148.5, "max_memory_gb": 14.86, "ce": 0.1826907871000003, "kl": 0.027337749022990465, "context": 0.19635965290945023}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.13656571507453918, "loss_scale": 4.0, "retries": 0, "seconds": 148.9, "max_memory_gb": 14.86, "ce": 0.05342346616089344, "kl": 0.02773258312663529, "context": 0.06728975707665086}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.3121465742588043, "loss_scale": 4.0, "retries": 0, "seconds": 149.3, "max_memory_gb": 14.86, "ce": 0.08417171728797257, "kl": 0.0275801733660046, "context": 0.09796180785633624}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 0.2184104174375534, "loss_scale": 4.0, "retries": 0, "seconds": 149.7, "max_memory_gb": 14.86, "ce": 0.22582473373040557, "kl": 0.013430666062049568, "context": 0.23254006868228316}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.3042944073677063, "loss_scale": 4.0, "retries": 0, "seconds": 150.1, "max_memory_gb": 14.86, "ce": 0.02764503820799291, "kl": 0.025955942750442773, "context": 0.040623008739203215}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.17694957554340363, "loss_scale": 4.0, "retries": 0, "seconds": 150.5, "max_memory_gb": 14.86, "ce": 0.08558681490831077, "kl": 0.018766415683785453, "context": 0.09497002069838345}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.110836386680603, "loss_scale": 4.0, "retries": 0, "seconds": 150.9, "max_memory_gb": 14.86, "ce": 0.08153290062909946, "kl": 0.3883802195341559, "context": 0.27572301332838833}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 1.7852606773376465, "loss_scale": 4.0, "retries": 0, "seconds": 151.4, "max_memory_gb": 14.86, "ce": 0.27946928632445633, "kl": 0.046516306611010805, "context": 0.30272744107060134}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.2886970341205597, "loss_scale": 4.0, "retries": 0, "seconds": 151.8, "max_memory_gb": 14.86, "ce": 0.08259084960445762, "kl": 0.06552629265934229, "context": 0.11535399314016104}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 1.8727667331695557, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 14.86, "ce": 0.01112126704538241, "kl": 0.4578044672853139, "context": 0.24002348992507905}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.24859683215618134, "loss_scale": 4.0, "retries": 0, "seconds": 152.6, "max_memory_gb": 14.86, "ce": 0.17588610923849046, "kl": 0.01904966071015224, "context": 0.18541093543171883}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.6944988369941711, "loss_scale": 4.0, "retries": 0, "seconds": 153.0, "max_memory_gb": 14.86, "ce": 0.07743036141619086, "kl": 0.27830940805142745, "context": 0.21658506197854877}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.22729234397411346, "loss_scale": 4.0, "retries": 0, "seconds": 153.4, "max_memory_gb": 14.86, "ce": 0.1416954395826906, "kl": 0.11143118483596481, "context": 0.1974110242445022}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.5213542580604553, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 14.86, "ce": 0.02827541925944388, "kl": 0.03358867262249987, "context": 0.04506975528784096}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.189531609416008, "loss_scale": 4.0, "retries": 0, "seconds": 154.2, "max_memory_gb": 14.86, "ce": 0.0694353700382635, "kl": 0.012316355074290186, "context": 0.07559354836121202}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 0.5597993731498718, "loss_scale": 4.0, "retries": 0, "seconds": 154.6, "max_memory_gb": 14.86, "ce": 0.42450636625289917, "kl": 0.15631947643123567, "context": 0.502666100859642}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.6476411819458008, "loss_scale": 4.0, "retries": 0, "seconds": 155.0, "max_memory_gb": 14.86, "ce": 0.12406664772424847, "kl": 0.024424980001640506, "context": 0.13627913582604378}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.21489949524402618, "loss_scale": 4.0, "retries": 0, "seconds": 155.4, "max_memory_gb": 14.86, "ce": 0.04999258951283991, "kl": 0.02484312347951345, "context": 0.06241415115073323}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.13884729146957397, "loss_scale": 4.0, "retries": 0, "seconds": 155.8, "max_memory_gb": 14.86, "ce": 0.10505622811615467, "kl": 0.042913456505630165, "context": 0.12651295540854335}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.34123075008392334, "loss_scale": 4.0, "retries": 0, "seconds": 156.3, "max_memory_gb": 14.86, "ce": 0.18407791620120406, "kl": 0.10308144458758761, "context": 0.235618646023795}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 1.2088724374771118, "loss_scale": 4.0, "retries": 0, "seconds": 156.7, "max_memory_gb": 14.86, "ce": 0.19725731248036027, "kl": 0.1522790889866883, "context": 0.2733968519605696}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.3078901767730713, "loss_scale": 4.0, "retries": 0, "seconds": 157.1, "max_memory_gb": 14.86, "ce": 0.13739835261367261, "kl": 0.0382101968571078, "context": 0.15650345454923809}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.15660808980464935, "loss_scale": 4.0, "retries": 0, "seconds": 157.5, "max_memory_gb": 14.86, "ce": 0.07945932587608695, "kl": 0.09618055023020133, "context": 0.12754960404708982}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 0.7512223124504089, "loss_scale": 4.0, "retries": 0, "seconds": 157.9, "max_memory_gb": 14.86, "ce": 0.0998790804296732, "kl": 0.040871459437767044, "context": 0.12031481321901083}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.1252959966659546, "loss_scale": 4.0, "retries": 0, "seconds": 158.3, "max_memory_gb": 14.86, "ce": 0.10290699097095057, "kl": 0.04230215464485809, "context": 0.12405807006871328}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.6520062685012817, "loss_scale": 4.0, "retries": 0, "seconds": 158.7, "max_memory_gb": 14.86, "ce": 0.04532071715220809, "kl": 0.11061866031377576, "context": 0.10063004726544023}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.3200888931751251, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 14.86, "ce": 0.2775288596749306, "kl": 0.09614488971419632, "context": 0.3256013160571456}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 1.8154245615005493, "loss_scale": 4.0, "retries": 0, "seconds": 159.5, "max_memory_gb": 14.86, "ce": 0.306701592868194, "kl": 0.29018583383003715, "context": 0.4517945065163076}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.19216400384902954, "loss_scale": 4.0, "retries": 0, "seconds": 159.9, "max_memory_gb": 14.86, "ce": 0.08937375689856708, "kl": 0.06518009769206401, "context": 0.1219638055190444}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.24395054578781128, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 14.86, "ce": 0.016719643725082278, "kl": 0.016714098426746204, "context": 0.02507669315673411}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.12717188894748688, "loss_scale": 4.0, "retries": 0, "seconds": 160.8, "max_memory_gb": 14.86, "ce": 0.2076834114268422, "kl": 0.005898143135709688, "context": 0.21063248300924897}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.17054446041584015, "loss_scale": 4.0, "retries": 0, "seconds": 161.2, "max_memory_gb": 14.86, "ce": 0.07764029875397682, "kl": 0.02110776940753567, "context": 0.08819418353959918}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.2169015258550644, "loss_scale": 4.0, "retries": 0, "seconds": 161.6, "max_memory_gb": 14.86, "ce": 0.08816505526192486, "kl": 0.016422378190327436, "context": 0.09637624199967831}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.14991846680641174, "loss_scale": 4.0, "retries": 0, "seconds": 162.0, "max_memory_gb": 14.86, "ce": 0.050724513828754425, "kl": 0.005284772953018546, "context": 0.05336689809337258}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.37381628155708313, "loss_scale": 4.0, "retries": 0, "seconds": 162.4, "max_memory_gb": 14.86, "ce": 0.023587828734889627, "kl": 0.019665971412905492, "context": 0.033420813735574484}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 2.1833572387695312, "loss_scale": 4.0, "retries": 0, "seconds": 162.8, "max_memory_gb": 14.86, "ce": 0.1460278294980526, "kl": 0.10097683686763048, "context": 0.19651625212281942}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 0.896695613861084, "loss_scale": 4.0, "retries": 0, "seconds": 163.2, "max_memory_gb": 14.86, "ce": 0.2868974432349205, "kl": 0.15195276902522892, "context": 0.3628738336265087}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.17765149474143982, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 14.86, "ce": 0.010626683244481683, "kl": 0.006923969340277836, "context": 0.01408866822021082}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 0.7949938774108887, "loss_scale": 4.0, "retries": 0, "seconds": 164.1, "max_memory_gb": 14.86, "ce": 0.02636375732254237, "kl": 0.6950195345561951, "context": 0.3738735388033092}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 0.4360295832157135, "loss_scale": 4.0, "retries": 0, "seconds": 164.5, "max_memory_gb": 14.86, "ce": 0.11557218781672418, "kl": 0.023190666921436787, "context": 0.12716751941479743}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.18608467280864716, "loss_scale": 4.0, "retries": 0, "seconds": 164.9, "max_memory_gb": 14.86, "ce": 0.06436235667206347, "kl": 0.0170921670505777, "context": 0.07290844107046723}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 1.1321728229522705, "loss_scale": 4.0, "retries": 0, "seconds": 165.3, "max_memory_gb": 14.86, "ce": 0.012834644294343889, "kl": 0.22428503980336245, "context": 0.12497716373763978}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.2723679840564728, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 14.86, "ce": 0.15994829154806212, "kl": 0.0815588927216595, "context": 0.20072773256106302}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.5724166035652161, "loss_scale": 4.0, "retries": 0, "seconds": 166.1, "max_memory_gb": 14.86, "ce": 0.11702773906290531, "kl": 0.006762372795492411, "context": 0.12040892522782087}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.2695544958114624, "loss_scale": 4.0, "retries": 0, "seconds": 166.5, "max_memory_gb": 14.86, "ce": 0.14862703258404508, "kl": 0.0014366440882440656, "context": 0.149345354991965}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 2.559893846511841, "loss_scale": 4.0, "retries": 0, "seconds": 166.9, "max_memory_gb": 14.86, "ce": 0.14549429644830525, "kl": 0.14340385812101886, "context": 0.2171962254215032}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.663904070854187, "loss_scale": 4.0, "retries": 0, "seconds": 167.4, "max_memory_gb": 14.86, "ce": 0.08776514045894146, "kl": 0.5450931214727461, "context": 0.36031169816851616}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.2524626553058624, "loss_scale": 4.0, "retries": 0, "seconds": 167.8, "max_memory_gb": 14.86, "ce": 0.16244160197675228, "kl": 0.08323543844744563, "context": 0.2040593232959509}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.06975269317626953, "loss_scale": 4.0, "retries": 0, "seconds": 168.2, "max_memory_gb": 14.86, "ce": 0.011708382982760668, "kl": 0.011323385988362134, "context": 0.017370076151564717}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 0.23338612914085388, "loss_scale": 4.0, "retries": 0, "seconds": 168.6, "max_memory_gb": 14.86, "ce": 0.16827066976111382, "kl": 0.06142315057149972, "context": 0.19898224819917232}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.4181959927082062, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 14.86, "ce": 0.07116555399261415, "kl": 0.039814784307964146, "context": 0.09107294352725148}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.3135356605052948, "loss_scale": 4.0, "retries": 0, "seconds": 169.4, "max_memory_gb": 14.86, "ce": 0.205257156630978, "kl": 0.027713944320566952, "context": 0.2191141250077635}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 0.815422773361206, "loss_scale": 4.0, "retries": 0, "seconds": 169.8, "max_memory_gb": 14.86, "ce": 0.737885728594847, "kl": 0.336609459889587, "context": 0.9061904781265184}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 0.41710802912712097, "loss_scale": 4.0, "retries": 0, "seconds": 170.2, "max_memory_gb": 14.86, "ce": 0.25369223672896624, "kl": 0.22485495620639995, "context": 0.3661197144538164}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.7677748799324036, "loss_scale": 4.0, "retries": 0, "seconds": 170.6, "max_memory_gb": 14.86, "ce": 0.21143996191676706, "kl": 0.06212848042923724, "context": 0.24250420776661485}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.90324866771698, "loss_scale": 4.0, "retries": 0, "seconds": 171.0, "max_memory_gb": 14.86, "ce": 0.04643371974816546, "kl": 0.07809779516537674, "context": 0.08548261597752571}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 3.899306058883667, "loss_scale": 4.0, "retries": 0, "seconds": 171.5, "max_memory_gb": 14.86, "ce": 0.277242444222793, "kl": 0.21047251265554223, "context": 0.3824786855839193}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.21732491254806519, "loss_scale": 4.0, "retries": 0, "seconds": 171.9, "max_memory_gb": 14.86, "ce": 0.16343559860251844, "kl": 0.05967726855305955, "context": 0.19327422557398677}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.2828925848007202, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 14.86, "ce": 0.1062815235927701, "kl": 0.06526344089070335, "context": 0.1389132458716631}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 0.3545205891132355, "loss_scale": 4.0, "retries": 0, "seconds": 172.7, "max_memory_gb": 14.86, "ce": 0.20001213089562953, "kl": 0.007898059295257553, "context": 0.20396116084884852}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 0.6720882654190063, "loss_scale": 4.0, "retries": 0, "seconds": 173.1, "max_memory_gb": 14.86, "ce": 0.28155767370481044, "kl": 0.044676860445179045, "context": 0.30389609676785767}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.1954086869955063, "loss_scale": 4.0, "retries": 0, "seconds": 173.5, "max_memory_gb": 14.86, "ce": 0.01709678489714861, "kl": 0.0042933024269586895, "context": 0.019243436167016625}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 0.22794774174690247, "loss_scale": 4.0, "retries": 0, "seconds": 173.9, "max_memory_gb": 14.86, "ce": 0.37017335277050734, "kl": 0.06679888861253858, "context": 0.4035727884620428}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.13740617036819458, "loss_scale": 4.0, "retries": 0, "seconds": 174.3, "max_memory_gb": 14.86, "ce": 0.042107060085982084, "kl": 0.03385084887850098, "context": 0.05903248209506273}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.1682116687297821, "loss_scale": 4.0, "retries": 0, "seconds": 174.7, "max_memory_gb": 14.86, "ce": 0.08491525216959417, "kl": 0.05201072877389379, "context": 0.11092061607632786}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 1.2775629758834839, "loss_scale": 4.0, "retries": 0, "seconds": 175.1, "max_memory_gb": 14.86, "ce": 0.1397036548005417, "kl": 0.8523766280850396, "context": 0.5658919624984264}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.200135737657547, "loss_scale": 4.0, "retries": 0, "seconds": 175.5, "max_memory_gb": 14.86, "ce": 0.08140872954390943, "kl": 0.025096289711655118, "context": 0.09395687258802354}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 1.3549695014953613, "loss_scale": 4.0, "retries": 0, "seconds": 175.9, "max_memory_gb": 14.86, "ce": 0.21866221248637885, "kl": 0.015574995290080551, "context": 0.22644970210967585}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 1.423341155052185, "loss_scale": 4.0, "retries": 0, "seconds": 176.3, "max_memory_gb": 14.86, "ce": 0.3005595048889518, "kl": 0.4314032800029963, "context": 0.5162611305713654}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.2867395281791687, "loss_scale": 4.0, "retries": 0, "seconds": 176.7, "max_memory_gb": 14.86, "ce": 0.06229299516417086, "kl": 0.027112456155009568, "context": 0.07584922388195992}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.27526533603668213, "loss_scale": 4.0, "retries": 0, "seconds": 177.2, "max_memory_gb": 14.86, "ce": 0.11564278812147677, "kl": 0.11664897430091514, "context": 0.1739672808907926}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 0.7469889521598816, "loss_scale": 4.0, "retries": 0, "seconds": 177.6, "max_memory_gb": 14.86, "ce": 0.2836464215070009, "kl": 0.12164196139201522, "context": 0.3444674015045166}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.12395965307950974, "loss_scale": 4.0, "retries": 0, "seconds": 178.0, "max_memory_gb": 14.86, "ce": 0.06735379493329674, "kl": 0.013675369198608678, "context": 0.07419148145709187}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 1.598434329032898, "loss_scale": 4.0, "retries": 0, "seconds": 178.4, "max_memory_gb": 14.86, "ce": 0.011997615918517113, "kl": 0.3102851050571189, "context": 0.1671401704661548}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.9219620823860168, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 14.86, "ce": 0.021634785109199584, "kl": 0.13896069215843454, "context": 0.0911151310428977}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.2458164095878601, "loss_scale": 4.0, "retries": 0, "seconds": 179.2, "max_memory_gb": 14.86, "ce": 0.14398791291750968, "kl": 0.010109574301168323, "context": 0.14904270553961396}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 3.264986515045166, "loss_scale": 4.0, "retries": 0, "seconds": 179.6, "max_memory_gb": 14.86, "ce": 0.18981604813598096, "kl": 0.35745623124967096, "context": 0.3685441636480391}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.7671024799346924, "loss_scale": 4.0, "retries": 0, "seconds": 180.0, "max_memory_gb": 14.86, "ce": 0.09231911669485271, "kl": 0.10449244105257094, "context": 0.14456533268094063}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 1.024631381034851, "loss_scale": 4.0, "retries": 0, "seconds": 180.4, "max_memory_gb": 14.86, "ce": 0.13630788889713585, "kl": 0.10643325303681195, "context": 0.18952450982760638}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.07156272232532501, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 14.86, "ce": 0.013922140933573246, "kl": 0.00045743357986793853, "context": 0.014150857925415039}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.5676729083061218, "loss_scale": 4.0, "retries": 0, "seconds": 181.3, "max_memory_gb": 14.86, "ce": 0.11220456869341433, "kl": 0.05755524435517145, "context": 0.14098218723665923}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.39517122507095337, "loss_scale": 4.0, "retries": 0, "seconds": 181.7, "max_memory_gb": 14.86, "ce": 0.12003505136817694, "kl": 0.018614582833833992, "context": 0.12934234645217657}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 1.2607879638671875, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 14.86, "ce": 0.16524900728836656, "kl": 0.10589564993279055, "context": 0.21819683071225882}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.6948508024215698, "loss_scale": 4.0, "retries": 0, "seconds": 182.5, "max_memory_gb": 14.86, "ce": 0.10364479257259518, "kl": 0.03705302980961278, "context": 0.12217131024226546}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.5458190441131592, "loss_scale": 4.0, "retries": 0, "seconds": 182.9, "max_memory_gb": 14.86, "ce": 0.16968780383467674, "kl": 0.21084707276895642, "context": 0.27511134184896946}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.11398521065711975, "loss_scale": 4.0, "retries": 0, "seconds": 183.3, "max_memory_gb": 14.86, "ce": 0.06209392985329032, "kl": 0.059918035913142376, "context": 0.09205294703133404}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 1.3544965982437134, "loss_scale": 4.0, "retries": 0, "seconds": 183.7, "max_memory_gb": 14.86, "ce": 0.20777658955194056, "kl": 0.1866701357648708, "context": 0.3011116513516754}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.48637762665748596, "loss_scale": 4.0, "retries": 0, "seconds": 184.1, "max_memory_gb": 14.86, "ce": 0.10871813213452697, "kl": 0.040172059554606676, "context": 0.12880416144616902}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.21149756014347076, "loss_scale": 4.0, "retries": 0, "seconds": 184.5, "max_memory_gb": 14.86, "ce": 0.09395810915157199, "kl": 0.03570177341316594, "context": 0.11180899618193507}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.23674394190311432, "loss_scale": 4.0, "retries": 0, "seconds": 185.0, "max_memory_gb": 14.86, "ce": 0.1409035325050354, "kl": 0.033129609655588865, "context": 0.1574683398939669}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 0.3221457600593567, "loss_scale": 4.0, "retries": 0, "seconds": 185.4, "max_memory_gb": 14.86, "ce": 0.3540155455120839, "kl": 0.0230234825903608, "context": 0.3655273003387265}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 2.5684590339660645, "loss_scale": 4.0, "retries": 0, "seconds": 185.8, "max_memory_gb": 14.86, "ce": 0.19016135623678565, "kl": 0.07972985922242515, "context": 0.23002628982067108}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.5337099432945251, "loss_scale": 4.0, "retries": 0, "seconds": 186.2, "max_memory_gb": 14.86, "ce": 0.2653513220138848, "kl": 0.02088432861273759, "context": 0.2757934909313917}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 1.8154926300048828, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 14.86, "ce": 0.2565942220389843, "kl": 0.13019058550707996, "context": 0.32168950885534286}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.750442385673523, "loss_scale": 4.0, "retries": 0, "seconds": 187.0, "max_memory_gb": 14.86, "ce": 0.042287052725441754, "kl": 0.014866427867673337, "context": 0.0497202652040869}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 0.787094235420227, "loss_scale": 4.0, "retries": 0, "seconds": 187.4, "max_memory_gb": 14.86, "ce": 0.023354994948022068, "kl": 0.02858581543114269, "context": 0.03764790255809203}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.5562893748283386, "loss_scale": 4.0, "retries": 0, "seconds": 187.8, "max_memory_gb": 14.86, "ce": 0.025640989653766155, "kl": 0.056187967682490125, "context": 0.053734972374513745}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.746288537979126, "loss_scale": 4.0, "retries": 0, "seconds": 188.2, "max_memory_gb": 14.86, "ce": 0.1953078352380544, "kl": 0.09188994015858043, "context": 0.24125280044972897}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.4895996153354645, "loss_scale": 4.0, "retries": 0, "seconds": 188.6, "max_memory_gb": 14.86, "ce": 0.053539965534582734, "kl": 0.036862820852547884, "context": 0.0719713750295341}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 0.8084518909454346, "loss_scale": 4.0, "retries": 0, "seconds": 189.0, "max_memory_gb": 14.86, "ce": 0.1728081600740552, "kl": 0.005929145525442436, "context": 0.17577272793278098}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 2.251080274581909, "loss_scale": 4.0, "retries": 0, "seconds": 189.4, "max_memory_gb": 14.86, "ce": 0.1615276550874114, "kl": 0.1815160585101694, "context": 0.2522856881842017}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.18229906260967255, "loss_scale": 4.0, "retries": 0, "seconds": 189.8, "max_memory_gb": 14.86, "ce": 0.033673886209726334, "kl": 0.010491304681636393, "context": 0.038919538259506226}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.19823263585567474, "loss_scale": 4.0, "retries": 0, "seconds": 190.2, "max_memory_gb": 14.86, "ce": 0.06889571552164853, "kl": 0.024869861197657883, "context": 0.08133064431603998}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 1.003426432609558, "loss_scale": 4.0, "retries": 0, "seconds": 190.6, "max_memory_gb": 14.86, "ce": 0.15502539183944464, "kl": 0.14267235931765754, "context": 0.226361567620188}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 0.504291832447052, "loss_scale": 4.0, "retries": 0, "seconds": 191.1, "max_memory_gb": 14.86, "ce": 0.18442460265941918, "kl": 0.08825057701324113, "context": 0.22854989068582654}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.2427421659231186, "loss_scale": 4.0, "retries": 0, "seconds": 191.5, "max_memory_gb": 14.86, "ce": 0.03893841523677111, "kl": 0.020045185869093984, "context": 0.04896100936457515}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 0.11989126354455948, "loss_scale": 4.0, "retries": 0, "seconds": 191.9, "max_memory_gb": 14.86, "ce": 0.17971133982064202, "kl": 0.14791062310905545, "context": 0.2536666502128355}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.1349426954984665, "loss_scale": 4.0, "retries": 0, "seconds": 192.3, "max_memory_gb": 14.86, "ce": 0.04471686400938779, "kl": 0.007927194543299265, "context": 0.04868046211777255}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.14437513053417206, "loss_scale": 4.0, "retries": 0, "seconds": 192.7, "max_memory_gb": 14.86, "ce": 0.04480242484714836, "kl": 0.00470098356163362, "context": 0.0471529180649668}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.19395259022712708, "loss_scale": 4.0, "retries": 0, "seconds": 193.1, "max_memory_gb": 14.86, "ce": 0.01929733215365559, "kl": 0.0037817923293914646, "context": 0.021188228158280253}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 0.8822113871574402, "loss_scale": 4.0, "retries": 0, "seconds": 193.6, "max_memory_gb": 14.86, "ce": 0.041492449410725385, "kl": 0.6313133292514976, "context": 0.3571491066250019}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 0.8159250617027283, "loss_scale": 4.0, "retries": 0, "seconds": 194.0, "max_memory_gb": 14.86, "ce": 0.017284569272305816, "kl": 0.2244496082621481, "context": 0.1295093764201738}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.23153968155384064, "loss_scale": 4.0, "retries": 0, "seconds": 194.4, "max_memory_gb": 14.86, "ce": 0.05718494486063719, "kl": 0.0210158271365799, "context": 0.06769285839982331}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.15156501531600952, "loss_scale": 4.0, "retries": 0, "seconds": 194.8, "max_memory_gb": 14.86, "ce": 0.04050194571027532, "kl": 0.011233245575567707, "context": 0.04611856956034899}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 1.8160806894302368, "loss_scale": 4.0, "retries": 0, "seconds": 195.2, "max_memory_gb": 14.86, "ce": 0.11325545480940491, "kl": 0.03053851705044508, "context": 0.12852471345104277}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.239786982536316, "loss_scale": 4.0, "retries": 0, "seconds": 195.6, "max_memory_gb": 14.86, "ce": 0.006988148263189942, "kl": 0.4400674140197225, "context": 0.22702184924855828}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.1792786121368408, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 14.86, "ce": 0.15763664164114743, "kl": 0.1622131324984366, "context": 0.23874320974573493}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.3279078006744385, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 14.86, "ce": 0.018579164752736688, "kl": 0.009995696804253384, "context": 0.023577013052999973}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 1.2170466184616089, "loss_scale": 4.0, "retries": 0, "seconds": 197.1, "max_memory_gb": 14.86, "ce": 0.07192317640874535, "kl": 0.033044738782336935, "context": 0.08844554575625807}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.1664353311061859, "loss_scale": 4.0, "retries": 0, "seconds": 197.5, "max_memory_gb": 14.86, "ce": 0.0985702145844698, "kl": 0.015230475532007404, "context": 0.10618545184843242}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 0.2591731548309326, "loss_scale": 4.0, "retries": 0, "seconds": 198.0, "max_memory_gb": 14.86, "ce": 0.1270109375473112, "kl": 0.012819700139516499, "context": 0.13342078926507384}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.9270052313804626, "loss_scale": 4.0, "retries": 0, "seconds": 198.4, "max_memory_gb": 14.86, "ce": 0.05366336909355596, "kl": 0.008411765011260286, "context": 0.05786925239954144}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 2.6131784915924072, "loss_scale": 4.0, "retries": 0, "seconds": 198.8, "max_memory_gb": 14.86, "ce": 0.021265372983179986, "kl": 0.6264511678600684, "context": 0.3344909611623734}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.18999607861042023, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 14.86, "ce": 0.042257212626282126, "kl": 0.07627629474154674, "context": 0.08039536001160741}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 2.6807923316955566, "loss_scale": 4.0, "retries": 0, "seconds": 199.6, "max_memory_gb": 14.86, "ce": 0.41568251652643085, "kl": 0.2456777014886029, "context": 0.5385213631670922}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.31784334778785706, "loss_scale": 4.0, "retries": 0, "seconds": 200.0, "max_memory_gb": 14.86, "ce": 0.21572206309065223, "kl": 0.008845329633913934, "context": 0.2201447351835668}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.22415615618228912, "loss_scale": 4.0, "retries": 0, "seconds": 200.4, "max_memory_gb": 14.86, "ce": 0.013292365008965135, "kl": 0.015613229857990518, "context": 0.021098979865200818}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 0.6680529117584229, "loss_scale": 4.0, "retries": 0, "seconds": 200.8, "max_memory_gb": 14.86, "ce": 0.13249779818579555, "kl": 0.015386235201731324, "context": 0.14019091706722975}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 0.5969504117965698, "loss_scale": 4.0, "retries": 0, "seconds": 201.3, "max_memory_gb": 14.86, "ce": 0.16885812091641128, "kl": 0.507675152271986, "context": 0.42269570380449295}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 0.4538951814174652, "loss_scale": 4.0, "retries": 0, "seconds": 201.7, "max_memory_gb": 14.86, "ce": 0.37372985016554594, "kl": 0.10653443983756006, "context": 0.4269970660097897}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.4528258442878723, "loss_scale": 4.0, "retries": 0, "seconds": 202.1, "max_memory_gb": 14.86, "ce": 0.16481868270784616, "kl": 0.06506976915989071, "context": 0.19735356327146292}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 0.547650933265686, "loss_scale": 4.0, "retries": 0, "seconds": 202.5, "max_memory_gb": 14.86, "ce": 0.15112879418302327, "kl": 0.028874379582703114, "context": 0.16556598246097565}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 1.6946336030960083, "loss_scale": 4.0, "retries": 0, "seconds": 202.9, "max_memory_gb": 14.86, "ce": 0.08307675400283188, "kl": 0.06773342075757682, "context": 0.11694346461445093}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.33550548553466797, "loss_scale": 4.0, "retries": 0, "seconds": 203.3, "max_memory_gb": 14.86, "ce": 0.15365015529096127, "kl": 0.0206522989392397, "context": 0.1639762967824936}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 2.2901711463928223, "loss_scale": 4.0, "retries": 0, "seconds": 203.7, "max_memory_gb": 14.86, "ce": 0.24652372463606298, "kl": 0.06420729923411272, "context": 0.27862736966926605}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.20814992487430573, "loss_scale": 4.0, "retries": 0, "seconds": 204.1, "max_memory_gb": 14.86, "ce": 0.05942975822836161, "kl": 0.0250732360764232, "context": 0.07196637417655438}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.12935447692871094, "loss_scale": 4.0, "retries": 0, "seconds": 204.5, "max_memory_gb": 14.86, "ce": 0.09548761742189527, "kl": 0.09401163860457018, "context": 0.14249343832489103}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.104515939950943, "loss_scale": 4.0, "retries": 0, "seconds": 204.9, "max_memory_gb": 14.86, "ce": 0.015248992247506976, "kl": 0.0062050863925833255, "context": 0.01835153577849269}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 1.4018115997314453, "loss_scale": 4.0, "retries": 0, "seconds": 205.4, "max_memory_gb": 14.86, "ce": 0.16370017640292645, "kl": 0.03886621305719018, "context": 0.18313328363001347}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.0401121377944946, "loss_scale": 4.0, "retries": 0, "seconds": 205.8, "max_memory_gb": 14.86, "ce": 0.014447966415900737, "kl": 0.599539739843749, "context": 0.3142178275156766}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.5270738005638123, "loss_scale": 4.0, "retries": 0, "seconds": 206.2, "max_memory_gb": 14.86, "ce": 0.06958947610110044, "kl": 0.007290000532520935, "context": 0.07323447801172733}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 1.0221604108810425, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 14.86, "ce": 0.061598122818395495, "kl": 0.042511392617598176, "context": 0.08285381877794862}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 0.5631242394447327, "loss_scale": 4.0, "retries": 0, "seconds": 207.0, "max_memory_gb": 14.86, "ce": 0.1344808042049408, "kl": 0.013063708844128996, "context": 0.14101265999488533}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.32082563638687134, "loss_scale": 4.0, "retries": 0, "seconds": 207.4, "max_memory_gb": 14.86, "ce": 0.1382975386804901, "kl": 0.02711188036482781, "context": 0.1518534732167609}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 0.511121928691864, "loss_scale": 4.0, "retries": 0, "seconds": 207.8, "max_memory_gb": 14.86, "ce": 0.10395622521173209, "kl": 0.05021510331062018, "context": 0.1290637793717906}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 0.43860000371932983, "loss_scale": 4.0, "retries": 0, "seconds": 208.2, "max_memory_gb": 14.86, "ce": 0.24957700027152896, "kl": 0.013650748238433152, "context": 0.2564023705199361}
diff --git a/runs/single_llava/context/L30_s13/adapter/README.md b/runs/single_llava/context/L30_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/context/L30_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/context/L30_s13/adapter/adapter_config.json b/runs/single_llava/context/L30_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5cecc836dfc76057a0aa30ff9eac10b71db9e8e8
--- /dev/null
+++ b/runs/single_llava/context/L30_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.30.self_attn.v_proj",
+    "model.language_model.layers.30.self_attn.k_proj",
+    "model.language_model.layers.30.mlp.up_proj",
+    "model.language_model.layers.30.self_attn.o_proj",
+    "model.language_model.layers.30.self_attn.q_proj",
+    "model.language_model.layers.30.mlp.down_proj",
+    "model.language_model.layers.30.mlp.gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L30_s13/config.json b/runs/single_llava/context/L30_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c7053c10dd6c6fbf61e2fa52050d8722e94736c7
--- /dev/null
+++ b/runs/single_llava/context/L30_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "context",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/context.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 30,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/context/L30_s13/steps.jsonl b/runs/single_llava/context/L30_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..deb93cd35fffa6dbb20287c7874116dccc040f7e
--- /dev/null
+++ b/runs/single_llava/context/L30_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.2289900779724121, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 14.29, "ce": 1.6455207280814648, "kl": 0.1433797632344067, "context": 1.717210564762354}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.39134645462036133, "loss_scale": 4.0, "retries": 0, "seconds": 1.7, "max_memory_gb": 14.3, "ce": 2.6811966276727617, "kl": 0.010254374748910777, "context": 2.6863238259684294}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.9371638894081116, "loss_scale": 4.0, "retries": 0, "seconds": 2.2, "max_memory_gb": 14.3, "ce": 5.391814947128296, "kl": 0.17894858889485477, "context": 5.481289267539978}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.6485188007354736, "loss_scale": 4.0, "retries": 0, "seconds": 2.6, "max_memory_gb": 14.3, "ce": 3.8244699658825994, "kl": 0.0511308640052448, "context": 3.8500353419221938}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.03698917478322983, "loss_scale": 4.0, "retries": 0, "seconds": 3.0, "max_memory_gb": 14.3, "ce": 0.08586880937218666, "kl": 0.036669437075033784, "context": 0.10420352779328823}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.3148009479045868, "loss_scale": 4.0, "retries": 0, "seconds": 3.4, "max_memory_gb": 14.3, "ce": 1.4666164703667164, "kl": 0.18353382486384362, "context": 1.558383397758007}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.7166507840156555, "loss_scale": 4.0, "retries": 0, "seconds": 3.8, "max_memory_gb": 14.3, "ce": 3.441585225984454, "kl": 0.10535141557920724, "context": 3.4942609211429954}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.9248384833335876, "loss_scale": 4.0, "retries": 0, "seconds": 4.2, "max_memory_gb": 14.3, "ce": 3.694105416536331, "kl": 0.16107333451509476, "context": 3.7746421098709106}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.5462819933891296, "loss_scale": 4.0, "retries": 0, "seconds": 4.6, "max_memory_gb": 14.3, "ce": 2.8416371569037437, "kl": 0.16243910370394588, "context": 2.9228566735982895}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.35286757349967957, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 14.3, "ce": 1.6320101097226143, "kl": 0.060014683404006064, "context": 1.6620174646377563}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.7754027843475342, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 14.3, "ce": 2.6294446773827076, "kl": 0.009410963459231425, "context": 2.634150189347565}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.4124595522880554, "loss_scale": 4.0, "retries": 0, "seconds": 5.8, "max_memory_gb": 14.3, "ce": 1.5355944791808724, "kl": 0.01760384693625383, "context": 1.5443964062724262}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.7334040403366089, "loss_scale": 4.0, "retries": 0, "seconds": 6.2, "max_memory_gb": 14.3, "ce": 2.6722874445840716, "kl": 0.007567666310933419, "context": 2.676071368623525}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.0800443887710571, "loss_scale": 4.0, "retries": 0, "seconds": 6.6, "max_memory_gb": 14.3, "ce": 3.8914758525788784, "kl": 0.11911428952589631, "context": 3.9510329961776733}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.8287585973739624, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 14.3, "ce": 2.58516900439281, "kl": 0.26777740186662413, "context": 2.719057622423861}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.4822750389575958, "loss_scale": 4.0, "retries": 0, "seconds": 7.3, "max_memory_gb": 14.3, "ce": 1.373133915825747, "kl": 0.009503336819761898, "context": 1.3778856196440756}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.47544190287590027, "loss_scale": 4.0, "retries": 0, "seconds": 7.7, "max_memory_gb": 14.3, "ce": 1.594458781182766, "kl": 0.07859890267718583, "context": 1.6337582906708121}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.3994526863098145, "loss_scale": 4.0, "retries": 0, "seconds": 8.1, "max_memory_gb": 14.3, "ce": 3.5621126294136047, "kl": 0.09517520875670016, "context": 3.609700247645378}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 2.1589858531951904, "loss_scale": 4.0, "retries": 0, "seconds": 8.4, "max_memory_gb": 14.3, "ce": 5.315246820449829, "kl": 0.33486068196361884, "context": 5.482677221298218}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.011330093257129192, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 14.3, "ce": 0.012958253500983119, "kl": 0.0021908174312557094, "context": 0.0140536620747298}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.1383650302886963, "loss_scale": 4.0, "retries": 0, "seconds": 9.2, "max_memory_gb": 14.3, "ce": 2.7646445918362588, "kl": 0.02967978309970931, "context": 2.7794844415038824}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.1288189888000488, "loss_scale": 4.0, "retries": 0, "seconds": 9.6, "max_memory_gb": 14.3, "ce": 2.933684242889285, "kl": 0.05395905859768391, "context": 2.9606637936085463}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.3000068664550781, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 14.3, "ce": 2.632131038233638, "kl": 0.5090692287485581, "context": 2.886665691388771}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.219081163406372, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 14.3, "ce": 4.983158707618713, "kl": 0.014542733319103718, "context": 4.990429997444153}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 2.1066980361938477, "loss_scale": 4.0, "retries": 0, "seconds": 10.6, "max_memory_gb": 14.3, "ce": 4.209846138954163, "kl": 0.03636050992645323, "context": 4.228026308119297}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.461491107940674, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 14.3, "ce": 5.20552933216095, "kl": 0.06490952521562576, "context": 5.23798406124115}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.8738130331039429, "loss_scale": 4.0, "retries": 0, "seconds": 11.4, "max_memory_gb": 14.3, "ce": 4.172730825841427, "kl": 0.08385294990148395, "context": 4.214657228440046}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.3194401264190674, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 14.3, "ce": 2.7807007282972336, "kl": 0.23572147451341152, "context": 2.898561477661133}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.2884618043899536, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 14.3, "ce": 2.5880256490781903, "kl": 0.5165029412601143, "context": 2.846277087694034}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.5499991178512573, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 14.3, "ce": 2.73415772523731, "kl": 0.03763914084993303, "context": 2.752977281808853}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.634221076965332, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 14.3, "ce": 2.8223568201065063, "kl": 0.02766989590600133, "context": 2.8361917212605476}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 3.019904613494873, "loss_scale": 4.0, "retries": 0, "seconds": 13.2, "max_memory_gb": 14.3, "ce": 5.269173622131348, "kl": 0.015369085595011711, "context": 5.27685809135437}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.5734726190567017, "loss_scale": 4.0, "retries": 0, "seconds": 13.6, "max_memory_gb": 14.3, "ce": 2.443401973694563, "kl": 0.1777093536220491, "context": 2.5322566274553537}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.8256704807281494, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 14.3, "ce": 2.5872257305309176, "kl": 0.6249743616208434, "context": 2.899712972342968}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.726402997970581, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 14.3, "ce": 2.6934829428792, "kl": 0.02984513802221045, "context": 2.7084055761806667}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.6482828855514526, "loss_scale": 4.0, "retries": 0, "seconds": 14.7, "max_memory_gb": 14.3, "ce": 2.5242222025990486, "kl": 0.006833879160694778, "context": 2.527639202773571}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.8090766668319702, "loss_scale": 4.0, "retries": 0, "seconds": 15.1, "max_memory_gb": 14.3, "ce": 2.4340606704354286, "kl": 0.33768188895192, "context": 2.602901615202427}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.3192384243011475, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 14.3, "ce": 4.612398207187653, "kl": 0.48630564979976043, "context": 4.855551064014435}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.1508489847183228, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 14.3, "ce": 1.418574451468885, "kl": 0.034941352787427604, "context": 1.4360451814718544}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.0867340564727783, "loss_scale": 4.0, "retries": 0, "seconds": 16.3, "max_memory_gb": 14.3, "ce": 2.6609750390052795, "kl": 0.1199690398061648, "context": 2.720959484577179}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.3377171754837036, "loss_scale": 4.0, "retries": 0, "seconds": 16.7, "max_memory_gb": 14.3, "ce": 1.4150831960141659, "kl": 0.17347779031842947, "context": 1.5018220469355583}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.7837425470352173, "loss_scale": 4.0, "retries": 0, "seconds": 17.1, "max_memory_gb": 14.3, "ce": 2.310387712903321, "kl": 0.007091416482580826, "context": 2.3139333771541715}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.0892534255981445, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 14.3, "ce": 2.3881990262307227, "kl": 0.01173194861257798, "context": 2.3940649777650833}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.3329730033874512, "loss_scale": 4.0, "retries": 0, "seconds": 17.9, "max_memory_gb": 14.3, "ce": 1.3362381784245372, "kl": 0.2765493946499191, "context": 1.4745129202492535}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.0271822214126587, "loss_scale": 4.0, "retries": 0, "seconds": 18.3, "max_memory_gb": 14.3, "ce": 1.3808272331953049, "kl": 0.05631918297149241, "context": 1.4089868441224098}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.276090621948242, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 14.3, "ce": 2.2484071031212807, "kl": 0.07416184092289768, "context": 2.2854879796504974}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.4445221424102783, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 14.3, "ce": 1.9694055886939168, "kl": 0.03579524871020112, "context": 1.9873032558243722}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 3.6557140350341797, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 14.3, "ce": 3.7138805985450745, "kl": 0.017206098382303026, "context": 3.7224837094545364}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.5563764572143555, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 14.3, "ce": 3.4362458921968937, "kl": 0.0981087670661509, "context": 3.4853002512827516}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 3.572920322418213, "loss_scale": 4.0, "retries": 0, "seconds": 20.3, "max_memory_gb": 14.3, "ce": 3.510745106264949, "kl": 0.05205062287859619, "context": 3.5367704136297107}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.5295408964157104, "loss_scale": 4.0, "retries": 0, "seconds": 20.6, "max_memory_gb": 14.3, "ce": 1.3300953842699528, "kl": 0.131685710279271, "context": 1.3959382665343583}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 3.193310022354126, "loss_scale": 4.0, "retries": 0, "seconds": 21.0, "max_memory_gb": 14.3, "ce": 2.2112158685922623, "kl": 0.1184011481818743, "context": 2.270416483283043}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 3.5430188179016113, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 14.3, "ce": 3.5998626723885536, "kl": 0.18962881760671735, "context": 3.6946771666407585}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 3.876922369003296, "loss_scale": 4.0, "retries": 0, "seconds": 21.7, "max_memory_gb": 14.3, "ce": 3.714969426393509, "kl": 0.04223079408984631, "context": 3.7360848039388657}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 2.4350664615631104, "loss_scale": 4.0, "retries": 0, "seconds": 22.0, "max_memory_gb": 14.3, "ce": 2.395491946488619, "kl": 0.018256531737279147, "context": 2.404620212968439}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 5.106991291046143, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 14.3, "ce": 4.6121591329574585, "kl": 0.305683747632429, "context": 4.765000939369202}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.242140769958496, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 14.3, "ce": 1.1814895160496235, "kl": 0.10222585231531411, "context": 1.2326024314388633}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 2.9543044567108154, "loss_scale": 4.0, "retries": 0, "seconds": 23.1, "max_memory_gb": 14.3, "ce": 2.2874992839060724, "kl": 0.004318988207160146, "context": 2.289658677764237}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.4968922138214111, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 14.3, "ce": 1.364886486902833, "kl": 0.06713155168108642, "context": 1.3984523173421621}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 2.668856620788574, "loss_scale": 4.0, "retries": 0, "seconds": 23.8, "max_memory_gb": 14.3, "ce": 2.305298467632383, "kl": 0.01280157471774146, "context": 2.311699249781668}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 4.377236366271973, "loss_scale": 4.0, "retries": 0, "seconds": 24.2, "max_memory_gb": 14.3, "ce": 3.2732213363051414, "kl": 0.06318126834230497, "context": 3.30481193959713}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.6099975109100342, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 14.3, "ce": 1.2648401334881783, "kl": 0.03991486434824765, "context": 1.284797502681613}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 4.289663314819336, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 14.3, "ce": 3.4116190606728196, "kl": 0.0906927758915117, "context": 3.456965400837362}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 4.5005292892456055, "loss_scale": 4.0, "retries": 0, "seconds": 25.3, "max_memory_gb": 14.31, "ce": 3.418600235134363, "kl": 0.036400106153450906, "context": 3.436800327152014}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 4.210296630859375, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 14.31, "ce": 3.5216535180807114, "kl": 0.13809503288939595, "context": 3.5907010436058044}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.5033122301101685, "loss_scale": 4.0, "retries": 0, "seconds": 26.0, "max_memory_gb": 14.31, "ce": 0.8810305246151984, "kl": 0.06150615104706958, "context": 0.9117836280493066}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 4.897002696990967, "loss_scale": 4.0, "retries": 0, "seconds": 26.4, "max_memory_gb": 14.31, "ce": 3.213158890604973, "kl": 0.01888186900760047, "context": 3.2225998789072037}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.6568934917449951, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 14.31, "ce": 1.1145925652235746, "kl": 0.04515340831130743, "context": 1.1371693196706474}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 2.8655033111572266, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 14.31, "ce": 2.103509480599314, "kl": 0.01354763816925697, "context": 2.1102832704782486}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 3.4425570964813232, "loss_scale": 4.0, "retries": 0, "seconds": 27.5, "max_memory_gb": 14.31, "ce": 2.067600494250655, "kl": 0.00403144933807198, "context": 2.069616274908185}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 5.127130031585693, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 14.31, "ce": 3.05930151976645, "kl": 0.010464784340001643, "context": 3.064533971250057}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 5.792276859283447, "loss_scale": 4.0, "retries": 0, "seconds": 28.2, "max_memory_gb": 14.31, "ce": 3.630815327167511, "kl": 0.14584170281887054, "context": 3.703736186027527}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.10941027849912643, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 14.31, "ce": 0.22978946939110756, "kl": 0.0362423793412745, "context": 0.24791066069155931}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 3.405322790145874, "loss_scale": 4.0, "retries": 0, "seconds": 28.9, "max_memory_gb": 14.31, "ce": 2.0642635258845985, "kl": 0.026631265704054385, "context": 2.0775791443884373}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 2.214541435241699, "loss_scale": 4.0, "retries": 0, "seconds": 29.3, "max_memory_gb": 14.31, "ce": 1.5338476598262787, "kl": 0.19831510446965694, "context": 1.63300521671772}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 3.3239693641662598, "loss_scale": 4.0, "retries": 0, "seconds": 29.7, "max_memory_gb": 14.31, "ce": 1.9072661586105824, "kl": 0.447786970064044, "context": 2.1311596669256687}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 5.226433753967285, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 14.31, "ce": 3.1973337531089783, "kl": 0.041349735110998154, "context": 3.218008652329445}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 3.6630303859710693, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 14.31, "ce": 1.6383516592904925, "kl": 0.06655374402180314, "context": 1.6716284975409508}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 2.151071786880493, "loss_scale": 4.0, "retries": 0, "seconds": 30.7, "max_memory_gb": 14.31, "ce": 0.828016591258347, "kl": 0.07856609765076428, "context": 0.867299631703645}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 3.7965805530548096, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 14.31, "ce": 1.8385652215220034, "kl": 0.00905035895993933, "context": 1.8430903868284076}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 5.557736873626709, "loss_scale": 4.0, "retries": 0, "seconds": 31.4, "max_memory_gb": 14.31, "ce": 2.7140288800001144, "kl": 0.04922359739430249, "context": 2.738640695810318}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 5.660688877105713, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 14.31, "ce": 2.6933023929595947, "kl": 0.06194018479436636, "context": 2.7242724746465683}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 5.335175037384033, "loss_scale": 4.0, "retries": 0, "seconds": 32.2, "max_memory_gb": 14.31, "ce": 2.2864248789846897, "kl": 0.1913210917264223, "context": 2.3820854453369975}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.8688679933547974, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 14.31, "ce": 1.0144327821908519, "kl": 0.07689576933626086, "context": 1.0528806445654482}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.0699431300163269, "loss_scale": 4.0, "retries": 0, "seconds": 32.9, "max_memory_gb": 14.31, "ce": 0.07328790286555886, "kl": 0.035700373893632786, "context": 0.09113808954134583}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 5.429012298583984, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 14.31, "ce": 2.4829642847180367, "kl": 0.1777131879935041, "context": 2.5718208737671375}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 5.602683067321777, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 14.31, "ce": 2.485250301659107, "kl": 0.061946295434609056, "context": 2.5162233971059322}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 3.955671787261963, "loss_scale": 4.0, "retries": 0, "seconds": 34.0, "max_memory_gb": 14.31, "ce": 1.6736042422708124, "kl": 0.02487055769597646, "context": 1.6860395830590278}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.05085847154259682, "loss_scale": 4.0, "retries": 0, "seconds": 34.4, "max_memory_gb": 14.31, "ce": 0.11357916728593409, "kl": 0.00594682790688239, "context": 0.11655258387327194}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 3.3148179054260254, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 14.31, "ce": 1.456522284541279, "kl": 0.020341675728559494, "context": 1.4666931321844459}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.15702250599861145, "loss_scale": 4.0, "retries": 0, "seconds": 35.1, "max_memory_gb": 14.31, "ce": 0.33606983721256256, "kl": 0.09240465378388762, "context": 0.38227216713130474}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.2381515502929688, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 14.31, "ce": 0.8590459765400738, "kl": 0.07955659832805395, "context": 0.8988242575433105}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 5.3554911613464355, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 14.31, "ce": 1.3175860855262727, "kl": 0.08107168578135315, "context": 1.3581219562329352}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 6.310910701751709, "loss_scale": 4.0, "retries": 0, "seconds": 36.2, "max_memory_gb": 14.31, "ce": 2.363576352596283, "kl": 0.10659679770469666, "context": 2.416874758899212}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 4.0754170417785645, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 14.31, "ce": 1.6072327215224504, "kl": 0.07370667555369437, "context": 1.644086066633463}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 6.412315368652344, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 14.31, "ce": 2.3315525203943253, "kl": 0.038475254812510684, "context": 2.3507901281118393}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.21236252784729, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 14.31, "ce": 1.0427400693297386, "kl": 0.023143993224948645, "context": 1.0543120838701725}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 4.331377983093262, "loss_scale": 4.0, "retries": 0, "seconds": 37.6, "max_memory_gb": 14.31, "ce": 1.4011375294066966, "kl": 0.048743168053988484, "context": 1.4255090937949717}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.1555136889219284, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 14.31, "ce": 0.23339410708285868, "kl": 0.0667599473381415, "context": 0.26677407743409276}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 5.8256659507751465, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 14.31, "ce": 2.1363233029842377, "kl": 0.09075814820243977, "context": 2.181702349334955}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 6.269915580749512, "loss_scale": 4.0, "retries": 0, "seconds": 38.7, "max_memory_gb": 14.31, "ce": 1.9457865878939629, "kl": 0.3814681868534535, "context": 2.1365206837654114}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 6.106744766235352, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 14.31, "ce": 2.2966260761022568, "kl": 0.146362517029047, "context": 2.369807332754135}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 4.255862712860107, "loss_scale": 4.0, "retries": 0, "seconds": 39.5, "max_memory_gb": 14.31, "ce": 1.2849089584778994, "kl": 0.08543470571748912, "context": 1.3276262991130352}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.428131580352783, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 14.31, "ce": 0.6933591613778844, "kl": 0.04215444851433858, "context": 0.7144363721599802}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 6.589532852172852, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 14.31, "ce": 1.8690517842769623, "kl": 0.19374882979900576, "context": 1.9659262150526047}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 6.697399616241455, "loss_scale": 4.0, "retries": 0, "seconds": 40.5, "max_memory_gb": 14.31, "ce": 1.6455232203006744, "kl": 0.07831126637756824, "context": 1.6846788376569748}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.216848611831665, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 14.31, "ce": 0.7227002484723926, "kl": 0.021933741751126945, "context": 0.733667110092938}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.3034894466400146, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 14.31, "ce": 0.7474905047565699, "kl": 0.026319821772631258, "context": 0.7606504042632878}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 6.808823585510254, "loss_scale": 4.0, "retries": 0, "seconds": 41.6, "max_memory_gb": 14.31, "ce": 1.819172516465187, "kl": 0.3361210352159105, "context": 1.9872329831123352}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 4.414247989654541, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 14.31, "ce": 1.209002673625946, "kl": 0.042033424018882215, "context": 1.2300193905830383}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 4.217545509338379, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 14.31, "ce": 1.0341072180308402, "kl": 0.07537607869016938, "context": 1.0717952344566584}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 4.615652084350586, "loss_scale": 4.0, "retries": 0, "seconds": 42.7, "max_memory_gb": 14.31, "ce": 1.049206992611289, "kl": 0.016759908088715747, "context": 1.0575869530439377}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 4.940865993499756, "loss_scale": 4.0, "retries": 0, "seconds": 43.1, "max_memory_gb": 14.31, "ce": 1.1924769654870033, "kl": 0.045281559228897095, "context": 1.2151177376508713}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 7.058496475219727, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 14.31, "ce": 1.543251872062683, "kl": 0.6292313473531976, "context": 1.8578675389289856}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 4.587731838226318, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 14.31, "ce": 1.0734931230545044, "kl": 0.23366502579301596, "context": 1.1903256326913834}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.11267320811748505, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 14.31, "ce": 0.12034041294828057, "kl": 0.005946456738456618, "context": 0.12331363977864385}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 4.524134635925293, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 14.31, "ce": 1.104179485118948, "kl": 0.011221159365959466, "context": 1.1097900471650064}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 2.5139167308807373, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 14.31, "ce": 0.559375838143751, "kl": 0.0479953082249267, "context": 0.5833734860643744}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.7068424224853516, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 14.31, "ce": 0.3959027798846364, "kl": 0.07008269126527011, "context": 0.4309441316872835}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 4.318217754364014, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 14.31, "ce": 0.8981509692966938, "kl": 0.06763369659893215, "context": 0.9319678102619946}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.7388596534729004, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 14.31, "ce": 0.6691756434738636, "kl": 0.08614557122200495, "context": 0.7122484208084643}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 3.7665698528289795, "loss_scale": 4.0, "retries": 0, "seconds": 46.4, "max_memory_gb": 14.31, "ce": 0.730181201826781, "kl": 0.022232817485928535, "context": 0.7412976094055921}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.4854915142059326, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 14.31, "ce": 0.830801147967577, "kl": 0.04074912593932822, "context": 0.8511757235974073}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 2.1351826190948486, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 14.31, "ce": 0.37364812660962343, "kl": 0.030426498327869922, "context": 0.38886138051748276}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.182034969329834, "loss_scale": 4.0, "retries": 0, "seconds": 47.5, "max_memory_gb": 14.31, "ce": 0.4139210907742381, "kl": 0.019853711244650185, "context": 0.4238479360938072}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 4.391542434692383, "loss_scale": 4.0, "retries": 0, "seconds": 47.8, "max_memory_gb": 14.31, "ce": 0.9517335370182991, "kl": 0.09067459288053215, "context": 0.9970708265900612}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 2.0426158905029297, "loss_scale": 4.0, "retries": 0, "seconds": 48.2, "max_memory_gb": 14.31, "ce": 0.22761356621049345, "kl": 0.004721679029898951, "context": 0.22997440793551505}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 4.1660847663879395, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 14.31, "ce": 0.4909844798967242, "kl": 0.30311757268646033, "context": 0.6425432551186532}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 7.442375183105469, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 14.31, "ce": 0.9244335442781448, "kl": 0.07916580862365663, "context": 0.9640164524316788}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 4.1100382804870605, "loss_scale": 4.0, "retries": 0, "seconds": 49.2, "max_memory_gb": 14.31, "ce": 0.7160664796829224, "kl": 0.3647457400802523, "context": 0.8984393756836653}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 2.253911256790161, "loss_scale": 4.0, "retries": 0, "seconds": 49.6, "max_memory_gb": 14.31, "ce": 0.41368288919329643, "kl": 0.04777782829478383, "context": 0.43757180869579315}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 5.249325752258301, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 14.31, "ce": 0.8640775866806507, "kl": 0.18188438203651458, "context": 0.9550197725184262}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 5.288639545440674, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 14.31, "ce": 0.7467288672924042, "kl": 0.1350650079548359, "context": 0.8142613917589188}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.18183238804340363, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 14.31, "ce": 0.21214433386921883, "kl": 0.014782615879084915, "context": 0.21953563764691353}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 4.640082836151123, "loss_scale": 4.0, "retries": 0, "seconds": 51.1, "max_memory_gb": 14.31, "ce": 0.5901923552155495, "kl": 0.027015888364985585, "context": 0.6037002950906754}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 4.149867057800293, "loss_scale": 4.0, "retries": 0, "seconds": 51.4, "max_memory_gb": 14.31, "ce": 0.42178842052817345, "kl": 0.04444181779399514, "context": 0.44400932639837265}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 2.3929896354675293, "loss_scale": 4.0, "retries": 0, "seconds": 51.8, "max_memory_gb": 14.31, "ce": 0.2685955837368965, "kl": 0.024291136767715216, "context": 0.28074114955961704}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 2.6759746074676514, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 14.31, "ce": 0.3691636063158512, "kl": 0.059013326186686754, "context": 0.3986702747642994}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 3.648707866668701, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 14.31, "ce": 0.5565346665680408, "kl": 0.0838800065102987, "context": 0.5984746739268303}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 1.5581660270690918, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 14.31, "ce": 0.1782350391149521, "kl": 0.02462693676352501, "context": 0.19054850935935974}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 1.4191551208496094, "loss_scale": 4.0, "retries": 0, "seconds": 53.3, "max_memory_gb": 14.31, "ce": 0.23217611154541373, "kl": 0.020291403867304325, "context": 0.24232181534171104}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.009685754776001, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 14.31, "ce": 0.27327278163284063, "kl": 0.012309520738199353, "context": 0.27942753676325083}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.18131378293037415, "loss_scale": 4.0, "retries": 0, "seconds": 54.0, "max_memory_gb": 14.31, "ce": 0.14106146921403706, "kl": 0.058558131830068305, "context": 0.17034053592942655}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 3.239063262939453, "loss_scale": 4.0, "retries": 0, "seconds": 54.4, "max_memory_gb": 14.31, "ce": 0.585295632481575, "kl": 0.11985167488455772, "context": 0.6452214736491442}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 1.8959896564483643, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 14.31, "ce": 0.1781841591000557, "kl": 0.005308761894411873, "context": 0.18083854345604777}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 1.2107703685760498, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 14.31, "ce": 0.2161313397809863, "kl": 0.015242318389937282, "context": 0.22375249862670898}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 1.3702558279037476, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 14.31, "ce": 0.29492544452659786, "kl": 0.22463239956414327, "context": 0.4072416462004185}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.676734447479248, "loss_scale": 4.0, "retries": 0, "seconds": 55.8, "max_memory_gb": 14.31, "ce": 0.0707385039422661, "kl": 0.1001721149077639, "context": 0.12082456168718636}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 1.2622441053390503, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 14.31, "ce": 0.17450885940343142, "kl": 0.010192023270064965, "context": 0.17960487492382526}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 1.4781657457351685, "loss_scale": 4.0, "retries": 0, "seconds": 56.6, "max_memory_gb": 14.31, "ce": 0.3133686017245054, "kl": 0.024537514778785408, "context": 0.3256373591721058}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.9187262654304504, "loss_scale": 4.0, "retries": 0, "seconds": 56.9, "max_memory_gb": 14.31, "ce": 0.1875210041180253, "kl": 0.014445892971707508, "context": 0.19474395085126162}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 0.8840740323066711, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 14.31, "ce": 0.1420611683279276, "kl": 0.3349678700324148, "context": 0.30954510532319546}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 1.3347097635269165, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 14.31, "ce": 0.17771779280155897, "kl": 0.16599574068095535, "context": 0.2607156643643975}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.0306742191314697, "loss_scale": 4.0, "retries": 0, "seconds": 58.0, "max_memory_gb": 14.31, "ce": 0.5858702789992094, "kl": 0.25927392579615116, "context": 0.7155072186142206}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.976001024246216, "loss_scale": 4.0, "retries": 0, "seconds": 58.4, "max_memory_gb": 14.31, "ce": 0.5949969822540879, "kl": 0.1250334620126523, "context": 0.6575137050822377}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.5591610670089722, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 14.31, "ce": 0.1603121762163937, "kl": 0.10622193230665289, "context": 0.2134231321979314}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.8328376412391663, "loss_scale": 4.0, "retries": 0, "seconds": 59.1, "max_memory_gb": 14.31, "ce": 0.21810240764170885, "kl": 0.04780138484784402, "context": 0.24200309440493584}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.5850553512573242, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 14.31, "ce": 0.3053565453737974, "kl": 0.16972913080826402, "context": 0.3902211207896471}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.6074352264404297, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 14.31, "ce": 0.08954492397606373, "kl": 0.0840493236319162, "context": 0.131569585762918}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.7439916133880615, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 14.31, "ce": 0.15350461192429066, "kl": 0.058356255292892456, "context": 0.18268273770809174}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.886082649230957, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 14.31, "ce": 0.25169289391487837, "kl": 0.047315733434516005, "context": 0.2753507560119033}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.31339794397354126, "loss_scale": 4.0, "retries": 0, "seconds": 60.9, "max_memory_gb": 14.31, "ce": 0.12197498790919781, "kl": 0.026490576448850334, "context": 0.13522027805447578}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 1.31551992893219, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 14.31, "ce": 0.14096925221383572, "kl": 0.07625917843688512, "context": 0.17909884173423052}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 1.631754755973816, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 14.31, "ce": 0.34304163604974747, "kl": 0.02056676836218685, "context": 0.3533250167965889}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.41932225227355957, "loss_scale": 4.0, "retries": 0, "seconds": 62.0, "max_memory_gb": 14.31, "ce": 0.19329665414988995, "kl": 0.05248445817960601, "context": 0.21953888423740864}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 1.0277032852172852, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 14.31, "ce": 0.20748246274888515, "kl": 0.05498601507861167, "context": 0.23497547395527363}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.7139797806739807, "loss_scale": 4.0, "retries": 0, "seconds": 62.8, "max_memory_gb": 14.31, "ce": 0.2507446492090821, "kl": 0.012884051713626832, "context": 0.25718667451292276}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.5597507357597351, "loss_scale": 4.0, "retries": 0, "seconds": 63.1, "max_memory_gb": 14.31, "ce": 0.09609345067292452, "kl": 0.05390328180510551, "context": 0.12304508779197931}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.3138673305511475, "loss_scale": 4.0, "retries": 0, "seconds": 63.5, "max_memory_gb": 14.31, "ce": 0.2510137055069208, "kl": 0.7090984696988016, "context": 0.6055629551410675}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.31536543369293213, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 14.31, "ce": 0.03968963632360101, "kl": 0.006251447077374905, "context": 0.042815360240638256}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 1.200451135635376, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 14.31, "ce": 0.07901640306226909, "kl": 0.1600575891789049, "context": 0.15904520032927394}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 1.3307174444198608, "loss_scale": 4.0, "retries": 0, "seconds": 64.6, "max_memory_gb": 14.31, "ce": 0.15677057206630707, "kl": 0.011225644615478814, "context": 0.1623833943158388}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 1.4647512435913086, "loss_scale": 4.0, "retries": 0, "seconds": 65.0, "max_memory_gb": 14.31, "ce": 0.16973073408007622, "kl": 0.06931785691995174, "context": 0.2043896596878767}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.9467251896858215, "loss_scale": 4.0, "retries": 0, "seconds": 65.3, "max_memory_gb": 14.31, "ce": 0.10976653499528766, "kl": 0.05310881696641445, "context": 0.13632094534114003}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.7840304374694824, "loss_scale": 4.0, "retries": 0, "seconds": 65.7, "max_memory_gb": 14.31, "ce": 0.22706101555377245, "kl": 0.16217045718804002, "context": 0.3081462513655424}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.6100072860717773, "loss_scale": 4.0, "retries": 0, "seconds": 66.1, "max_memory_gb": 14.31, "ce": 0.06743651069700718, "kl": 0.012882889132015407, "context": 0.07387795625254512}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 1.610650658607483, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 14.31, "ce": 0.24172873748466372, "kl": 0.07750171929365024, "context": 0.2804795945994556}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.324404239654541, "loss_scale": 4.0, "retries": 0, "seconds": 66.8, "max_memory_gb": 14.31, "ce": 0.20774022489786148, "kl": 0.0724047155235894, "context": 0.2439425871707499}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.3621128797531128, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 14.31, "ce": 0.3418285599909723, "kl": 0.005360172799555585, "context": 0.3445086325518787}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 1.6500643491744995, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 14.31, "ce": 0.28325581550598145, "kl": 0.04035891543026082, "context": 0.30343526508659124}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 1.1072075366973877, "loss_scale": 4.0, "retries": 0, "seconds": 67.9, "max_memory_gb": 14.31, "ce": 0.10681662987917662, "kl": 0.03873229608871043, "context": 0.12618277966976166}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.0902366638183594, "loss_scale": 4.0, "retries": 0, "seconds": 68.2, "max_memory_gb": 14.31, "ce": 0.18235276360064745, "kl": 0.005591990789980628, "context": 0.1851487592794001}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.21957728266716003, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 14.31, "ce": 0.02458601852413267, "kl": 0.007418217846861808, "context": 0.02829512779135257}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 1.4814341068267822, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 14.31, "ce": 0.09384978888556361, "kl": 0.10249267259496264, "context": 0.1450961281079799}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.49958765506744385, "loss_scale": 4.0, "retries": 0, "seconds": 69.3, "max_memory_gb": 14.31, "ce": 0.04726639622822404, "kl": 0.018221008183900267, "context": 0.05637689959257841}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.7031246423721313, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 14.31, "ce": 0.07765007996931672, "kl": 0.0061367739690467715, "context": 0.08071846701204777}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 1.6011841297149658, "loss_scale": 4.0, "retries": 0, "seconds": 70.1, "max_memory_gb": 14.31, "ce": 0.11229895194992423, "kl": 0.23912725871196017, "context": 0.2318625864572823}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.19880148768424988, "loss_scale": 4.0, "retries": 0, "seconds": 70.4, "max_memory_gb": 14.31, "ce": 0.14973110985010862, "kl": 0.044252055708057014, "context": 0.1718571368837729}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 0.48189595341682434, "loss_scale": 4.0, "retries": 0, "seconds": 70.8, "max_memory_gb": 14.31, "ce": 0.0525537533685565, "kl": 0.006952076539164409, "context": 0.05602979240939021}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.5812016725540161, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 14.31, "ce": 0.047903696075081825, "kl": 0.2929761721752584, "context": 0.1943917814642191}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 0.2598552107810974, "loss_scale": 4.0, "retries": 0, "seconds": 71.5, "max_memory_gb": 14.31, "ce": 0.2552829319611192, "kl": 0.049900167621672153, "context": 0.28023301903158426}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.3023613393306732, "loss_scale": 4.0, "retries": 0, "seconds": 71.9, "max_memory_gb": 14.31, "ce": 0.19012046046555042, "kl": 0.005244608400971629, "context": 0.19274275936186314}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.738612174987793, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 14.31, "ce": 0.06712989090010524, "kl": 0.02992989882477559, "context": 0.0820948388427496}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 0.7337422370910645, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 14.31, "ce": 0.28922446351498365, "kl": 0.01918446971103549, "context": 0.29881669487804174}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.6178027391433716, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 14.31, "ce": 0.11569479154422879, "kl": 0.020261618541553617, "context": 0.12582560162991285}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.7295201420783997, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 14.31, "ce": 0.05077624274417758, "kl": 0.027088077840744518, "context": 0.06432028161361814}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.5181013345718384, "loss_scale": 4.0, "retries": 0, "seconds": 73.7, "max_memory_gb": 14.31, "ce": 0.15504268649965525, "kl": 0.02730124199297279, "context": 0.16869330825284123}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.8039306402206421, "loss_scale": 4.0, "retries": 0, "seconds": 74.0, "max_memory_gb": 14.31, "ce": 0.21618076227605343, "kl": 0.050749495305353776, "context": 0.24155551008880138}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 0.9429273009300232, "loss_scale": 4.0, "retries": 0, "seconds": 74.4, "max_memory_gb": 14.31, "ce": 0.0876607233658433, "kl": 0.154656677346793, "context": 0.1649890560656786}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.34840771555900574, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 14.31, "ce": 0.25599752087146044, "kl": 0.010111331823281944, "context": 0.2610531775280833}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.2863204777240753, "loss_scale": 4.0, "retries": 0, "seconds": 75.2, "max_memory_gb": 14.31, "ce": 0.12979473359882832, "kl": 0.01148532540537417, "context": 0.13553739804774523}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.22024479508399963, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 14.31, "ce": 0.015382420155219734, "kl": 0.008040044878725894, "context": 0.01940244249999523}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 0.7877633571624756, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 14.31, "ce": 0.19189965119585395, "kl": 0.09243882447481155, "context": 0.23811906203627586}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 1.5544191598892212, "loss_scale": 4.0, "retries": 0, "seconds": 76.2, "max_memory_gb": 14.31, "ce": 0.13936496200039983, "kl": 0.02059978956822306, "context": 0.1496648583561182}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 1.6247100830078125, "loss_scale": 4.0, "retries": 0, "seconds": 76.6, "max_memory_gb": 14.31, "ce": 0.1317943292669952, "kl": 0.2663839172164444, "context": 0.2649862876860425}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 1.927116870880127, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 14.31, "ce": 0.11897995695471764, "kl": 0.0491083606320899, "context": 0.14353413926437497}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 0.7208269238471985, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 14.31, "ce": 0.5857829535380006, "kl": 0.41867409873520955, "context": 0.7951199766248465}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.6346136331558228, "loss_scale": 4.0, "retries": 0, "seconds": 77.7, "max_memory_gb": 14.31, "ce": 0.0395804385188967, "kl": 0.009841217019129544, "context": 0.0445010457187891}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.41418972611427307, "loss_scale": 4.0, "retries": 0, "seconds": 78.1, "max_memory_gb": 14.31, "ce": 0.21889414777979255, "kl": 0.014227627078071237, "context": 0.2260079625993967}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 0.6884024143218994, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 14.31, "ce": 0.21298280300106853, "kl": 0.05570861865999177, "context": 0.240837111370638}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.9980534315109253, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 14.31, "ce": 0.20409731846302748, "kl": 0.09521937719546258, "context": 0.2517070029862225}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 0.9447720646858215, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 14.31, "ce": 0.19648216664791107, "kl": 0.541251156013459, "context": 0.4671077523380518}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.4630303978919983, "loss_scale": 4.0, "retries": 0, "seconds": 79.6, "max_memory_gb": 14.31, "ce": 0.2833150364458561, "kl": 0.048560190363787115, "context": 0.307595144957304}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 1.9002724885940552, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 14.31, "ce": 0.2725129062309861, "kl": 0.2724657170474529, "context": 0.40874575916677713}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 0.31250280141830444, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 14.31, "ce": 0.35544331674464047, "kl": 0.0048705728113418445, "context": 0.3578786051366478}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.1646144390106201, "loss_scale": 4.0, "retries": 0, "seconds": 80.7, "max_memory_gb": 14.31, "ce": 0.3107203128747642, "kl": 0.11295230890391394, "context": 0.36719647189602256}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.5488419532775879, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 14.31, "ce": 0.08732822956517339, "kl": 0.04689668113132939, "context": 0.11077656876295805}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 0.7121362090110779, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 14.31, "ce": 0.18301584268920124, "kl": 0.3234568436164409, "context": 0.344744271133095}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.7935445308685303, "loss_scale": 4.0, "retries": 0, "seconds": 81.8, "max_memory_gb": 14.31, "ce": 0.12372324010357261, "kl": 0.029267872858326882, "context": 0.13835717551410198}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 1.2085371017456055, "loss_scale": 4.0, "retries": 0, "seconds": 82.1, "max_memory_gb": 14.31, "ce": 0.09578972263261676, "kl": 0.02251544618047774, "context": 0.1070474456064403}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 1.4073293209075928, "loss_scale": 4.0, "retries": 0, "seconds": 82.5, "max_memory_gb": 14.31, "ce": 0.12124963896349072, "kl": 0.06151316536124796, "context": 0.15200622100383043}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 4.314302444458008, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 14.31, "ce": 0.3188896430656314, "kl": 0.3067894992273068, "context": 0.47228437941521406}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.27324873208999634, "loss_scale": 4.0, "retries": 0, "seconds": 83.2, "max_memory_gb": 14.31, "ce": 0.23506876546889544, "kl": 0.0383466777857393, "context": 0.254242108669132}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.8777986764907837, "loss_scale": 4.0, "retries": 0, "seconds": 83.6, "max_memory_gb": 14.31, "ce": 0.11044262396171689, "kl": 0.07890778983710334, "context": 0.14989652065560222}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 0.4061729907989502, "loss_scale": 4.0, "retries": 0, "seconds": 84.0, "max_memory_gb": 14.31, "ce": 0.10090667894110084, "kl": 0.03777188784442842, "context": 0.11979262623935938}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.2900192439556122, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 14.31, "ce": 0.14210922829806805, "kl": 0.023999395198188722, "context": 0.15410892851650715}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 0.5092571973800659, "loss_scale": 4.0, "retries": 0, "seconds": 84.7, "max_memory_gb": 14.31, "ce": 0.024116550805047154, "kl": 0.009003284736536443, "context": 0.028618192533031106}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.5645196437835693, "loss_scale": 4.0, "retries": 0, "seconds": 85.1, "max_memory_gb": 14.31, "ce": 0.1299323122948408, "kl": 0.0853915112093091, "context": 0.17262806743383408}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.4124627709388733, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 14.31, "ce": 0.26155986124649644, "kl": 0.13564784824848175, "context": 0.3293837886303663}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.4985226094722748, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 14.31, "ce": 0.3054252912988886, "kl": 0.08818048238754272, "context": 0.34951552748680115}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.4120984971523285, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 14.31, "ce": 0.20699930214323103, "kl": 0.022054164961446077, "context": 0.21802638471126556}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 0.6795827150344849, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 14.31, "ce": 0.04958684789016843, "kl": 0.008282918919576332, "context": 0.05372830806300044}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 3.5416276454925537, "loss_scale": 4.0, "retries": 0, "seconds": 86.9, "max_memory_gb": 14.31, "ce": 0.2523865457624197, "kl": 0.12652865404743352, "context": 0.31565086357295513}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 0.34011149406433105, "loss_scale": 4.0, "retries": 0, "seconds": 87.2, "max_memory_gb": 14.31, "ce": 0.21156207518652081, "kl": 0.020067391742486507, "context": 0.22159577254205942}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 0.3266112208366394, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 14.31, "ce": 0.2730535580776632, "kl": 0.035672329540830106, "context": 0.29088972555473447}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.4139215648174286, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 14.31, "ce": 0.0845337687060237, "kl": 0.01991665698005818, "context": 0.09449209528975189}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.6775527000427246, "loss_scale": 4.0, "retries": 0, "seconds": 88.3, "max_memory_gb": 14.31, "ce": 0.2119411574676633, "kl": 0.09463575151130499, "context": 0.2592590262647718}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.6920862197875977, "loss_scale": 4.0, "retries": 0, "seconds": 88.7, "max_memory_gb": 14.31, "ce": 0.0512789161875844, "kl": 0.09909419673931552, "context": 0.10082601441536099}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.8974929451942444, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 14.31, "ce": 0.25772577431052923, "kl": 0.03400473250076175, "context": 0.2747281463816762}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 0.8802974820137024, "loss_scale": 4.0, "retries": 0, "seconds": 89.4, "max_memory_gb": 14.31, "ce": 0.14927278133109212, "kl": 0.039732874953188, "context": 0.16913921665400267}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.23846302926540375, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 14.31, "ce": 0.2259913608431816, "kl": 0.01927762507693842, "context": 0.23563017323613167}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.7612995505332947, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 14.31, "ce": 0.16336804442107677, "kl": 0.028769874945282936, "context": 0.17775297560729086}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 0.6297929883003235, "loss_scale": 4.0, "retries": 0, "seconds": 90.5, "max_memory_gb": 14.31, "ce": 0.03127297479659319, "kl": 0.06454836821649224, "context": 0.06354716094210744}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 0.4847443103790283, "loss_scale": 4.0, "retries": 0, "seconds": 90.9, "max_memory_gb": 14.31, "ce": 0.25422486884053797, "kl": 0.06328434275928885, "context": 0.28586703236214817}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 0.37811994552612305, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 14.31, "ce": 0.3820593885611743, "kl": 0.010927230934612453, "context": 0.38752299500629306}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.4111311435699463, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 14.31, "ce": 0.030614103889092803, "kl": 0.026103565003722906, "context": 0.04366588685661554}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.6251454949378967, "loss_scale": 4.0, "retries": 0, "seconds": 92.0, "max_memory_gb": 14.31, "ce": 0.11050748638808727, "kl": 0.040393507573753595, "context": 0.1307042392436415}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.703927218914032, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 14.31, "ce": 0.07976291514933109, "kl": 0.038652412476949394, "context": 0.09908912423998117}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 2.075810432434082, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 14.31, "ce": 0.17021225672215223, "kl": 0.033977999293711036, "context": 0.18720125500112772}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 1.1628206968307495, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 14.31, "ce": 0.07935505220666528, "kl": 0.015159784205025062, "context": 0.08693494251929224}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.265769898891449, "loss_scale": 4.0, "retries": 0, "seconds": 93.4, "max_memory_gb": 14.31, "ce": 0.16110702813602984, "kl": 0.029403779044514522, "context": 0.175808914587833}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.7475007772445679, "loss_scale": 4.0, "retries": 0, "seconds": 93.8, "max_memory_gb": 14.31, "ce": 0.04663923941552639, "kl": 0.11073211207985878, "context": 0.10200529545545578}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.25080054998397827, "loss_scale": 4.0, "retries": 0, "seconds": 94.2, "max_memory_gb": 14.31, "ce": 0.10357741103507578, "kl": 0.008263706462457776, "context": 0.10770926461555064}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 1.656009316444397, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 14.31, "ce": 0.1875477540306747, "kl": 0.2164029646664858, "context": 0.2957492284476757}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.4301762878894806, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 14.31, "ce": 0.19130316004157066, "kl": 0.20762819796800613, "context": 0.29511725530028343}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.4057801365852356, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 14.31, "ce": 0.10163538996130228, "kl": 0.0287636890043359, "context": 0.11601723358035088}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 0.21425668895244598, "loss_scale": 4.0, "retries": 0, "seconds": 95.6, "max_memory_gb": 14.31, "ce": 0.0757092492422089, "kl": 0.015482979459193302, "context": 0.08345074055250734}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 1.1261001825332642, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 14.31, "ce": 0.05969253974035382, "kl": 0.07184564700582996, "context": 0.09561535948887467}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.09781219065189362, "loss_scale": 4.0, "retries": 0, "seconds": 96.4, "max_memory_gb": 14.31, "ce": 0.03539888234809041, "kl": 0.03773068462032825, "context": 0.05426422320306301}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 0.5105493664741516, "loss_scale": 4.0, "retries": 0, "seconds": 96.7, "max_memory_gb": 14.31, "ce": 0.14563451474532485, "kl": 0.11612071329727769, "context": 0.20369486510753632}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 1.428447961807251, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 14.31, "ce": 0.1853469517081976, "kl": 0.12457763240672648, "context": 0.2476357650011778}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 0.567068338394165, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 14.31, "ce": 0.2636899910867214, "kl": 0.09108768892474473, "context": 0.30923382798209786}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 1.0640580654144287, "loss_scale": 4.0, "retries": 0, "seconds": 97.8, "max_memory_gb": 14.31, "ce": 0.2568715428933501, "kl": 0.037699414417147636, "context": 0.2757212556898594}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 0.5187311172485352, "loss_scale": 4.0, "retries": 0, "seconds": 98.2, "max_memory_gb": 14.31, "ce": 0.2141475083772093, "kl": 0.08445507648866624, "context": 0.25637504598125815}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 0.7007029056549072, "loss_scale": 4.0, "retries": 0, "seconds": 98.5, "max_memory_gb": 14.31, "ce": 0.10237612784840167, "kl": 0.07106641109567136, "context": 0.13790933322161436}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 3.107866048812866, "loss_scale": 4.0, "retries": 0, "seconds": 98.9, "max_memory_gb": 14.31, "ce": 0.2952358862385154, "kl": 0.11972690524999052, "context": 0.35509933484718204}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.2279573678970337, "loss_scale": 4.0, "retries": 0, "seconds": 99.3, "max_memory_gb": 14.31, "ce": 0.1606477713212371, "kl": 0.00706301664467901, "context": 0.16417927597649395}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.22967198491096497, "loss_scale": 4.0, "retries": 0, "seconds": 99.7, "max_memory_gb": 14.31, "ce": 0.25996352545917034, "kl": 0.05485515995678725, "context": 0.2873911107890308}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 3.4944145679473877, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 14.31, "ce": 0.3191868420690298, "kl": 0.4576716638293874, "context": 0.5480226706713438}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 0.5750933885574341, "loss_scale": 4.0, "retries": 0, "seconds": 100.4, "max_memory_gb": 14.31, "ce": 0.14073604391887784, "kl": 0.00919715991767589, "context": 0.1453346242196858}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 1.3173221349716187, "loss_scale": 4.0, "retries": 0, "seconds": 100.7, "max_memory_gb": 14.31, "ce": 0.09907513542566448, "kl": 0.04210229846648872, "context": 0.12012628675438464}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 1.0318779945373535, "loss_scale": 4.0, "retries": 0, "seconds": 101.1, "max_memory_gb": 14.31, "ce": 0.043617765652015805, "kl": 0.2907116948044859, "context": 0.18897361028939486}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 1.1470351219177246, "loss_scale": 4.0, "retries": 0, "seconds": 101.5, "max_memory_gb": 14.31, "ce": 0.038038400118239224, "kl": 0.19471991230966523, "context": 0.13539836183190346}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.9670892953872681, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 14.31, "ce": 0.16837702691555023, "kl": 0.05670011416077614, "context": 0.19672708213329315}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 0.7477917075157166, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 14.31, "ce": 0.4847379932180047, "kl": 0.38007332826964557, "context": 0.674774655723013}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 0.1738370805978775, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 14.31, "ce": 0.007148922770284116, "kl": 0.008581826914451085, "context": 0.011439836234785616}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.6390420198440552, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 14.31, "ce": 0.017208385514095426, "kl": 0.07880402728915215, "context": 0.0566103991586715}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 1.1881718635559082, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 14.31, "ce": 0.15329584700521082, "kl": 0.12534734916698653, "context": 0.21596952312393114}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 1.8662885427474976, "loss_scale": 4.0, "retries": 0, "seconds": 103.7, "max_memory_gb": 14.31, "ce": 0.11030941922217607, "kl": 0.0905863436346408, "context": 0.15560258831828833}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 1.0597494840621948, "loss_scale": 4.0, "retries": 0, "seconds": 104.0, "max_memory_gb": 14.31, "ce": 0.04792485130019486, "kl": 0.012217429844895378, "context": 0.054033564403653145}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.7854390144348145, "loss_scale": 4.0, "retries": 0, "seconds": 104.4, "max_memory_gb": 14.31, "ce": 0.2580411732196808, "kl": 0.06054002000018954, "context": 0.28831117786467075}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.19532524049282074, "loss_scale": 4.0, "retries": 0, "seconds": 104.7, "max_memory_gb": 14.31, "ce": 0.12847971508745104, "kl": 0.0074496598481346155, "context": 0.13220454845577478}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 1.2574868202209473, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 14.31, "ce": 0.09179927746299654, "kl": 0.13312601424695458, "context": 0.15836229163687676}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 1.8617385625839233, "loss_scale": 4.0, "retries": 0, "seconds": 105.5, "max_memory_gb": 14.31, "ce": 0.1393977713305503, "kl": 0.018266410968863056, "context": 0.1485309803392738}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 0.9618881344795227, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 14.31, "ce": 0.12194802181329578, "kl": 0.036399903445271775, "context": 0.14014797331765294}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 0.6180820465087891, "loss_scale": 4.0, "retries": 0, "seconds": 106.2, "max_memory_gb": 14.31, "ce": 0.021196749759837985, "kl": 0.017411263863323256, "context": 0.029902380891144276}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 0.5289891362190247, "loss_scale": 4.0, "retries": 0, "seconds": 106.6, "max_memory_gb": 14.31, "ce": 0.1837848387658596, "kl": 0.042744535545352846, "context": 0.20515710208564997}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.2967289388179779, "loss_scale": 4.0, "retries": 0, "seconds": 106.9, "max_memory_gb": 14.31, "ce": 0.009363664663396776, "kl": 0.015017723155324347, "context": 0.016872526612132788}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 0.5985813736915588, "loss_scale": 4.0, "retries": 0, "seconds": 107.3, "max_memory_gb": 14.31, "ce": 0.09225414087995887, "kl": 0.054489724221639335, "context": 0.1194989993236959}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.4387948215007782, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 14.31, "ce": 0.17968773999018595, "kl": 0.06185844324136269, "context": 0.2106169594626408}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 0.1572047919034958, "loss_scale": 4.0, "retries": 0, "seconds": 108.0, "max_memory_gb": 14.31, "ce": 0.009086038218811154, "kl": 0.0037856036669836612, "context": 0.010978839942254126}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 1.6165573596954346, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 14.31, "ce": 0.08268450130708516, "kl": 0.59466602653265, "context": 0.3800175227224827}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.8431177139282227, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 14.31, "ce": 0.0957316083367914, "kl": 0.09183094017498661, "context": 0.14164707786403596}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 1.218407154083252, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 14.31, "ce": 0.08282697852700949, "kl": 0.03709206450730562, "context": 0.10137300938367844}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 0.9065765142440796, "loss_scale": 4.0, "retries": 0, "seconds": 109.5, "max_memory_gb": 14.31, "ce": 0.29978213366121054, "kl": 0.07797644595848396, "context": 0.33877036394551396}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 1.398134469985962, "loss_scale": 4.0, "retries": 0, "seconds": 109.8, "max_memory_gb": 14.31, "ce": 0.20074373297393322, "kl": 0.3937569020781666, "context": 0.3976221792399883}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 0.1346425861120224, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 14.31, "ce": 0.090867965715006, "kl": 0.02527189743705094, "context": 0.1035039119888097}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 3.4023680686950684, "loss_scale": 4.0, "retries": 0, "seconds": 110.6, "max_memory_gb": 14.31, "ce": 0.44954719906672835, "kl": 0.13377415761351585, "context": 0.5164342783391476}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.7263487577438354, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 14.31, "ce": 0.013799991575069726, "kl": 0.09208661613229197, "context": 0.059843298920895904}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 0.35208067297935486, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 14.31, "ce": 0.2649983528535813, "kl": 0.039798114928998984, "context": 0.28489741683006287}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 0.3304535746574402, "loss_scale": 4.0, "retries": 0, "seconds": 111.6, "max_memory_gb": 14.31, "ce": 0.1569490220863372, "kl": 0.1105306938115973, "context": 0.21221437386702746}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 1.0289208889007568, "loss_scale": 4.0, "retries": 0, "seconds": 112.0, "max_memory_gb": 14.31, "ce": 0.056565254693850875, "kl": 0.011670444837363902, "context": 0.06240047770552337}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 1.0625702142715454, "loss_scale": 4.0, "retries": 0, "seconds": 112.4, "max_memory_gb": 14.31, "ce": 0.04135882778791711, "kl": 0.03521427287341794, "context": 0.05896596540696919}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 0.4749503433704376, "loss_scale": 4.0, "retries": 0, "seconds": 112.7, "max_memory_gb": 14.31, "ce": 0.03214879659935832, "kl": 0.00849083130742656, "context": 0.03639421286061406}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 2.1293697357177734, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 14.31, "ce": 0.1082959771156311, "kl": 0.12869871605653316, "context": 0.1726453424198553}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 1.5417654514312744, "loss_scale": 4.0, "retries": 0, "seconds": 113.5, "max_memory_gb": 14.31, "ce": 0.02334614936262369, "kl": 0.11992990621365607, "context": 0.08331110188737512}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 0.2373412847518921, "loss_scale": 4.0, "retries": 0, "seconds": 113.8, "max_memory_gb": 14.31, "ce": 0.0766428573988378, "kl": 0.03043405874632299, "context": 0.09185988525860012}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 0.8368380069732666, "loss_scale": 4.0, "retries": 0, "seconds": 114.2, "max_memory_gb": 14.31, "ce": 0.029449158813804388, "kl": 0.5460846464266069, "context": 0.30249149235896766}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 0.3640943765640259, "loss_scale": 4.0, "retries": 0, "seconds": 114.6, "max_memory_gb": 14.31, "ce": 0.18945383315440267, "kl": 0.022609102335991338, "context": 0.20075838384218514}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 1.2983474731445312, "loss_scale": 4.0, "retries": 0, "seconds": 114.9, "max_memory_gb": 14.31, "ce": 0.03286489297170192, "kl": 0.18571251671528444, "context": 0.12572114844806492}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 0.2208680361509323, "loss_scale": 4.0, "retries": 0, "seconds": 115.3, "max_memory_gb": 14.31, "ce": 0.13151530222967267, "kl": 0.019220738678995986, "context": 0.14112566830590367}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.3271016776561737, "loss_scale": 4.0, "retries": 0, "seconds": 115.7, "max_memory_gb": 14.31, "ce": 0.020607862039469182, "kl": 0.014438828395213932, "context": 0.027827276615425944}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 0.7521793842315674, "loss_scale": 4.0, "retries": 0, "seconds": 116.0, "max_memory_gb": 14.31, "ce": 0.13183222338557243, "kl": 0.043987963697873056, "context": 0.15382620366290212}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 0.4843585193157196, "loss_scale": 4.0, "retries": 0, "seconds": 116.4, "max_memory_gb": 14.31, "ce": 0.09860664140433073, "kl": 0.09456087989383377, "context": 0.14588708127848804}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.6143202781677246, "loss_scale": 4.0, "retries": 0, "seconds": 116.8, "max_memory_gb": 14.31, "ce": 0.06616710394155234, "kl": 0.05160799389705062, "context": 0.0919710963498801}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 1.2168720960617065, "loss_scale": 4.0, "retries": 0, "seconds": 117.1, "max_memory_gb": 14.31, "ce": 0.1841940744780004, "kl": 0.5457314937375486, "context": 0.4570598192512989}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 1.416592001914978, "loss_scale": 4.0, "retries": 0, "seconds": 117.5, "max_memory_gb": 14.31, "ce": 0.23009744100272655, "kl": 0.029336932580918074, "context": 0.24476590380072594}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 0.8932021260261536, "loss_scale": 4.0, "retries": 0, "seconds": 117.8, "max_memory_gb": 14.31, "ce": 0.049461312126368284, "kl": 0.34705474153452087, "context": 0.22298868279904127}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 1.4875855445861816, "loss_scale": 4.0, "retries": 0, "seconds": 118.2, "max_memory_gb": 14.31, "ce": 0.10840940056368709, "kl": 0.08883620891720057, "context": 0.15282750688493252}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 2.458745241165161, "loss_scale": 4.0, "retries": 0, "seconds": 118.6, "max_memory_gb": 14.31, "ce": 0.15366362128406763, "kl": 0.08894505236821715, "context": 0.1981361408252269}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 0.33133450150489807, "loss_scale": 4.0, "retries": 0, "seconds": 118.9, "max_memory_gb": 14.31, "ce": 0.13072243193164468, "kl": 0.09349553845822811, "context": 0.17747020116075873}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 0.5574939250946045, "loss_scale": 4.0, "retries": 0, "seconds": 119.3, "max_memory_gb": 14.31, "ce": 0.025847767246887088, "kl": 0.03947418072493747, "context": 0.04558485746383667}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 1.9480500221252441, "loss_scale": 4.0, "retries": 0, "seconds": 119.7, "max_memory_gb": 14.31, "ce": 0.02255943138152361, "kl": 0.14223233447410166, "context": 0.09367559757083654}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.5507637858390808, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 14.31, "ce": 0.02141769591253251, "kl": 0.01489413998206146, "context": 0.028864766471087933}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 0.3324097990989685, "loss_scale": 4.0, "retries": 0, "seconds": 120.4, "max_memory_gb": 14.31, "ce": 0.22947086638305336, "kl": 0.015563913140795194, "context": 0.23725281993392855}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.32668495178222656, "loss_scale": 4.0, "retries": 0, "seconds": 120.8, "max_memory_gb": 14.31, "ce": 0.05330491531640291, "kl": 0.033916539410711266, "context": 0.07026318460702896}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 2.9075100421905518, "loss_scale": 4.0, "retries": 0, "seconds": 121.1, "max_memory_gb": 14.31, "ce": 0.15664495178498328, "kl": 0.07422880785088637, "context": 0.19375935569405556}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 0.9821370840072632, "loss_scale": 4.0, "retries": 0, "seconds": 121.5, "max_memory_gb": 14.31, "ce": 0.11175296315923333, "kl": 0.14518685568327783, "context": 0.18434638809412718}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 0.40726879239082336, "loss_scale": 4.0, "retries": 0, "seconds": 121.9, "max_memory_gb": 14.31, "ce": 0.07697215676307678, "kl": 0.17958385980455205, "context": 0.1667640944942832}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 0.7652692794799805, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 14.31, "ce": 0.17340291605796665, "kl": 0.0066880641534226015, "context": 0.17674694361630827}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 0.9514976143836975, "loss_scale": 4.0, "retries": 0, "seconds": 122.6, "max_memory_gb": 14.31, "ce": 0.2583748325705528, "kl": 0.09006859554210678, "context": 0.3034091340377927}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 2.5861003398895264, "loss_scale": 4.0, "retries": 0, "seconds": 123.0, "max_memory_gb": 14.31, "ce": 0.18120430945418775, "kl": 0.03718762827338651, "context": 0.1997981274034828}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 0.3271676003932953, "loss_scale": 4.0, "retries": 0, "seconds": 123.3, "max_memory_gb": 14.31, "ce": 0.27532730298116803, "kl": 0.029601240064948797, "context": 0.2901279181241989}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 0.6358893513679504, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 14.31, "ce": 0.1719094100408256, "kl": 0.1139100578147918, "context": 0.22886443883180618}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 4.112718105316162, "loss_scale": 4.0, "retries": 0, "seconds": 124.1, "max_memory_gb": 14.31, "ce": 0.31666594720445573, "kl": 0.19207550294231623, "context": 0.4127036896534264}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 1.2248202562332153, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 14.31, "ce": 0.039254104951396585, "kl": 0.09801467362194671, "context": 0.08826144319027662}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 2.1300652027130127, "loss_scale": 4.0, "retries": 0, "seconds": 124.8, "max_memory_gb": 14.31, "ce": 0.25019092578440905, "kl": 0.12140785623341799, "context": 0.31089484971016645}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.2612190246582031, "loss_scale": 4.0, "retries": 0, "seconds": 125.2, "max_memory_gb": 14.31, "ce": 0.1830805684439838, "kl": 0.17324923817068338, "context": 0.26970518846064806}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.37631702423095703, "loss_scale": 4.0, "retries": 0, "seconds": 125.5, "max_memory_gb": 14.31, "ce": 0.13199244032148272, "kl": 0.0155252909171395, "context": 0.1397550837136805}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 0.5367510914802551, "loss_scale": 4.0, "retries": 0, "seconds": 125.9, "max_memory_gb": 14.31, "ce": 0.19871649658307433, "kl": 0.07001280515396502, "context": 0.23372290330007672}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 2.207207202911377, "loss_scale": 4.0, "retries": 0, "seconds": 126.3, "max_memory_gb": 14.31, "ce": 0.25079531501978636, "kl": 0.05153688881546259, "context": 0.27656376734375954}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 0.8738502860069275, "loss_scale": 4.0, "retries": 0, "seconds": 126.6, "max_memory_gb": 14.31, "ce": 0.23917656671255827, "kl": 0.7860674561234191, "context": 0.6322103030979633}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.4276536703109741, "loss_scale": 4.0, "retries": 0, "seconds": 127.0, "max_memory_gb": 14.31, "ce": 0.018885910511016846, "kl": 0.035362723167054355, "context": 0.0365672716870904}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.7881066203117371, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 14.31, "ce": 0.03809846518561244, "kl": 0.019014363235328346, "context": 0.04760564723983407}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.25847065448760986, "loss_scale": 4.0, "retries": 0, "seconds": 127.7, "max_memory_gb": 14.31, "ce": 0.13008596215513535, "kl": 0.1854943026555702, "context": 0.22283311345381662}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 0.6206569671630859, "loss_scale": 4.0, "retries": 0, "seconds": 128.1, "max_memory_gb": 14.31, "ce": 0.23141734290402383, "kl": 0.010732407157775015, "context": 0.23678355058655143}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 1.393944501876831, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 14.31, "ce": 0.17379465769045055, "kl": 0.022874692105688155, "context": 0.18523200461640954}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.08962032198905945, "loss_scale": 4.0, "retries": 0, "seconds": 128.8, "max_memory_gb": 14.31, "ce": 0.009749540186021477, "kl": 0.004014008361991728, "context": 0.011756544525269419}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.214150533080101, "loss_scale": 4.0, "retries": 0, "seconds": 129.2, "max_memory_gb": 14.31, "ce": 0.09411904832813889, "kl": 0.016777964679931756, "context": 0.10250802885275334}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.20643174648284912, "loss_scale": 4.0, "retries": 0, "seconds": 129.6, "max_memory_gb": 14.31, "ce": 0.13930746167898178, "kl": 0.05697516747750342, "context": 0.16779504157602787}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 1.5694211721420288, "loss_scale": 4.0, "retries": 0, "seconds": 129.9, "max_memory_gb": 14.31, "ce": 0.029609482269734144, "kl": 0.28644418925978243, "context": 0.17283157259225845}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 1.771167278289795, "loss_scale": 4.0, "retries": 0, "seconds": 130.3, "max_memory_gb": 14.31, "ce": 0.3066183845512569, "kl": 0.330358037725091, "context": 0.4717974103987217}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 0.33217865228652954, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 14.31, "ce": 0.14389180217403919, "kl": 0.06436280643538339, "context": 0.1760732129914686}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 1.6945315599441528, "loss_scale": 4.0, "retries": 0, "seconds": 131.0, "max_memory_gb": 14.31, "ce": 0.24320237385109067, "kl": 0.02093954554584343, "context": 0.25367214623838663}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.9553373456001282, "loss_scale": 4.0, "retries": 0, "seconds": 131.3, "max_memory_gb": 14.31, "ce": 0.14951607072725892, "kl": 0.02839632681570947, "context": 0.16371423832606524}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 0.20307816565036774, "loss_scale": 4.0, "retries": 0, "seconds": 131.7, "max_memory_gb": 14.31, "ce": 0.052342968760058284, "kl": 0.031587483623297885, "context": 0.06813671253621578}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 0.4005887806415558, "loss_scale": 4.0, "retries": 0, "seconds": 132.1, "max_memory_gb": 14.31, "ce": 0.08961481100413948, "kl": 0.030282962427008897, "context": 0.10475629032589495}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 0.3307274878025055, "loss_scale": 4.0, "retries": 0, "seconds": 132.4, "max_memory_gb": 14.31, "ce": 0.21269873133860528, "kl": 0.015549990348517895, "context": 0.22047373151872307}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 0.44913825392723083, "loss_scale": 4.0, "retries": 0, "seconds": 132.8, "max_memory_gb": 14.31, "ce": 0.02295802836306393, "kl": 0.022142572721350007, "context": 0.03402931476011872}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 0.26576539874076843, "loss_scale": 4.0, "retries": 0, "seconds": 133.1, "max_memory_gb": 14.31, "ce": 0.08470721705816686, "kl": 0.02154073180281557, "context": 0.09547758218832314}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.3165290355682373, "loss_scale": 4.0, "retries": 0, "seconds": 133.5, "max_memory_gb": 14.31, "ce": 0.08349262457340956, "kl": 0.30008280809852295, "context": 0.23353402130305767}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 4.229637622833252, "loss_scale": 4.0, "retries": 0, "seconds": 133.8, "max_memory_gb": 14.31, "ce": 0.3188904217677191, "kl": 0.04482032726809848, "context": 0.3413005833281204}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.39534470438957214, "loss_scale": 4.0, "retries": 0, "seconds": 134.2, "max_memory_gb": 14.31, "ce": 0.07601873483508825, "kl": 0.09538323991000652, "context": 0.12371035851538181}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 2.2292442321777344, "loss_scale": 4.0, "retries": 0, "seconds": 134.6, "max_memory_gb": 14.31, "ce": 0.009189114934997633, "kl": 0.48065935038539465, "context": 0.2495187831518706}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 0.34801536798477173, "loss_scale": 4.0, "retries": 0, "seconds": 134.9, "max_memory_gb": 14.31, "ce": 0.15901263838168234, "kl": 0.01911199356254656, "context": 0.1685686365235597}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 0.9731701016426086, "loss_scale": 4.0, "retries": 0, "seconds": 135.3, "max_memory_gb": 14.31, "ce": 0.07723060296848416, "kl": 0.28161870362237096, "context": 0.218039957806468}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.420248806476593, "loss_scale": 4.0, "retries": 0, "seconds": 135.6, "max_memory_gb": 14.31, "ce": 0.11719435569830239, "kl": 0.09928725883946754, "context": 0.16683799237944186}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.9914522171020508, "loss_scale": 4.0, "retries": 0, "seconds": 136.0, "max_memory_gb": 14.31, "ce": 0.0321676058229059, "kl": 0.036199160225805826, "context": 0.0502671857830137}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.26457440853118896, "loss_scale": 4.0, "retries": 0, "seconds": 136.4, "max_memory_gb": 14.31, "ce": 0.06673511117696762, "kl": 0.013676275935722515, "context": 0.07357324555050582}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 0.9023983478546143, "loss_scale": 4.0, "retries": 0, "seconds": 136.7, "max_memory_gb": 14.31, "ce": 0.4784196466207504, "kl": 0.16283246187958866, "context": 0.5598358754068613}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 2.1875522136688232, "loss_scale": 4.0, "retries": 0, "seconds": 137.1, "max_memory_gb": 14.31, "ce": 0.09544078970793635, "kl": 0.019724039444554364, "context": 0.1053028057795018}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.4428478181362152, "loss_scale": 4.0, "retries": 0, "seconds": 137.4, "max_memory_gb": 14.31, "ce": 0.0453999787569046, "kl": 0.03332269644306507, "context": 0.06206132797524333}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 0.24657104909420013, "loss_scale": 4.0, "retries": 0, "seconds": 137.8, "max_memory_gb": 14.31, "ce": 0.10154191753827035, "kl": 0.05452938535017893, "context": 0.12880660919472575}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 0.4586780369281769, "loss_scale": 4.0, "retries": 0, "seconds": 138.2, "max_memory_gb": 14.31, "ce": 0.18423031619749963, "kl": 0.11493370201424113, "context": 0.24169716745382175}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 1.3857790231704712, "loss_scale": 4.0, "retries": 0, "seconds": 138.5, "max_memory_gb": 14.31, "ce": 0.19733261072542518, "kl": 0.13233651369228028, "context": 0.26350086473394185}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.3957177400588989, "loss_scale": 4.0, "retries": 0, "seconds": 138.9, "max_memory_gb": 14.31, "ce": 0.1482635955326259, "kl": 0.04222115103038959, "context": 0.16937417478766292}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.2336151897907257, "loss_scale": 4.0, "retries": 0, "seconds": 139.2, "max_memory_gb": 14.31, "ce": 0.0808139011496678, "kl": 0.11328532565676142, "context": 0.13745655911043286}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 1.1699085235595703, "loss_scale": 4.0, "retries": 0, "seconds": 139.6, "max_memory_gb": 14.31, "ce": 0.09962365496903658, "kl": 0.045992499392014, "context": 0.12261990457773209}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.1887078583240509, "loss_scale": 4.0, "retries": 0, "seconds": 140.0, "max_memory_gb": 14.31, "ce": 0.09935211413539946, "kl": 0.05455891140445601, "context": 0.12663156911730766}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 0.7793645858764648, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 14.31, "ce": 0.040252924780361354, "kl": 0.08842619619099423, "context": 0.08446602197363973}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 0.44144994020462036, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 14.31, "ce": 0.2613605195656419, "kl": 0.1111265558283776, "context": 0.316923797596246}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 2.4771461486816406, "loss_scale": 4.0, "retries": 0, "seconds": 141.0, "max_memory_gb": 14.31, "ce": 0.2989145520259626, "kl": 0.2418777879793197, "context": 0.41985345439752564}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.2720656991004944, "loss_scale": 4.0, "retries": 0, "seconds": 141.4, "max_memory_gb": 14.31, "ce": 0.0708554545417428, "kl": 0.06454160519933794, "context": 0.1031262535834685}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 0.4206909239292145, "loss_scale": 4.0, "retries": 0, "seconds": 141.7, "max_memory_gb": 14.31, "ce": 0.01850821857806295, "kl": 0.011670637206407264, "context": 0.024343537399545312}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.1901288479566574, "loss_scale": 4.0, "retries": 0, "seconds": 142.1, "max_memory_gb": 14.31, "ce": 0.2044480019249022, "kl": 0.005210148970945738, "context": 0.20705307484604418}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.30181533098220825, "loss_scale": 4.0, "retries": 0, "seconds": 142.5, "max_memory_gb": 14.31, "ce": 0.08523623552173376, "kl": 0.025295216853919555, "context": 0.0978838438168168}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.5557927489280701, "loss_scale": 4.0, "retries": 0, "seconds": 142.8, "max_memory_gb": 14.31, "ce": 0.0877339135913644, "kl": 0.018573898472823203, "context": 0.09702086297329515}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 0.23887835443019867, "loss_scale": 4.0, "retries": 0, "seconds": 143.2, "max_memory_gb": 14.31, "ce": 0.04772240965394303, "kl": 0.005760920932516456, "context": 0.05060287017840892}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.3747493326663971, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 14.31, "ce": 0.01578448247164488, "kl": 0.01628956914646551, "context": 0.023929266957566142}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 3.752995491027832, "loss_scale": 4.0, "retries": 0, "seconds": 143.9, "max_memory_gb": 14.31, "ce": 0.13334393128752708, "kl": 0.10343599878251553, "context": 0.18506192229688168}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 1.2717971801757812, "loss_scale": 4.0, "retries": 0, "seconds": 144.3, "max_memory_gb": 14.31, "ce": 0.27158039063215256, "kl": 0.17026906873797998, "context": 0.3567149303853512}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.4263129234313965, "loss_scale": 4.0, "retries": 0, "seconds": 144.7, "max_memory_gb": 14.31, "ce": 0.00835593169904314, "kl": 0.013308703972143121, "context": 0.015010283677838743}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 1.3947045803070068, "loss_scale": 4.0, "retries": 0, "seconds": 145.0, "max_memory_gb": 14.31, "ce": 0.029959620849695057, "kl": 0.6747110828291625, "context": 0.3673151582479477}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 1.036910891532898, "loss_scale": 4.0, "retries": 0, "seconds": 145.4, "max_memory_gb": 14.31, "ce": 0.12170990597223863, "kl": 0.026296840514987707, "context": 0.1348583239596337}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.2552401125431061, "loss_scale": 4.0, "retries": 0, "seconds": 145.7, "max_memory_gb": 14.31, "ce": 0.058646538062021136, "kl": 0.019037576741538942, "context": 0.06816532730590552}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 1.6262271404266357, "loss_scale": 4.0, "retries": 0, "seconds": 146.1, "max_memory_gb": 14.31, "ce": 0.009786523529328406, "kl": 0.20109608047641814, "context": 0.11033456330187619}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.48375335335731506, "loss_scale": 4.0, "retries": 0, "seconds": 146.4, "max_memory_gb": 14.31, "ce": 0.14794959110440686, "kl": 0.08833315323863644, "context": 0.19211616867687553}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.6920018196105957, "loss_scale": 4.0, "retries": 0, "seconds": 146.8, "max_memory_gb": 14.31, "ce": 0.09684320911765099, "kl": 0.005329925799742341, "context": 0.09950817399658263}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 0.42058077454566956, "loss_scale": 4.0, "retries": 0, "seconds": 147.2, "max_memory_gb": 14.31, "ce": 0.13795829885202693, "kl": 0.0015713785851403372, "context": 0.1387439952013665}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 3.2477211952209473, "loss_scale": 4.0, "retries": 0, "seconds": 147.5, "max_memory_gb": 14.31, "ce": 0.10170532111078501, "kl": 0.1217685786250513, "context": 0.16258961777202785}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 1.7196271419525146, "loss_scale": 4.0, "retries": 0, "seconds": 147.9, "max_memory_gb": 14.31, "ce": 0.05990470573306084, "kl": 0.5979971770429984, "context": 0.358903284650296}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 0.47833743691444397, "loss_scale": 4.0, "retries": 0, "seconds": 148.3, "max_memory_gb": 14.31, "ce": 0.16095658391714096, "kl": 0.09641778748482466, "context": 0.20916547812521458}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.08211224526166916, "loss_scale": 4.0, "retries": 0, "seconds": 148.6, "max_memory_gb": 14.31, "ce": 0.006918733182828873, "kl": 0.011340068856952712, "context": 0.012588767800480127}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 0.5356481075286865, "loss_scale": 4.0, "retries": 0, "seconds": 149.0, "max_memory_gb": 14.31, "ce": 0.15026970533654094, "kl": 0.07176358670039917, "context": 0.18615149200195447}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.6254556179046631, "loss_scale": 4.0, "retries": 0, "seconds": 149.4, "max_memory_gb": 14.31, "ce": 0.07628146355273202, "kl": 0.04445605759974569, "context": 0.09850948909297585}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.4212399125099182, "loss_scale": 4.0, "retries": 0, "seconds": 149.7, "max_memory_gb": 14.31, "ce": 0.20877026510424912, "kl": 0.030447128927335143, "context": 0.2239938270067796}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 1.2519645690917969, "loss_scale": 4.0, "retries": 0, "seconds": 150.1, "max_memory_gb": 14.31, "ce": 0.7718445058562793, "kl": 0.3850274084834382, "context": 0.9643582365242764}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 0.5011269450187683, "loss_scale": 4.0, "retries": 0, "seconds": 150.5, "max_memory_gb": 14.31, "ce": 0.2632060842588544, "kl": 0.2550264187302673, "context": 0.3907192829065025}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 1.525651454925537, "loss_scale": 4.0, "retries": 0, "seconds": 150.9, "max_memory_gb": 14.31, "ce": 0.21175905130803585, "kl": 0.07785016216621443, "context": 0.2506841277354397}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 1.3685733079910278, "loss_scale": 4.0, "retries": 0, "seconds": 151.2, "max_memory_gb": 14.31, "ce": 0.0405505548696965, "kl": 0.0849538648762973, "context": 0.0830274845357053}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 4.418729782104492, "loss_scale": 4.0, "retries": 0, "seconds": 151.6, "max_memory_gb": 14.31, "ce": 0.23794287513010204, "kl": 0.20495051980833523, "context": 0.34041813132353127}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.337637722492218, "loss_scale": 4.0, "retries": 0, "seconds": 152.0, "max_memory_gb": 14.31, "ce": 0.15761601354461163, "kl": 0.06008016207488254, "context": 0.18765609082765877}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 0.3944678008556366, "loss_scale": 4.0, "retries": 0, "seconds": 152.3, "max_memory_gb": 14.31, "ce": 0.10536407283507288, "kl": 0.07644789159530774, "context": 0.14358801674097776}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 0.4783066213130951, "loss_scale": 4.0, "retries": 0, "seconds": 152.7, "max_memory_gb": 14.31, "ce": 0.17260691191768274, "kl": 0.011460136520327069, "context": 0.17833698529284447}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 1.0536915063858032, "loss_scale": 4.0, "retries": 0, "seconds": 153.0, "max_memory_gb": 14.31, "ce": 0.2676277848659083, "kl": 0.056402639020234346, "context": 0.2958291086833924}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.2564678490161896, "loss_scale": 4.0, "retries": 0, "seconds": 153.4, "max_memory_gb": 14.31, "ce": 0.011766023933887482, "kl": 0.004036650717353041, "context": 0.013784348731860518}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 0.3815477788448334, "loss_scale": 4.0, "retries": 0, "seconds": 153.7, "max_memory_gb": 14.31, "ce": 0.37473728973418474, "kl": 0.07203003275208175, "context": 0.4107523011043668}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.27782705426216125, "loss_scale": 4.0, "retries": 0, "seconds": 154.1, "max_memory_gb": 14.31, "ce": 0.0349049714859575, "kl": 0.03772663284325972, "context": 0.05376828694716096}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.2547284960746765, "loss_scale": 4.0, "retries": 0, "seconds": 154.5, "max_memory_gb": 14.31, "ce": 0.06842984486138448, "kl": 0.04873300555482274, "context": 0.09279634791892022}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 1.5966328382492065, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 14.31, "ce": 0.12485499994363636, "kl": 0.7950892513617873, "context": 0.5223996071144938}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.2771642804145813, "loss_scale": 4.0, "retries": 0, "seconds": 155.2, "max_memory_gb": 14.31, "ce": 0.07367190660443157, "kl": 0.02953979493759107, "context": 0.08844180358573794}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 2.177163600921631, "loss_scale": 4.0, "retries": 0, "seconds": 155.5, "max_memory_gb": 14.31, "ce": 0.22511215909617022, "kl": 0.014580205614038277, "context": 0.23240225267363712}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 1.9310963153839111, "loss_scale": 4.0, "retries": 0, "seconds": 155.9, "max_memory_gb": 14.31, "ce": 0.27606816857587546, "kl": 0.439575276337564, "context": 0.49585580825805664}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.44921159744262695, "loss_scale": 4.0, "retries": 0, "seconds": 156.3, "max_memory_gb": 14.31, "ce": 0.06837618356803432, "kl": 0.035494816838763654, "context": 0.0861235938500613}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.4394936263561249, "loss_scale": 4.0, "retries": 0, "seconds": 156.6, "max_memory_gb": 14.31, "ce": 0.13270846474915743, "kl": 0.12451629179122392, "context": 0.1949666105210781}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 0.8818485140800476, "loss_scale": 4.0, "retries": 0, "seconds": 157.0, "max_memory_gb": 14.31, "ce": 0.2947672316804528, "kl": 0.11470122542232275, "context": 0.35211784578859806}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.13454756140708923, "loss_scale": 4.0, "retries": 0, "seconds": 157.4, "max_memory_gb": 14.31, "ce": 0.059335653044399805, "kl": 0.012300595338274434, "context": 0.06548595205822494}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 2.058337926864624, "loss_scale": 4.0, "retries": 0, "seconds": 157.8, "max_memory_gb": 14.31, "ce": 0.009751963196322322, "kl": 0.3446338545272738, "context": 0.1820688956649974}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 1.432750940322876, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 14.31, "ce": 0.015429208462592214, "kl": 0.18754253338556737, "context": 0.10920047201216221}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.37034788727760315, "loss_scale": 4.0, "retries": 0, "seconds": 158.5, "max_memory_gb": 14.31, "ce": 0.1356837273342535, "kl": 0.011408312901039608, "context": 0.14138787996489555}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 4.689816951751709, "loss_scale": 4.0, "retries": 0, "seconds": 158.9, "max_memory_gb": 14.31, "ce": 0.2163012232631445, "kl": 0.42106051403970923, "context": 0.4268314951332286}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 1.1071134805679321, "loss_scale": 4.0, "retries": 0, "seconds": 159.2, "max_memory_gb": 14.31, "ce": 0.08413726260187104, "kl": 0.10624081746209413, "context": 0.13725767098367214}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 1.6622703075408936, "loss_scale": 4.0, "retries": 0, "seconds": 159.6, "max_memory_gb": 14.31, "ce": 0.12360554677434266, "kl": 0.1095723319449462, "context": 0.17839171458035707}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.07854849100112915, "loss_scale": 4.0, "retries": 0, "seconds": 159.9, "max_memory_gb": 14.31, "ce": 0.008915090467780828, "kl": 0.00038564899114135187, "context": 0.00910791486967355}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 0.6574382185935974, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 14.31, "ce": 0.09386458480730653, "kl": 0.05710152303799987, "context": 0.12241534900385886}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 0.6256470680236816, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 14.31, "ce": 0.13202826865017414, "kl": 0.020632082363590598, "context": 0.14234431087970734}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 1.431325912475586, "loss_scale": 4.0, "retries": 0, "seconds": 161.0, "max_memory_gb": 14.31, "ce": 0.14084942010231316, "kl": 0.09634891572204651, "context": 0.18902387632988393}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.9976646900177002, "loss_scale": 4.0, "retries": 0, "seconds": 161.4, "max_memory_gb": 14.31, "ce": 0.11803725769277662, "kl": 0.03835298601188697, "context": 0.13721375190652907}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 3.650914192199707, "loss_scale": 4.0, "retries": 0, "seconds": 161.8, "max_memory_gb": 14.31, "ce": 0.19949788693338633, "kl": 0.22112897667102516, "context": 0.31006237119436264}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.1536405235528946, "loss_scale": 4.0, "retries": 0, "seconds": 162.1, "max_memory_gb": 14.31, "ce": 0.05513369326945394, "kl": 0.07240246908258996, "context": 0.09133492759428918}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 1.7043322324752808, "loss_scale": 4.0, "retries": 0, "seconds": 162.5, "max_memory_gb": 14.31, "ce": 0.20387478329939768, "kl": 0.13470695618889295, "context": 0.2712282572174445}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.751120924949646, "loss_scale": 4.0, "retries": 0, "seconds": 162.8, "max_memory_gb": 14.31, "ce": 0.09615042363293469, "kl": 0.04790935915661976, "context": 0.12010510347317904}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.2563915252685547, "loss_scale": 4.0, "retries": 0, "seconds": 163.2, "max_memory_gb": 14.31, "ce": 0.08886298420839012, "kl": 0.037679223765735514, "context": 0.10770259494893253}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.4027453064918518, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 14.31, "ce": 0.12563326326198876, "kl": 0.03915674751624465, "context": 0.1452116386499256}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 0.472164124250412, "loss_scale": 4.0, "retries": 0, "seconds": 163.9, "max_memory_gb": 14.31, "ce": 0.3401802098087501, "kl": 0.028507111218459613, "context": 0.3544337747298414}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 3.0616610050201416, "loss_scale": 4.0, "retries": 0, "seconds": 164.3, "max_memory_gb": 14.31, "ce": 0.15462511847727, "kl": 0.11259099621383939, "context": 0.21092061791568995}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.5897706151008606, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 14.31, "ce": 0.266563777346164, "kl": 0.021482867326994892, "context": 0.2773052058182657}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 3.0473432540893555, "loss_scale": 4.0, "retries": 0, "seconds": 165.0, "max_memory_gb": 14.31, "ce": 0.3094961568713188, "kl": 0.16885441064368933, "context": 0.3939233534038067}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.9796382188796997, "loss_scale": 4.0, "retries": 0, "seconds": 165.4, "max_memory_gb": 14.31, "ce": 0.03157853183802217, "kl": 0.017131876993516926, "context": 0.04014446947257966}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 1.5787616968154907, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 14.31, "ce": 0.031067695847013965, "kl": 0.03599263932846952, "context": 0.04906401364132762}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 0.4190909266471863, "loss_scale": 4.0, "retries": 0, "seconds": 166.1, "max_memory_gb": 14.31, "ce": 0.013593235751613975, "kl": 0.04688144288957119, "context": 0.03703395673073828}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.8497346639633179, "loss_scale": 4.0, "retries": 0, "seconds": 166.5, "max_memory_gb": 14.31, "ce": 0.1741468843538314, "kl": 0.07112571125617251, "context": 0.2097097419318743}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 0.5833030343055725, "loss_scale": 4.0, "retries": 0, "seconds": 166.8, "max_memory_gb": 14.31, "ce": 0.042177289025858045, "kl": 0.0332665394525975, "context": 0.0588105577044189}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 1.7772012948989868, "loss_scale": 4.0, "retries": 0, "seconds": 167.2, "max_memory_gb": 14.31, "ce": 0.1802591043524444, "kl": 0.009655906818807125, "context": 0.18508706288412213}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 2.980997323989868, "loss_scale": 4.0, "retries": 0, "seconds": 167.5, "max_memory_gb": 14.31, "ce": 0.16159636061638594, "kl": 0.17967915791086853, "context": 0.2514359448105097}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.2188466191291809, "loss_scale": 4.0, "retries": 0, "seconds": 167.9, "max_memory_gb": 14.31, "ce": 0.02434321795590222, "kl": 0.010902780224569142, "context": 0.02979460754431784}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.305998831987381, "loss_scale": 4.0, "retries": 0, "seconds": 168.3, "max_memory_gb": 14.31, "ce": 0.06034334469586611, "kl": 0.024020695622311905, "context": 0.07235369144473225}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 1.38188636302948, "loss_scale": 4.0, "retries": 0, "seconds": 168.6, "max_memory_gb": 14.31, "ce": 0.11857452266849577, "kl": 0.1498154144792352, "context": 0.19348222436383367}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 0.6848616003990173, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 14.31, "ce": 0.1909286780282855, "kl": 0.12256951874587685, "context": 0.252213436877355}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.3733515739440918, "loss_scale": 4.0, "retries": 0, "seconds": 169.4, "max_memory_gb": 14.31, "ce": 0.024858703720383346, "kl": 0.0171385697030928, "context": 0.033427988179028034}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 0.19948382675647736, "loss_scale": 4.0, "retries": 0, "seconds": 169.7, "max_memory_gb": 14.31, "ce": 0.17951648740563542, "kl": 0.15498385881073773, "context": 0.2570084151811898}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.2245630919933319, "loss_scale": 4.0, "retries": 0, "seconds": 170.1, "max_memory_gb": 14.31, "ce": 0.04313016580999829, "kl": 0.009026593106682412, "context": 0.047643460624385625}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.15402637422084808, "loss_scale": 4.0, "retries": 0, "seconds": 170.5, "max_memory_gb": 14.31, "ce": 0.033106357033830136, "kl": 0.0035388900359976105, "context": 0.03487580269575119}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.5018011927604675, "loss_scale": 4.0, "retries": 0, "seconds": 170.8, "max_memory_gb": 14.31, "ce": 0.018623962299898267, "kl": 0.004937177218380384, "context": 0.021092551411129534}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 1.2262115478515625, "loss_scale": 4.0, "retries": 0, "seconds": 171.2, "max_memory_gb": 14.31, "ce": 0.05465295963222161, "kl": 0.5918818068676046, "context": 0.35059387046203483}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 1.155137300491333, "loss_scale": 4.0, "retries": 0, "seconds": 171.5, "max_memory_gb": 14.31, "ce": 0.02157239020743873, "kl": 0.18336327015049392, "context": 0.11325402575312182}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 0.3197593092918396, "loss_scale": 4.0, "retries": 0, "seconds": 171.9, "max_memory_gb": 14.31, "ce": 0.05511110386578366, "kl": 0.02730874098779168, "context": 0.06876547436695546}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 0.21227103471755981, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 14.31, "ce": 0.034428386366926134, "kl": 0.01204546430381015, "context": 0.04045111918821931}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 2.612386465072632, "loss_scale": 4.0, "retries": 0, "seconds": 172.6, "max_memory_gb": 14.31, "ce": 0.12267971399705857, "kl": 0.0290307643590495, "context": 0.137195092625916}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.2511541843414307, "loss_scale": 4.0, "retries": 0, "seconds": 173.0, "max_memory_gb": 14.31, "ce": 0.006374268559738994, "kl": 0.41028685178025626, "context": 0.2115176945226267}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 1.8370200395584106, "loss_scale": 4.0, "retries": 0, "seconds": 173.3, "max_memory_gb": 14.31, "ce": 0.1299874068936333, "kl": 0.19984170170209836, "context": 0.2299082602839917}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 0.29427215456962585, "loss_scale": 4.0, "retries": 0, "seconds": 173.7, "max_memory_gb": 14.31, "ce": 0.010394750162959099, "kl": 0.007874210074078292, "context": 0.014331855345517397}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 1.3492783308029175, "loss_scale": 4.0, "retries": 0, "seconds": 174.1, "max_memory_gb": 14.31, "ce": 0.048851749219466, "kl": 0.026683078656787984, "context": 0.06219328992301598}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 0.31484946608543396, "loss_scale": 4.0, "retries": 0, "seconds": 174.4, "max_memory_gb": 14.31, "ce": 0.07973302982281893, "kl": 0.01743622138747014, "context": 0.08845114067662507}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 0.5237850546836853, "loss_scale": 4.0, "retries": 0, "seconds": 174.8, "max_memory_gb": 14.31, "ce": 0.13048790892935358, "kl": 0.015503283993894001, "context": 0.13823955337284133}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.7555148601531982, "loss_scale": 4.0, "retries": 0, "seconds": 175.2, "max_memory_gb": 14.31, "ce": 0.03231120129930787, "kl": 0.009796919417567551, "context": 0.03720966214314103}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 3.7713375091552734, "loss_scale": 4.0, "retries": 0, "seconds": 175.5, "max_memory_gb": 14.31, "ce": 0.017327452427707613, "kl": 0.7126100258392398, "context": 0.3736324631900061}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 0.21693965792655945, "loss_scale": 4.0, "retries": 0, "seconds": 175.9, "max_memory_gb": 14.31, "ce": 0.029008303390583023, "kl": 0.09100311980910192, "context": 0.07450986074400134}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 4.8051676750183105, "loss_scale": 4.0, "retries": 0, "seconds": 176.3, "max_memory_gb": 14.31, "ce": 0.456742333015427, "kl": 0.2962417582748458, "context": 0.6048632103484124}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.5572792887687683, "loss_scale": 4.0, "retries": 0, "seconds": 176.6, "max_memory_gb": 14.31, "ce": 0.21848873642738909, "kl": 0.011363326339051127, "context": 0.22417040343862027}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 0.5372088551521301, "loss_scale": 4.0, "retries": 0, "seconds": 177.0, "max_memory_gb": 14.31, "ce": 0.011950148269534111, "kl": 0.02629591792356223, "context": 0.025098106591030955}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 1.0816675424575806, "loss_scale": 4.0, "retries": 0, "seconds": 177.3, "max_memory_gb": 14.31, "ce": 0.11360226781107485, "kl": 0.015761647140607238, "context": 0.12148309254553169}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 1.1938039064407349, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 14.31, "ce": 0.15781723288819194, "kl": 0.551276998128742, "context": 0.4334557428956032}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 0.61989825963974, "loss_scale": 4.0, "retries": 0, "seconds": 178.1, "max_memory_gb": 14.31, "ce": 0.34648249484598637, "kl": 0.129030869808048, "context": 0.41099792439490557}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.6308428645133972, "loss_scale": 4.0, "retries": 0, "seconds": 178.4, "max_memory_gb": 14.31, "ce": 0.18416378274559975, "kl": 0.08915861579589546, "context": 0.22874309681355953}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 0.8188433051109314, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 14.31, "ce": 0.13713610410923138, "kl": 0.03037145547568798, "context": 0.15232183184707537}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 2.121814250946045, "loss_scale": 4.0, "retries": 0, "seconds": 179.1, "max_memory_gb": 14.31, "ce": 0.06978106035967357, "kl": 0.059362008643802255, "context": 0.09946206491440535}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 0.5423011779785156, "loss_scale": 4.0, "retries": 0, "seconds": 179.5, "max_memory_gb": 14.31, "ce": 0.1632269195979461, "kl": 0.025980454156524502, "context": 0.17621714749839157}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 3.7320446968078613, "loss_scale": 4.0, "retries": 0, "seconds": 179.9, "max_memory_gb": 14.31, "ce": 0.27925523091107607, "kl": 0.06765742966672406, "context": 0.3130839489167556}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.34841760993003845, "loss_scale": 4.0, "retries": 0, "seconds": 180.2, "max_memory_gb": 14.31, "ce": 0.04653303117811447, "kl": 0.029033935850748094, "context": 0.06104999747185502}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.24993358552455902, "loss_scale": 4.0, "retries": 0, "seconds": 180.6, "max_memory_gb": 14.31, "ce": 0.08173105429159477, "kl": 0.12476450385292992, "context": 0.14411331131123006}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.08033321052789688, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 14.31, "ce": 0.008357566024642438, "kl": 0.004047383306897245, "context": 0.010381257627159357}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 1.3451359272003174, "loss_scale": 4.0, "retries": 0, "seconds": 181.3, "max_memory_gb": 14.31, "ce": 0.13312803115695715, "kl": 0.04776860820129514, "context": 0.15701233642175794}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.4744399785995483, "loss_scale": 4.0, "retries": 0, "seconds": 181.7, "max_memory_gb": 14.31, "ce": 0.01045814505778253, "kl": 0.5175983382505365, "context": 0.2692573118256405}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.7848135828971863, "loss_scale": 4.0, "retries": 0, "seconds": 182.0, "max_memory_gb": 14.31, "ce": 0.07544091204181314, "kl": 0.008969805901870131, "context": 0.0799258123151958}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 1.2076830863952637, "loss_scale": 4.0, "retries": 0, "seconds": 182.4, "max_memory_gb": 14.31, "ce": 0.05049202253576368, "kl": 0.035137244733050466, "context": 0.06806064420379698}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 0.785568356513977, "loss_scale": 4.0, "retries": 0, "seconds": 182.8, "max_memory_gb": 14.31, "ce": 0.13897473510587588, "kl": 0.012939803884364665, "context": 0.1454446397256106}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.5026681423187256, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 14.31, "ce": 0.13567900119232945, "kl": 0.029515165762859397, "context": 0.1504365894361399}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 0.7804989814758301, "loss_scale": 4.0, "retries": 0, "seconds": 183.5, "max_memory_gb": 14.31, "ce": 0.08280563267180696, "kl": 0.048395490117400186, "context": 0.10700337929301895}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 0.5729608535766602, "loss_scale": 4.0, "retries": 0, "seconds": 183.8, "max_memory_gb": 14.31, "ce": 0.22615669536753558, "kl": 0.014167892455589026, "context": 0.23324063722975552}
diff --git a/runs/single_llava/removal/L10_s13/adapter/README.md b/runs/single_llava/removal/L10_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/removal/L10_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/removal/L10_s13/adapter/adapter_config.json b/runs/single_llava/removal/L10_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..87f5c88a3478df437f76c61a40a360acda1401e6
--- /dev/null
+++ b/runs/single_llava/removal/L10_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.10.mlp.up_proj",
+    "model.language_model.layers.10.mlp.down_proj",
+    "model.language_model.layers.10.self_attn.v_proj",
+    "model.language_model.layers.10.self_attn.q_proj",
+    "model.language_model.layers.10.self_attn.k_proj",
+    "model.language_model.layers.10.mlp.gate_proj",
+    "model.language_model.layers.10.self_attn.o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/removal/L10_s13/config.json b/runs/single_llava/removal/L10_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7de1d2ac5cebba5353ed4f5023ffa7001a02de10
--- /dev/null
+++ b/runs/single_llava/removal/L10_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "removal",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/removal.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 10,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/removal/L10_s13/steps.jsonl b/runs/single_llava/removal/L10_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2c97cf79e573dcd92cdd5de38ccc18ebb0ea9d48
--- /dev/null
+++ b/runs/single_llava/removal/L10_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.2653537094593048, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 17.03, "ce": 1.6287612388841808, "rank": 0.13243506848812103}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.3255559504032135, "loss_scale": 4.0, "retries": 0, "seconds": 4.4, "max_memory_gb": 17.05, "ce": 2.6751189748756588, "rank": 0.0325009822845459}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.7395737171173096, "loss_scale": 4.0, "retries": 0, "seconds": 6.3, "max_memory_gb": 17.05, "ce": 5.27434634976089, "rank": 0.15134704113006592}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.4953780770301819, "loss_scale": 4.0, "retries": 0, "seconds": 8.1, "max_memory_gb": 17.05, "ce": 3.7998799041379243, "rank": 0.12265133857727051}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.09839017689228058, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 17.05, "ce": 0.13944480172358453, "rank": 0.0}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.26154622435569763, "loss_scale": 4.0, "retries": 0, "seconds": 11.7, "max_memory_gb": 17.05, "ce": 1.5112212565727532, "rank": 0.0}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.537808895111084, "loss_scale": 4.0, "retries": 0, "seconds": 13.5, "max_memory_gb": 17.05, "ce": 3.453142264857888, "rank": 0.019974112510681152}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.6581587195396423, "loss_scale": 4.0, "retries": 0, "seconds": 15.3, "max_memory_gb": 17.05, "ce": 3.693382900208235, "rank": 0.20045030117034912}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.5057637095451355, "loss_scale": 4.0, "retries": 0, "seconds": 17.1, "max_memory_gb": 17.05, "ce": 2.8252401500940323, "rank": 0.040686845779418945}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.32875996828079224, "loss_scale": 4.0, "retries": 0, "seconds": 18.9, "max_memory_gb": 17.05, "ce": 1.612287218682468, "rank": 0.0}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.6271861791610718, "loss_scale": 4.0, "retries": 0, "seconds": 20.8, "max_memory_gb": 17.05, "ce": 2.607816875912249, "rank": 0.2079923003911972}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.43488457798957825, "loss_scale": 4.0, "retries": 0, "seconds": 22.6, "max_memory_gb": 17.05, "ce": 1.4975509156938642, "rank": 0.048140883445739746}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.5896051526069641, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 17.05, "ce": 2.648462751531042, "rank": 0.0}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.9066510200500488, "loss_scale": 4.0, "retries": 0, "seconds": 26.2, "max_memory_gb": 17.05, "ce": 3.834084752947092, "rank": 0.142614483833313}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.6774967312812805, "loss_scale": 4.0, "retries": 0, "seconds": 28.0, "max_memory_gb": 17.05, "ce": 2.567360366578214, "rank": 0.17147791385650635}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.4948481619358063, "loss_scale": 4.0, "retries": 0, "seconds": 29.7, "max_memory_gb": 17.05, "ce": 1.3962891747942194, "rank": 0.0}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.47296541929244995, "loss_scale": 4.0, "retries": 0, "seconds": 31.5, "max_memory_gb": 17.05, "ce": 1.5712507923599333, "rank": 0.03492283821105957}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.0379005670547485, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 17.06, "ce": 3.5272043216973543, "rank": 0.18897509574890137}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.5730359554290771, "loss_scale": 4.0, "retries": 0, "seconds": 35.1, "max_memory_gb": 17.06, "ce": 5.179475121200085, "rank": 0.3090568780899048}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.0360497310757637, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 17.06, "ce": 0.06557388452347368, "rank": 0.10982850939035416}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.9881109595298767, "loss_scale": 4.0, "retries": 0, "seconds": 38.7, "max_memory_gb": 17.06, "ce": 2.6744827412767336, "rank": 0.14484751224517822}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.0270748138427734, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 17.06, "ce": 2.867347010411322, "rank": 0.18620800971984863}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.9474174380302429, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 17.06, "ce": 2.675312611158006, "rank": 0.12389874458312988}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.7487622499465942, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 17.06, "ce": 4.851570934988558, "rank": 0.26078903675079346}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.7818938493728638, "loss_scale": 4.0, "retries": 0, "seconds": 46.1, "max_memory_gb": 17.06, "ce": 3.9649509941227734, "rank": 0.03110659122467041}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.4068167209625244, "loss_scale": 4.0, "retries": 0, "seconds": 47.9, "max_memory_gb": 17.06, "ce": 4.932674236595631, "rank": 0.03971457481384277}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.9962598085403442, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 17.06, "ce": 4.005262214690447, "rank": 0.46315670013427734}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.2854247093200684, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 17.06, "ce": 2.6268937969580293, "rank": 0.0}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.1707673072814941, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 17.06, "ce": 2.5489916619844735, "rank": 0.0}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.6258230209350586, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 17.06, "ce": 2.567375147365965, "rank": 0.0}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.8605738878250122, "loss_scale": 4.0, "retries": 0, "seconds": 57.1, "max_memory_gb": 17.06, "ce": 2.606330675072968, "rank": 0.1318296194076538}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 3.284130334854126, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 17.06, "ce": 4.901888711377978, "rank": 0.20302224159240723}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.4956361055374146, "loss_scale": 4.0, "retries": 0, "seconds": 61.1, "max_memory_gb": 17.06, "ce": 2.3389590131118894, "rank": 0.00721915066242218}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.7855329513549805, "loss_scale": 4.0, "retries": 0, "seconds": 63.1, "max_memory_gb": 17.06, "ce": 2.5038613585056737, "rank": 0.0}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 2.2163290977478027, "loss_scale": 4.0, "retries": 0, "seconds": 65.0, "max_memory_gb": 17.06, "ce": 2.4256789283826947, "rank": 0.12752044200897217}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.6522941589355469, "loss_scale": 4.0, "retries": 0, "seconds": 66.9, "max_memory_gb": 17.06, "ce": 2.365787868387997, "rank": 0.0}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 2.055875062942505, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 17.06, "ce": 2.3350269524380565, "rank": 0.0}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.657377243041992, "loss_scale": 4.0, "retries": 0, "seconds": 70.4, "max_memory_gb": 17.06, "ce": 4.319590548984706, "rank": 0.1314762830734253}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.8457865715026855, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 17.06, "ce": 1.2229944390710443, "rank": 0.10780692100524902}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.3702375888824463, "loss_scale": 4.0, "retries": 0, "seconds": 74.0, "max_memory_gb": 17.06, "ce": 2.3462868672795594, "rank": 0.0}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.1078883409500122, "loss_scale": 4.0, "retries": 0, "seconds": 75.8, "max_memory_gb": 17.06, "ce": 1.3726576752960682, "rank": 0.0}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.8637397289276123, "loss_scale": 4.0, "retries": 0, "seconds": 77.6, "max_memory_gb": 17.06, "ce": 1.9637418282218277, "rank": 0.0}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.805659055709839, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 17.06, "ce": 2.0408092385623604, "rank": 0.0}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.4033801555633545, "loss_scale": 4.0, "retries": 0, "seconds": 81.2, "max_memory_gb": 17.06, "ce": 1.2386307143606246, "rank": 0.0}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.5849415063858032, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 17.06, "ce": 1.1921602706424892, "rank": 0.0}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.730700969696045, "loss_scale": 4.0, "retries": 0, "seconds": 84.7, "max_memory_gb": 17.06, "ce": 1.949591614305973, "rank": 0.0}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 3.1206295490264893, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 17.06, "ce": 1.5866909674368799, "rank": 0.0}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 6.118156433105469, "loss_scale": 4.0, "retries": 0, "seconds": 88.3, "max_memory_gb": 17.06, "ce": 2.861174626275897, "rank": 0.022043466567993164}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 5.640113353729248, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 17.06, "ce": 2.6931647709570825, "rank": 0.0}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 5.561234951019287, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 17.06, "ce": 2.7027317910688, "rank": 0.0}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 2.3553152084350586, "loss_scale": 4.0, "retries": 0, "seconds": 93.8, "max_memory_gb": 17.06, "ce": 1.0628476405981928, "rank": 0.0}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 4.116091251373291, "loss_scale": 4.0, "retries": 0, "seconds": 95.6, "max_memory_gb": 17.06, "ce": 1.6738802595064044, "rank": 0.0}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 6.564763069152832, "loss_scale": 4.0, "retries": 0, "seconds": 97.4, "max_memory_gb": 17.06, "ce": 2.700189939700067, "rank": 0.0}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 7.344996929168701, "loss_scale": 4.0, "retries": 0, "seconds": 99.4, "max_memory_gb": 17.06, "ce": 2.553074872121215, "rank": 0.0}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 4.873615741729736, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 17.06, "ce": 1.5389347891323268, "rank": 0.0}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 10.106751441955566, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 17.06, "ce": 2.9161088475957513, "rank": 0.0}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 2.4290692806243896, "loss_scale": 4.0, "retries": 0, "seconds": 105.2, "max_memory_gb": 17.06, "ce": 0.828968808054924, "rank": 0.010039225220680237}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 5.294338226318359, "loss_scale": 4.0, "retries": 0, "seconds": 107.1, "max_memory_gb": 17.06, "ce": 1.2473032695706934, "rank": 0.0}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.4851014614105225, "loss_scale": 4.0, "retries": 0, "seconds": 108.9, "max_memory_gb": 17.06, "ce": 1.0590149187482893, "rank": 0.0}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 5.569611549377441, "loss_scale": 4.0, "retries": 0, "seconds": 110.7, "max_memory_gb": 17.06, "ce": 1.3420988398138434, "rank": 0.0}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 7.717735290527344, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 17.06, "ce": 1.612314193509519, "rank": 0.0}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 3.23911714553833, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 17.06, "ce": 0.7734976173378527, "rank": 0.0}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 7.831045627593994, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 17.06, "ce": 1.8006023336201906, "rank": 0.03168439865112305}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 7.653633117675781, "loss_scale": 4.0, "retries": 0, "seconds": 118.0, "max_memory_gb": 17.06, "ce": 1.4808640282135457, "rank": 0.0}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 6.547536849975586, "loss_scale": 4.0, "retries": 0, "seconds": 119.8, "max_memory_gb": 17.06, "ce": 1.713633969426155, "rank": 0.0}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.744053602218628, "loss_scale": 4.0, "retries": 0, "seconds": 121.6, "max_memory_gb": 17.06, "ce": 0.6007788652786985, "rank": 0.20530902594327927}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 5.602546691894531, "loss_scale": 4.0, "retries": 0, "seconds": 123.4, "max_memory_gb": 17.06, "ce": 1.1417332235723734, "rank": 0.0}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 2.6935009956359863, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 17.06, "ce": 0.5144030337687582, "rank": 0.0}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 4.137938976287842, "loss_scale": 4.0, "retries": 0, "seconds": 127.1, "max_memory_gb": 17.06, "ce": 0.7994859828613698, "rank": 0.0}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 3.541605234146118, "loss_scale": 4.0, "retries": 0, "seconds": 128.9, "max_memory_gb": 17.06, "ce": 0.6496156393550336, "rank": 0.0}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 5.711341857910156, "loss_scale": 4.0, "retries": 0, "seconds": 130.7, "max_memory_gb": 17.06, "ce": 0.8422634452581406, "rank": 0.0}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 5.83536434173584, "loss_scale": 4.0, "retries": 0, "seconds": 132.6, "max_memory_gb": 17.06, "ce": 1.0015619779005647, "rank": 0.0}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.3484004735946655, "loss_scale": 4.0, "retries": 0, "seconds": 134.3, "max_memory_gb": 17.06, "ce": 0.24749177345074713, "rank": 0.0}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 1.9247084856033325, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 17.06, "ce": 0.5031799962744117, "rank": 0.0}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.6952288150787354, "loss_scale": 4.0, "retries": 0, "seconds": 137.9, "max_memory_gb": 17.06, "ce": 0.830964651890099, "rank": 0.27655136585235596}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 1.8844772577285767, "loss_scale": 4.0, "retries": 0, "seconds": 139.8, "max_memory_gb": 17.06, "ce": 0.5383810237981379, "rank": 0.0}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 2.9208016395568848, "loss_scale": 4.0, "retries": 0, "seconds": 141.6, "max_memory_gb": 17.06, "ce": 0.759094045497477, "rank": 0.0}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 1.6616888046264648, "loss_scale": 4.0, "retries": 0, "seconds": 143.3, "max_memory_gb": 17.06, "ce": 0.45800743578001857, "rank": 0.0}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 0.9012949466705322, "loss_scale": 4.0, "retries": 0, "seconds": 145.1, "max_memory_gb": 17.06, "ce": 0.30008298996835947, "rank": 0.0036847442388534546}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 1.277734637260437, "loss_scale": 4.0, "retries": 0, "seconds": 146.9, "max_memory_gb": 17.06, "ce": 0.37472984846681356, "rank": 0.0}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 1.4055503606796265, "loss_scale": 4.0, "retries": 0, "seconds": 148.7, "max_memory_gb": 17.06, "ce": 0.6223121527582407, "rank": 0.0}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 0.8037723302841187, "loss_scale": 4.0, "retries": 0, "seconds": 150.5, "max_memory_gb": 17.06, "ce": 0.50034453952685, "rank": 0.0}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 0.8398341536521912, "loss_scale": 4.0, "retries": 0, "seconds": 152.3, "max_memory_gb": 17.06, "ce": 0.5743882453534752, "rank": 0.0}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 0.6813604235649109, "loss_scale": 4.0, "retries": 0, "seconds": 154.0, "max_memory_gb": 17.06, "ce": 0.3013541620457545, "rank": 0.0}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.41978543996810913, "loss_scale": 4.0, "retries": 0, "seconds": 155.9, "max_memory_gb": 17.06, "ce": 0.12983459560200572, "rank": 0.0}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 1.0865713357925415, "loss_scale": 4.0, "retries": 0, "seconds": 157.7, "max_memory_gb": 17.06, "ce": 0.5939884511753917, "rank": 0.0}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 0.7596805095672607, "loss_scale": 4.0, "retries": 0, "seconds": 159.5, "max_memory_gb": 17.06, "ce": 0.4927105763927102, "rank": 0.0}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 0.5810946226119995, "loss_scale": 4.0, "retries": 0, "seconds": 161.3, "max_memory_gb": 17.06, "ce": 0.45623165159486234, "rank": 0.0}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.3421095609664917, "loss_scale": 4.0, "retries": 0, "seconds": 163.1, "max_memory_gb": 17.06, "ce": 0.17280211357865483, "rank": 0.1241958737373352}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 0.4941292405128479, "loss_scale": 4.0, "retries": 0, "seconds": 164.9, "max_memory_gb": 17.06, "ce": 0.39960496174171567, "rank": 0.0}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.4931747317314148, "loss_scale": 4.0, "retries": 0, "seconds": 166.7, "max_memory_gb": 17.06, "ce": 0.360349303111434, "rank": 0.0}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 0.36913952231407166, "loss_scale": 4.0, "retries": 0, "seconds": 168.5, "max_memory_gb": 17.06, "ce": 0.29016237088944763, "rank": 0.0}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 0.4943414628505707, "loss_scale": 4.0, "retries": 0, "seconds": 170.3, "max_memory_gb": 17.06, "ce": 0.343200626084581, "rank": 0.0}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 0.540108859539032, "loss_scale": 4.0, "retries": 0, "seconds": 172.2, "max_memory_gb": 17.06, "ce": 0.5166195062920451, "rank": 0.002531766891479492}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 0.6709714531898499, "loss_scale": 4.0, "retries": 0, "seconds": 174.0, "max_memory_gb": 17.06, "ce": 0.48902175994589925, "rank": 0.0}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 0.7049700617790222, "loss_scale": 4.0, "retries": 0, "seconds": 175.8, "max_memory_gb": 17.06, "ce": 0.5343865472823381, "rank": 0.062232330441474915}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 0.6360589265823364, "loss_scale": 4.0, "retries": 0, "seconds": 177.6, "max_memory_gb": 17.06, "ce": 0.31412726896815, "rank": 0.016329094767570496}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 0.35962173342704773, "loss_scale": 4.0, "retries": 0, "seconds": 179.4, "max_memory_gb": 17.06, "ce": 0.309364530723542, "rank": 0.0}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.7133542895317078, "loss_scale": 4.0, "retries": 0, "seconds": 181.2, "max_memory_gb": 17.06, "ce": 0.22086609352845699, "rank": 0.024518698453903198}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 0.5450439453125, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 17.06, "ce": 0.5255201244726777, "rank": 0.0}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 0.5074237585067749, "loss_scale": 4.0, "retries": 0, "seconds": 185.0, "max_memory_gb": 17.06, "ce": 0.4749835589900613, "rank": 0.0}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 0.9613125324249268, "loss_scale": 4.0, "retries": 0, "seconds": 186.9, "max_memory_gb": 17.06, "ce": 0.6470595630817115, "rank": 0.0}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 0.35918304324150085, "loss_scale": 4.0, "retries": 0, "seconds": 188.7, "max_memory_gb": 17.06, "ce": 0.3163674407405779, "rank": 0.1175977811217308}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 0.2453080713748932, "loss_scale": 4.0, "retries": 0, "seconds": 190.4, "max_memory_gb": 17.06, "ce": 0.21573205257300287, "rank": 0.08932866156101227}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 0.7522282600402832, "loss_scale": 4.0, "retries": 0, "seconds": 192.2, "max_memory_gb": 17.06, "ce": 0.4574836092069745, "rank": 0.05411801487207413}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 0.4991978406906128, "loss_scale": 4.0, "retries": 0, "seconds": 194.0, "max_memory_gb": 17.06, "ce": 0.479890544898808, "rank": 0.09016163647174835}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 0.3230806887149811, "loss_scale": 4.0, "retries": 0, "seconds": 195.8, "max_memory_gb": 17.06, "ce": 0.27099999971687794, "rank": 0.0}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 0.37596431374549866, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 17.06, "ce": 0.27759790723212063, "rank": 0.0}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 0.6493387222290039, "loss_scale": 4.0, "retries": 0, "seconds": 199.4, "max_memory_gb": 17.06, "ce": 0.6013336610049009, "rank": 0.0}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 0.5871270895004272, "loss_scale": 4.0, "retries": 0, "seconds": 201.2, "max_memory_gb": 17.06, "ce": 0.39380110474303365, "rank": 0.05510471761226654}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 0.4552331566810608, "loss_scale": 4.0, "retries": 0, "seconds": 203.0, "max_memory_gb": 17.06, "ce": 0.3310268234927207, "rank": 0.00012758374214172363}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 0.4337981045246124, "loss_scale": 4.0, "retries": 0, "seconds": 204.8, "max_memory_gb": 17.06, "ce": 0.3309075792785734, "rank": 0.0}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 0.36006665229797363, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 17.06, "ce": 0.3795342445373535, "rank": 0.0}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 0.7231184244155884, "loss_scale": 4.0, "retries": 0, "seconds": 208.5, "max_memory_gb": 17.06, "ce": 0.6707635335624218, "rank": 0.1327698975801468}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 0.4730296730995178, "loss_scale": 4.0, "retries": 0, "seconds": 210.3, "max_memory_gb": 17.06, "ce": 0.4467210005968809, "rank": 0.1201343908905983}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.23845316469669342, "loss_scale": 4.0, "retries": 0, "seconds": 212.0, "max_memory_gb": 17.06, "ce": 0.14859285682905465, "rank": 0.0}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 0.8263993263244629, "loss_scale": 4.0, "retries": 0, "seconds": 213.8, "max_memory_gb": 17.06, "ce": 0.4337487407028675, "rank": 0.0}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 0.277067095041275, "loss_scale": 4.0, "retries": 0, "seconds": 215.6, "max_memory_gb": 17.06, "ce": 0.2947067031636834, "rank": 0.0}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 0.3403720557689667, "loss_scale": 4.0, "retries": 0, "seconds": 217.4, "max_memory_gb": 17.06, "ce": 0.26700756838545203, "rank": 0.0}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 0.7506887316703796, "loss_scale": 4.0, "retries": 0, "seconds": 219.3, "max_memory_gb": 17.06, "ce": 0.4234173677396029, "rank": 0.0}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 0.5798329710960388, "loss_scale": 4.0, "retries": 0, "seconds": 221.1, "max_memory_gb": 17.06, "ce": 0.33532921213191, "rank": 0.0}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 0.3123505711555481, "loss_scale": 4.0, "retries": 0, "seconds": 222.9, "max_memory_gb": 17.06, "ce": 0.3668722028378397, "rank": 0.0}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 0.810811460018158, "loss_scale": 4.0, "retries": 0, "seconds": 224.6, "max_memory_gb": 17.06, "ce": 0.5688952417112887, "rank": 0.0}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 0.7536030411720276, "loss_scale": 4.0, "retries": 0, "seconds": 226.3, "max_memory_gb": 17.06, "ce": 0.39749180153012276, "rank": 0.0452866405248642}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 0.3285997807979584, "loss_scale": 4.0, "retries": 0, "seconds": 228.0, "max_memory_gb": 17.06, "ce": 0.29196617053821683, "rank": 0.0}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 0.5237862467765808, "loss_scale": 4.0, "retries": 0, "seconds": 229.8, "max_memory_gb": 17.06, "ce": 0.5323069337755442, "rank": 0.0}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 0.2156722992658615, "loss_scale": 4.0, "retries": 0, "seconds": 231.6, "max_memory_gb": 17.06, "ce": 0.17327834200114012, "rank": 0.0}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 0.48926183581352234, "loss_scale": 4.0, "retries": 0, "seconds": 233.5, "max_memory_gb": 17.06, "ce": 0.37722240679431707, "rank": 0.0}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 0.44464948773384094, "loss_scale": 4.0, "retries": 0, "seconds": 235.2, "max_memory_gb": 17.06, "ce": 0.5191492298617959, "rank": 0.0}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 0.8335224986076355, "loss_scale": 4.0, "retries": 0, "seconds": 237.0, "max_memory_gb": 17.06, "ce": 0.5335880378261209, "rank": 0.0}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 0.9830713868141174, "loss_scale": 4.0, "retries": 0, "seconds": 238.8, "max_memory_gb": 17.06, "ce": 0.36861937306821346, "rank": 0.0}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 0.968222439289093, "loss_scale": 4.0, "retries": 0, "seconds": 240.6, "max_memory_gb": 17.06, "ce": 0.6053362132515758, "rank": 0.0719042718410492}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 1.0353920459747314, "loss_scale": 4.0, "retries": 0, "seconds": 242.4, "max_memory_gb": 17.06, "ce": 0.5148463854566216, "rank": 0.041343286633491516}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.5655443072319031, "loss_scale": 4.0, "retries": 0, "seconds": 244.2, "max_memory_gb": 17.06, "ce": 0.19463899871334434, "rank": 0.0}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 0.5050235390663147, "loss_scale": 4.0, "retries": 0, "seconds": 246.0, "max_memory_gb": 17.06, "ce": 0.4411933533847332, "rank": 0.031171157956123352}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 0.6178869605064392, "loss_scale": 4.0, "retries": 0, "seconds": 247.8, "max_memory_gb": 17.06, "ce": 0.4342943197116256, "rank": 0.0}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 0.5767912864685059, "loss_scale": 4.0, "retries": 0, "seconds": 249.6, "max_memory_gb": 17.06, "ce": 0.3575440489221364, "rank": 0.0}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 0.3989766836166382, "loss_scale": 4.0, "retries": 0, "seconds": 251.5, "max_memory_gb": 17.06, "ce": 0.36565400566905737, "rank": 0.0}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 1.069201946258545, "loss_scale": 4.0, "retries": 0, "seconds": 253.5, "max_memory_gb": 17.06, "ce": 0.5772446077316999, "rank": 0.0}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 0.4011526107788086, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 17.06, "ce": 0.2657060065539554, "rank": 0.0}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 0.5480960607528687, "loss_scale": 4.0, "retries": 0, "seconds": 257.3, "max_memory_gb": 17.06, "ce": 0.38858985202386975, "rank": 0.0}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 0.3574036955833435, "loss_scale": 4.0, "retries": 0, "seconds": 259.1, "max_memory_gb": 17.06, "ce": 0.3530458190944046, "rank": 0.0}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.314845472574234, "loss_scale": 4.0, "retries": 0, "seconds": 261.0, "max_memory_gb": 17.06, "ce": 0.17333721253089607, "rank": 0.0}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 0.7380245327949524, "loss_scale": 4.0, "retries": 0, "seconds": 262.8, "max_memory_gb": 17.06, "ce": 0.3677304429002106, "rank": 0.0}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 0.2884214222431183, "loss_scale": 4.0, "retries": 0, "seconds": 264.6, "max_memory_gb": 17.06, "ce": 0.23225845757406205, "rank": 0.0}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 0.4080408811569214, "loss_scale": 4.0, "retries": 0, "seconds": 266.4, "max_memory_gb": 17.06, "ce": 0.3190428973175585, "rank": 0.0}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 0.6203790307044983, "loss_scale": 4.0, "retries": 0, "seconds": 268.2, "max_memory_gb": 17.06, "ce": 0.5144716261420399, "rank": 0.0}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 0.22727632522583008, "loss_scale": 4.0, "retries": 0, "seconds": 270.2, "max_memory_gb": 17.06, "ce": 0.157042468781583, "rank": 0.0}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 0.4264734089374542, "loss_scale": 4.0, "retries": 0, "seconds": 271.9, "max_memory_gb": 17.06, "ce": 0.22714193316642195, "rank": 0.13004708290100098}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 0.7655853033065796, "loss_scale": 4.0, "retries": 0, "seconds": 273.8, "max_memory_gb": 17.06, "ce": 0.47510278690606356, "rank": 0.0029802322387695312}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 0.5654059648513794, "loss_scale": 4.0, "retries": 0, "seconds": 275.6, "max_memory_gb": 17.06, "ce": 0.24621729145292193, "rank": 0.0}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 0.6040381193161011, "loss_scale": 4.0, "retries": 0, "seconds": 277.4, "max_memory_gb": 17.06, "ce": 0.4377418155781925, "rank": 0.0}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 0.49127042293548584, "loss_scale": 4.0, "retries": 0, "seconds": 279.2, "max_memory_gb": 17.06, "ce": 0.33668928185943514, "rank": 0.0}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 1.111899733543396, "loss_scale": 4.0, "retries": 0, "seconds": 281.0, "max_memory_gb": 17.06, "ce": 0.633968616835773, "rank": 0.11069561541080475}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 1.1119927167892456, "loss_scale": 4.0, "retries": 0, "seconds": 283.0, "max_memory_gb": 17.06, "ce": 0.5830888766795397, "rank": 0.0}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 0.4022523760795593, "loss_scale": 4.0, "retries": 0, "seconds": 284.9, "max_memory_gb": 17.06, "ce": 0.2567069035721943, "rank": 0.0}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 0.4134531021118164, "loss_scale": 4.0, "retries": 0, "seconds": 286.8, "max_memory_gb": 17.06, "ce": 0.2737285391194746, "rank": 0.0}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 0.6503630876541138, "loss_scale": 4.0, "retries": 0, "seconds": 288.7, "max_memory_gb": 17.06, "ce": 0.4192064991220832, "rank": 0.0}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 0.49696820974349976, "loss_scale": 4.0, "retries": 0, "seconds": 290.6, "max_memory_gb": 17.07, "ce": 0.27906945592258126, "rank": 0.0}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 0.591069757938385, "loss_scale": 4.0, "retries": 0, "seconds": 292.5, "max_memory_gb": 17.07, "ce": 0.35883168689906597, "rank": 0.0}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 0.702985942363739, "loss_scale": 4.0, "retries": 0, "seconds": 294.3, "max_memory_gb": 17.07, "ce": 0.34690025926101953, "rank": 0.0}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 0.36513763666152954, "loss_scale": 4.0, "retries": 0, "seconds": 296.0, "max_memory_gb": 17.07, "ce": 0.2183677013963461, "rank": 0.0}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 0.6617574691772461, "loss_scale": 4.0, "retries": 0, "seconds": 297.8, "max_memory_gb": 17.07, "ce": 0.3491540604736656, "rank": 0.0}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 0.728105366230011, "loss_scale": 4.0, "retries": 0, "seconds": 299.6, "max_memory_gb": 17.07, "ce": 0.3988931989297271, "rank": 0.030606761574745178}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 0.43257254362106323, "loss_scale": 4.0, "retries": 0, "seconds": 301.4, "max_memory_gb": 17.07, "ce": 0.21884999191388488, "rank": 0.0}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 0.694508969783783, "loss_scale": 4.0, "retries": 0, "seconds": 303.3, "max_memory_gb": 17.07, "ce": 0.3328569671139121, "rank": 0.0}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 0.8062127232551575, "loss_scale": 4.0, "retries": 0, "seconds": 305.1, "max_memory_gb": 17.07, "ce": 0.413093619979918, "rank": 0.0}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 0.419758141040802, "loss_scale": 4.0, "retries": 0, "seconds": 306.9, "max_memory_gb": 17.07, "ce": 0.2459282297641039, "rank": 0.0}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 1.2298731803894043, "loss_scale": 4.0, "retries": 0, "seconds": 308.7, "max_memory_gb": 17.07, "ce": 0.6875802236609161, "rank": 0.0}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 0.6318218111991882, "loss_scale": 4.0, "retries": 0, "seconds": 310.5, "max_memory_gb": 17.07, "ce": 0.26673670625314116, "rank": 0.0}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 0.7310258746147156, "loss_scale": 4.0, "retries": 0, "seconds": 312.3, "max_memory_gb": 17.07, "ce": 0.2753348869155161, "rank": 0.0}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 0.8699655532836914, "loss_scale": 4.0, "retries": 0, "seconds": 314.1, "max_memory_gb": 17.07, "ce": 0.40184302371926606, "rank": 0.0}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 0.8709095120429993, "loss_scale": 4.0, "retries": 0, "seconds": 315.9, "max_memory_gb": 17.07, "ce": 0.3568288432434201, "rank": 0.0}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 0.7714877128601074, "loss_scale": 4.0, "retries": 0, "seconds": 317.7, "max_memory_gb": 17.07, "ce": 0.2984608580591157, "rank": 0.0}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 0.6785001754760742, "loss_scale": 4.0, "retries": 0, "seconds": 319.4, "max_memory_gb": 17.07, "ce": 0.3577356254681945, "rank": 0.0}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 0.9343921542167664, "loss_scale": 4.0, "retries": 0, "seconds": 321.3, "max_memory_gb": 17.07, "ce": 0.3081168755888939, "rank": 0.0}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 0.9734153151512146, "loss_scale": 4.0, "retries": 0, "seconds": 323.1, "max_memory_gb": 17.07, "ce": 0.3668884492944926, "rank": 0.12646567821502686}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 0.5292765498161316, "loss_scale": 4.0, "retries": 0, "seconds": 324.9, "max_memory_gb": 17.07, "ce": 0.23341851984150708, "rank": 0.0}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 0.5627446174621582, "loss_scale": 4.0, "retries": 0, "seconds": 326.6, "max_memory_gb": 17.07, "ce": 0.4274149453267455, "rank": 0.0}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 1.3119733333587646, "loss_scale": 4.0, "retries": 0, "seconds": 328.4, "max_memory_gb": 17.07, "ce": 0.5287709203548729, "rank": 0.0}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 1.202962040901184, "loss_scale": 4.0, "retries": 0, "seconds": 330.3, "max_memory_gb": 17.07, "ce": 0.32800540467724204, "rank": 0.0}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.5466355085372925, "loss_scale": 4.0, "retries": 0, "seconds": 332.1, "max_memory_gb": 17.07, "ce": 0.3039122853660956, "rank": 0.0}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 0.5418186187744141, "loss_scale": 4.0, "retries": 0, "seconds": 333.9, "max_memory_gb": 17.07, "ce": 0.14617282332619652, "rank": 0.0}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 0.7454182505607605, "loss_scale": 4.0, "retries": 0, "seconds": 335.7, "max_memory_gb": 17.07, "ce": 0.21988520602462813, "rank": 0.09672726690769196}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 0.6906060576438904, "loss_scale": 4.0, "retries": 0, "seconds": 337.5, "max_memory_gb": 17.07, "ce": 0.22249354468658566, "rank": 0.0}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 0.6232712268829346, "loss_scale": 4.0, "retries": 0, "seconds": 339.4, "max_memory_gb": 17.07, "ce": 0.15910027222707868, "rank": 0.0}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 0.9907282590866089, "loss_scale": 4.0, "retries": 0, "seconds": 341.2, "max_memory_gb": 17.07, "ce": 0.3109053149819374, "rank": 0.13291069865226746}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.28445953130722046, "loss_scale": 4.0, "retries": 0, "seconds": 343.0, "max_memory_gb": 17.07, "ce": 0.1431518459576182, "rank": 0.0}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 1.0227481126785278, "loss_scale": 4.0, "retries": 0, "seconds": 344.8, "max_memory_gb": 17.07, "ce": 0.17614029441028833, "rank": 0.0}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 1.1541433334350586, "loss_scale": 4.0, "retries": 0, "seconds": 346.6, "max_memory_gb": 17.07, "ce": 0.25326236709952354, "rank": 0.0}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 1.0163580179214478, "loss_scale": 4.0, "retries": 0, "seconds": 348.3, "max_memory_gb": 17.07, "ce": 0.2965575170237571, "rank": 0.0}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 0.5788617134094238, "loss_scale": 4.0, "retries": 0, "seconds": 350.1, "max_memory_gb": 17.07, "ce": 0.28357991483062506, "rank": 0.0}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 0.7300437688827515, "loss_scale": 4.0, "retries": 0, "seconds": 352.0, "max_memory_gb": 17.07, "ce": 0.20343787432648242, "rank": 0.12524455785751343}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 1.901782512664795, "loss_scale": 4.0, "retries": 0, "seconds": 353.8, "max_memory_gb": 17.07, "ce": 0.35861006332561374, "rank": 0.0}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 0.8357698917388916, "loss_scale": 4.0, "retries": 0, "seconds": 355.6, "max_memory_gb": 17.07, "ce": 0.2675772560760379, "rank": 0.0}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 0.6202563047409058, "loss_scale": 4.0, "retries": 0, "seconds": 357.5, "max_memory_gb": 17.07, "ce": 0.16595241497270763, "rank": 0.0}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 0.6197959780693054, "loss_scale": 4.0, "retries": 0, "seconds": 359.3, "max_memory_gb": 17.07, "ce": 0.17292138887569308, "rank": 0.09890955686569214}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 0.6396095752716064, "loss_scale": 4.0, "retries": 0, "seconds": 361.1, "max_memory_gb": 17.07, "ce": 0.21246481966227293, "rank": 0.0}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 1.218388557434082, "loss_scale": 4.0, "retries": 0, "seconds": 362.9, "max_memory_gb": 17.07, "ce": 0.23103967809583992, "rank": 0.0}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 0.9482417106628418, "loss_scale": 4.0, "retries": 0, "seconds": 364.7, "max_memory_gb": 17.07, "ce": 0.27124008047394454, "rank": 0.0}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 0.6094931960105896, "loss_scale": 4.0, "retries": 0, "seconds": 366.5, "max_memory_gb": 17.07, "ce": 0.1408142144791782, "rank": 0.0}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 0.2668677568435669, "loss_scale": 4.0, "retries": 0, "seconds": 368.2, "max_memory_gb": 17.07, "ce": 0.07008761027827859, "rank": 0.0}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 1.1558542251586914, "loss_scale": 4.0, "retries": 0, "seconds": 370.0, "max_memory_gb": 17.07, "ce": 0.2276349274907261, "rank": 0.0}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 1.197653889656067, "loss_scale": 4.0, "retries": 0, "seconds": 371.8, "max_memory_gb": 17.07, "ce": 0.256359769962728, "rank": 0.0}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 0.8773912787437439, "loss_scale": 4.0, "retries": 0, "seconds": 373.6, "max_memory_gb": 17.07, "ce": 0.21622078301152214, "rank": 0.0}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 0.837119996547699, "loss_scale": 4.0, "retries": 0, "seconds": 375.4, "max_memory_gb": 17.07, "ce": 0.1657766808057204, "rank": 0.0}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.8075042963027954, "loss_scale": 4.0, "retries": 0, "seconds": 377.2, "max_memory_gb": 17.07, "ce": 0.500276559847407, "rank": 0.23895415663719177}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 0.6146867275238037, "loss_scale": 4.0, "retries": 0, "seconds": 379.0, "max_memory_gb": 17.07, "ce": 0.10501860844669864, "rank": 0.0}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 0.7108584642410278, "loss_scale": 4.0, "retries": 0, "seconds": 380.9, "max_memory_gb": 17.07, "ce": 0.21449734084308147, "rank": 0.12464970350265503}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 1.2613439559936523, "loss_scale": 4.0, "retries": 0, "seconds": 382.8, "max_memory_gb": 17.07, "ce": 0.29874704021494836, "rank": 0.0}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 0.7386890053749084, "loss_scale": 4.0, "retries": 0, "seconds": 384.6, "max_memory_gb": 17.07, "ce": 0.24830687209032476, "rank": 0.0}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 1.6110005378723145, "loss_scale": 4.0, "retries": 0, "seconds": 386.4, "max_memory_gb": 17.07, "ce": 0.3558224947191775, "rank": 0.0}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 0.904782235622406, "loss_scale": 4.0, "retries": 0, "seconds": 388.2, "max_memory_gb": 17.07, "ce": 0.32647723238915205, "rank": 0.0}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 0.9413356184959412, "loss_scale": 4.0, "retries": 0, "seconds": 390.0, "max_memory_gb": 17.07, "ce": 0.3098914442816749, "rank": 0.0}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 2.467745065689087, "loss_scale": 4.0, "retries": 0, "seconds": 391.9, "max_memory_gb": 17.07, "ce": 0.46963578544091433, "rank": 0.0}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.1583045721054077, "loss_scale": 4.0, "retries": 0, "seconds": 393.7, "max_memory_gb": 17.07, "ce": 0.1941739316098392, "rank": 0.0}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 0.553604006767273, "loss_scale": 4.0, "retries": 0, "seconds": 395.5, "max_memory_gb": 17.07, "ce": 0.08960537100210786, "rank": 0.0}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 1.8229995965957642, "loss_scale": 4.0, "retries": 0, "seconds": 397.3, "max_memory_gb": 17.07, "ce": 0.32240480254404247, "rank": 0.0}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 0.6764983534812927, "loss_scale": 4.0, "retries": 0, "seconds": 399.2, "max_memory_gb": 17.07, "ce": 0.12740231992211193, "rank": 0.0}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 0.9909998178482056, "loss_scale": 4.0, "retries": 0, "seconds": 401.2, "max_memory_gb": 17.07, "ce": 0.1443963189376518, "rank": 0.0}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 1.207172155380249, "loss_scale": 4.0, "retries": 0, "seconds": 403.1, "max_memory_gb": 17.07, "ce": 0.17232810659334064, "rank": 0.0}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.8533828258514404, "loss_scale": 4.0, "retries": 0, "seconds": 405.0, "max_memory_gb": 17.07, "ce": 0.38471254077740014, "rank": 0.0}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 0.7868759632110596, "loss_scale": 4.0, "retries": 0, "seconds": 407.0, "max_memory_gb": 17.07, "ce": 0.14611067704390734, "rank": 0.0}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 0.6796360015869141, "loss_scale": 4.0, "retries": 0, "seconds": 408.9, "max_memory_gb": 17.07, "ce": 0.10894985613413155, "rank": 0.09697539359331131}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 1.0888655185699463, "loss_scale": 4.0, "retries": 0, "seconds": 410.9, "max_memory_gb": 17.07, "ce": 0.08884500770363957, "rank": 0.0}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 0.6971631646156311, "loss_scale": 4.0, "retries": 0, "seconds": 412.7, "max_memory_gb": 17.07, "ce": 0.10466231952887028, "rank": 0.0}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 3.7534406185150146, "loss_scale": 4.0, "retries": 0, "seconds": 414.5, "max_memory_gb": 17.07, "ce": 0.10777763871010393, "rank": 0.0}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 0.728191077709198, "loss_scale": 4.0, "retries": 0, "seconds": 416.3, "max_memory_gb": 17.07, "ce": 0.07428923319093883, "rank": 0.0}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 0.999570369720459, "loss_scale": 4.0, "retries": 0, "seconds": 418.1, "max_memory_gb": 17.07, "ce": 0.1874048626050353, "rank": 0.0}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 0.8016409277915955, "loss_scale": 4.0, "retries": 0, "seconds": 419.9, "max_memory_gb": 17.07, "ce": 0.23999245290178806, "rank": 0.0}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 0.8695796728134155, "loss_scale": 4.0, "retries": 0, "seconds": 421.7, "max_memory_gb": 17.07, "ce": 0.2163337438250892, "rank": 0.0}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 1.7546100616455078, "loss_scale": 4.0, "retries": 0, "seconds": 423.5, "max_memory_gb": 17.07, "ce": 0.12544035795144737, "rank": 0.0}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 1.6965603828430176, "loss_scale": 4.0, "retries": 0, "seconds": 425.3, "max_memory_gb": 17.07, "ce": 0.191050790483132, "rank": 0.0}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 1.029819130897522, "loss_scale": 4.0, "retries": 0, "seconds": 427.2, "max_memory_gb": 17.07, "ce": 0.18246077001094818, "rank": 0.0}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 1.1684067249298096, "loss_scale": 4.0, "retries": 0, "seconds": 429.0, "max_memory_gb": 17.07, "ce": 0.31298447493463755, "rank": 0.0}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 0.5531538128852844, "loss_scale": 4.0, "retries": 0, "seconds": 430.8, "max_memory_gb": 17.07, "ce": 0.11267598974518478, "rank": 0.0}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 1.7671211957931519, "loss_scale": 4.0, "retries": 0, "seconds": 432.7, "max_memory_gb": 17.07, "ce": 0.21600073820445687, "rank": 0.0}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 0.9538184404373169, "loss_scale": 4.0, "retries": 0, "seconds": 434.5, "max_memory_gb": 17.07, "ce": 0.09468177054077387, "rank": 0.0}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 0.8235322833061218, "loss_scale": 4.0, "retries": 0, "seconds": 436.3, "max_memory_gb": 17.07, "ce": 0.14585359685588628, "rank": 0.0}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 1.2784875631332397, "loss_scale": 4.0, "retries": 0, "seconds": 438.1, "max_memory_gb": 17.07, "ce": 0.10288199700880796, "rank": 0.0}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 0.635307788848877, "loss_scale": 4.0, "retries": 0, "seconds": 440.0, "max_memory_gb": 17.07, "ce": 0.11335956468246877, "rank": 0.0}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.7831205129623413, "loss_scale": 4.0, "retries": 0, "seconds": 441.8, "max_memory_gb": 17.07, "ce": 0.11499843385536224, "rank": 0.0}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 3.9932093620300293, "loss_scale": 4.0, "retries": 0, "seconds": 443.6, "max_memory_gb": 17.07, "ce": 0.1941828140988946, "rank": 0.0}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 2.5498039722442627, "loss_scale": 4.0, "retries": 0, "seconds": 445.4, "max_memory_gb": 17.07, "ce": 0.36075691028963774, "rank": 0.005601376295089722}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 1.0944222211837769, "loss_scale": 4.0, "retries": 0, "seconds": 447.2, "max_memory_gb": 17.07, "ce": 0.33041958790272474, "rank": 0.0}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 0.33588993549346924, "loss_scale": 4.0, "retries": 0, "seconds": 449.0, "max_memory_gb": 17.07, "ce": 0.03202961460920051, "rank": 0.0}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 0.9235217571258545, "loss_scale": 4.0, "retries": 0, "seconds": 450.9, "max_memory_gb": 17.07, "ce": 0.13512313779210672, "rank": 0.0}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 0.8263058662414551, "loss_scale": 4.0, "retries": 0, "seconds": 452.7, "max_memory_gb": 17.07, "ce": 0.09953069721814245, "rank": 0.0}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 1.5477266311645508, "loss_scale": 4.0, "retries": 0, "seconds": 454.5, "max_memory_gb": 17.07, "ce": 0.09603034274186939, "rank": 0.0}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 0.8833674788475037, "loss_scale": 4.0, "retries": 0, "seconds": 456.3, "max_memory_gb": 17.07, "ce": 0.07726652413839474, "rank": 0.0}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 0.5663745403289795, "loss_scale": 4.0, "retries": 0, "seconds": 458.1, "max_memory_gb": 17.07, "ce": 0.10614828870166093, "rank": 0.0}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 0.6445903778076172, "loss_scale": 4.0, "retries": 0, "seconds": 459.9, "max_memory_gb": 17.07, "ce": 0.042012745747342706, "rank": 0.0}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 0.9675799012184143, "loss_scale": 4.0, "retries": 0, "seconds": 461.7, "max_memory_gb": 17.07, "ce": 0.13749151150113903, "rank": 0.0}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 13.477167129516602, "loss_scale": 4.0, "retries": 0, "seconds": 463.6, "max_memory_gb": 17.07, "ce": 0.27132691629230976, "rank": 0.0}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 0.6978095769882202, "loss_scale": 4.0, "retries": 0, "seconds": 465.3, "max_memory_gb": 17.07, "ce": 0.1642457713605836, "rank": 0.0}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 0.5371504426002502, "loss_scale": 4.0, "retries": 0, "seconds": 467.1, "max_memory_gb": 17.07, "ce": 0.07037992484401911, "rank": 0.0}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 1.038596510887146, "loss_scale": 4.0, "retries": 0, "seconds": 468.9, "max_memory_gb": 17.07, "ce": 0.05640987880178727, "rank": 0.0}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 1.3217719793319702, "loss_scale": 4.0, "retries": 0, "seconds": 470.7, "max_memory_gb": 17.07, "ce": 0.09912410360993817, "rank": 0.0}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 0.3922654986381531, "loss_scale": 4.0, "retries": 0, "seconds": 472.5, "max_memory_gb": 17.07, "ce": 0.02845927531598136, "rank": 0.0}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 1.4413464069366455, "loss_scale": 4.0, "retries": 0, "seconds": 474.3, "max_memory_gb": 17.07, "ce": 0.09429941413691267, "rank": 0.0}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 0.76523756980896, "loss_scale": 4.0, "retries": 0, "seconds": 476.1, "max_memory_gb": 17.07, "ce": 0.09819268248975277, "rank": 0.0}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 1.0661110877990723, "loss_scale": 4.0, "retries": 0, "seconds": 477.9, "max_memory_gb": 17.07, "ce": 0.16669618373271078, "rank": 0.0}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 1.2564232349395752, "loss_scale": 4.0, "retries": 0, "seconds": 479.7, "max_memory_gb": 17.07, "ce": 0.1702478100778535, "rank": 0.1250704973936081}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 2.61979603767395, "loss_scale": 4.0, "retries": 0, "seconds": 481.5, "max_memory_gb": 17.07, "ce": 0.2474226992635522, "rank": 0.0}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 1.2290652990341187, "loss_scale": 4.0, "retries": 0, "seconds": 483.3, "max_memory_gb": 17.07, "ce": 0.07601014006650075, "rank": 0.0}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 26.397762298583984, "loss_scale": 4.0, "retries": 0, "seconds": 485.1, "max_memory_gb": 17.07, "ce": 0.5885411353083327, "rank": 0.0}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 0.9796267151832581, "loss_scale": 4.0, "retries": 0, "seconds": 487.0, "max_memory_gb": 17.07, "ce": 0.1312100634095259, "rank": 0.0}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 1.0585566759109497, "loss_scale": 4.0, "retries": 0, "seconds": 488.8, "max_memory_gb": 17.07, "ce": 0.1944329718244262, "rank": 0.0}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.675136089324951, "loss_scale": 4.0, "retries": 0, "seconds": 490.6, "max_memory_gb": 17.07, "ce": 0.5364191876142286, "rank": 0.0}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 2.2616336345672607, "loss_scale": 4.0, "retries": 0, "seconds": 492.4, "max_memory_gb": 17.07, "ce": 0.18166883720550686, "rank": 0.0}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 1.115338921546936, "loss_scale": 4.0, "retries": 0, "seconds": 494.3, "max_memory_gb": 17.07, "ce": 0.04332581668859348, "rank": 0.11697549372911453}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 1.693501591682434, "loss_scale": 4.0, "retries": 0, "seconds": 496.1, "max_memory_gb": 17.07, "ce": 0.10867558629252017, "rank": 0.0}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 0.8389749526977539, "loss_scale": 4.0, "retries": 0, "seconds": 497.9, "max_memory_gb": 17.07, "ce": 0.04915525941760279, "rank": 0.12502126395702362}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 0.6943002939224243, "loss_scale": 4.0, "retries": 0, "seconds": 499.7, "max_memory_gb": 17.07, "ce": 0.0926717147231102, "rank": 0.0}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 54.95853042602539, "loss_scale": 4.0, "retries": 0, "seconds": 501.7, "max_memory_gb": 17.07, "ce": 1.0695279681240208, "rank": 0.0}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 1.5378692150115967, "loss_scale": 4.0, "retries": 0, "seconds": 503.5, "max_memory_gb": 17.07, "ce": 0.03505704148847144, "rank": 0.0}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 0.742716908454895, "loss_scale": 4.0, "retries": 0, "seconds": 505.4, "max_memory_gb": 17.07, "ce": 0.04046284357900731, "rank": 0.0}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 2.872694969177246, "loss_scale": 4.0, "retries": 0, "seconds": 507.2, "max_memory_gb": 17.07, "ce": 0.2029049691336695, "rank": 0.02446376532316208}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 5.094605922698975, "loss_scale": 4.0, "retries": 0, "seconds": 509.0, "max_memory_gb": 17.07, "ce": 0.05657307500950992, "rank": 0.0}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 0.4117088317871094, "loss_scale": 4.0, "retries": 0, "seconds": 510.8, "max_memory_gb": 17.07, "ce": 0.0222015762119554, "rank": 0.0}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.088632106781006, "loss_scale": 4.0, "retries": 0, "seconds": 512.6, "max_memory_gb": 17.07, "ce": 0.1393734377634246, "rank": 0.0}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 0.7824654579162598, "loss_scale": 4.0, "retries": 0, "seconds": 514.4, "max_memory_gb": 17.07, "ce": 0.07722522303811274, "rank": 0.0}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 0.9252685308456421, "loss_scale": 4.0, "retries": 0, "seconds": 516.3, "max_memory_gb": 17.07, "ce": 0.08631628754665144, "rank": 0.0}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 1.471548080444336, "loss_scale": 4.0, "retries": 0, "seconds": 518.1, "max_memory_gb": 17.07, "ce": 0.027119736914755777, "rank": 0.0}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 1.1858681440353394, "loss_scale": 4.0, "retries": 0, "seconds": 520.1, "max_memory_gb": 17.07, "ce": 0.08690816338639706, "rank": 0.0}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 1.0755336284637451, "loss_scale": 4.0, "retries": 0, "seconds": 522.1, "max_memory_gb": 17.07, "ce": 0.05342352596926503, "rank": 0.0}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 2.1239378452301025, "loss_scale": 4.0, "retries": 0, "seconds": 524.0, "max_memory_gb": 17.07, "ce": 0.12382787914248183, "rank": 0.0}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 0.1768985539674759, "loss_scale": 4.0, "retries": 0, "seconds": 525.9, "max_memory_gb": 17.07, "ce": 0.008555743144825101, "rank": 0.0}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 2.7011189460754395, "loss_scale": 4.0, "retries": 0, "seconds": 527.7, "max_memory_gb": 17.07, "ce": 0.06649641162948683, "rank": 0.0}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 0.7964928150177002, "loss_scale": 4.0, "retries": 0, "seconds": 529.5, "max_memory_gb": 17.07, "ce": 0.07938079886662308, "rank": 0.0}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 9.334131240844727, "loss_scale": 4.0, "retries": 0, "seconds": 531.4, "max_memory_gb": 17.07, "ce": 0.03195944341132417, "rank": 0.0}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 15.70351791381836, "loss_scale": 4.0, "retries": 0, "seconds": 533.2, "max_memory_gb": 17.07, "ce": 0.2056186793197412, "rank": 0.0}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 0.7804127335548401, "loss_scale": 4.0, "retries": 0, "seconds": 535.0, "max_memory_gb": 17.07, "ce": 0.0413195654109586, "rank": 0.0}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 42.559505462646484, "loss_scale": 4.0, "retries": 0, "seconds": 536.8, "max_memory_gb": 17.07, "ce": 0.7230882995063439, "rank": 0.0}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 68.01610565185547, "loss_scale": 4.0, "retries": 0, "seconds": 538.6, "max_memory_gb": 17.07, "ce": 0.8065586521988735, "rank": 0.0}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 15.884777069091797, "loss_scale": 4.0, "retries": 0, "seconds": 540.4, "max_memory_gb": 17.07, "ce": 0.401228605711367, "rank": 0.0}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 1.9231709241867065, "loss_scale": 4.0, "retries": 0, "seconds": 542.4, "max_memory_gb": 17.07, "ce": 0.14967313458328135, "rank": 0.0}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 3.08164119720459, "loss_scale": 4.0, "retries": 0, "seconds": 544.3, "max_memory_gb": 17.07, "ce": 0.28756556364533026, "rank": 0.0}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 0.628914475440979, "loss_scale": 4.0, "retries": 0, "seconds": 546.5, "max_memory_gb": 17.07, "ce": 0.033574091794434935, "rank": 0.0}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 1.63041090965271, "loss_scale": 4.0, "retries": 0, "seconds": 548.4, "max_memory_gb": 17.07, "ce": 0.22314166161231697, "rank": 0.0}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 1.3514456748962402, "loss_scale": 4.0, "retries": 0, "seconds": 550.3, "max_memory_gb": 17.07, "ce": 0.10273840703302994, "rank": 0.0}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 0.28051647543907166, "loss_scale": 4.0, "retries": 0, "seconds": 552.3, "max_memory_gb": 17.07, "ce": 0.012601219990756363, "rank": 0.0}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 5.153307914733887, "loss_scale": 4.0, "retries": 0, "seconds": 554.3, "max_memory_gb": 17.07, "ce": 0.055778907335479744, "rank": 0.0}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 2.571110725402832, "loss_scale": 4.0, "retries": 0, "seconds": 556.3, "max_memory_gb": 17.07, "ce": 0.10168240263010375, "rank": 0.0}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 3.658956289291382, "loss_scale": 4.0, "retries": 0, "seconds": 558.3, "max_memory_gb": 17.07, "ce": 0.13872838273528032, "rank": 0.0}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 0.7847083210945129, "loss_scale": 4.0, "retries": 0, "seconds": 560.3, "max_memory_gb": 17.07, "ce": 0.02778893441427499, "rank": 0.0}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 2.080176830291748, "loss_scale": 4.0, "retries": 0, "seconds": 562.1, "max_memory_gb": 17.07, "ce": 0.19744208896008786, "rank": 0.0}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 3.392178535461426, "loss_scale": 4.0, "retries": 0, "seconds": 564.1, "max_memory_gb": 17.07, "ce": 0.2156998931968701, "rank": 0.0}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 2.1392362117767334, "loss_scale": 4.0, "retries": 0, "seconds": 566.1, "max_memory_gb": 17.07, "ce": 0.17782247222203296, "rank": 0.0}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 0.6322652697563171, "loss_scale": 4.0, "retries": 0, "seconds": 568.1, "max_memory_gb": 17.07, "ce": 0.05901012473623268, "rank": 0.0}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 1.4192384481430054, "loss_scale": 4.0, "retries": 0, "seconds": 570.0, "max_memory_gb": 17.07, "ce": 0.1494857475772733, "rank": 0.0}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 0.37757357954978943, "loss_scale": 4.0, "retries": 0, "seconds": 571.8, "max_memory_gb": 17.07, "ce": 0.021352917843614705, "rank": 0.0}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 2.7939817905426025, "loss_scale": 4.0, "retries": 0, "seconds": 573.7, "max_memory_gb": 17.07, "ce": 0.19070083606493426, "rank": 0.0}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 2.0175364017486572, "loss_scale": 4.0, "retries": 0, "seconds": 575.7, "max_memory_gb": 17.07, "ce": 0.2464748747224803, "rank": 0.0}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 0.6453493237495422, "loss_scale": 4.0, "retries": 0, "seconds": 577.5, "max_memory_gb": 17.07, "ce": 0.02980159666913096, "rank": 0.10901421308517456}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 1.7508472204208374, "loss_scale": 4.0, "retries": 0, "seconds": 579.4, "max_memory_gb": 17.07, "ce": 0.3177697026039823, "rank": 0.0}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 0.9506458044052124, "loss_scale": 4.0, "retries": 0, "seconds": 581.2, "max_memory_gb": 17.07, "ce": 0.08489194946014322, "rank": 0.0}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 50.45574951171875, "loss_scale": 4.0, "retries": 0, "seconds": 583.1, "max_memory_gb": 17.07, "ce": 0.8853241297183558, "rank": 0.0}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 1.781835913658142, "loss_scale": 4.0, "retries": 0, "seconds": 585.2, "max_memory_gb": 17.07, "ce": 0.11787933336745482, "rank": 0.0}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.038446307182312, "loss_scale": 4.0, "retries": 0, "seconds": 587.1, "max_memory_gb": 17.07, "ce": 0.09307682531652972, "rank": 0.0}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 7.589044570922852, "loss_scale": 4.0, "retries": 0, "seconds": 589.1, "max_memory_gb": 17.07, "ce": 0.11414150684140623, "rank": 0.1251712292432785}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 1.2987220287322998, "loss_scale": 4.0, "retries": 0, "seconds": 591.2, "max_memory_gb": 17.07, "ce": 0.020313454602728598, "rank": 0.0}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 1.6606528759002686, "loss_scale": 4.0, "retries": 0, "seconds": 593.3, "max_memory_gb": 17.07, "ce": 0.020242636936018243, "rank": 0.0}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 0.9401059150695801, "loss_scale": 4.0, "retries": 0, "seconds": 595.3, "max_memory_gb": 17.07, "ce": 0.01790972917660838, "rank": 0.0}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 1.1931896209716797, "loss_scale": 4.0, "retries": 0, "seconds": 597.2, "max_memory_gb": 17.07, "ce": 0.19086760250502266, "rank": 0.0}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 0.7039098143577576, "loss_scale": 4.0, "retries": 0, "seconds": 599.0, "max_memory_gb": 17.07, "ce": 0.03644234454986872, "rank": 0.10209010541439056}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 6.536787509918213, "loss_scale": 4.0, "retries": 0, "seconds": 600.9, "max_memory_gb": 17.07, "ce": 0.1693599794380134, "rank": 0.0}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 1.1395412683486938, "loss_scale": 4.0, "retries": 0, "seconds": 602.9, "max_memory_gb": 17.07, "ce": 0.12233418538380647, "rank": 0.0}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 1.1820497512817383, "loss_scale": 4.0, "retries": 0, "seconds": 604.7, "max_memory_gb": 17.07, "ce": 0.14693437547248323, "rank": 0.0}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 45.76583480834961, "loss_scale": 4.0, "retries": 0, "seconds": 606.5, "max_memory_gb": 17.07, "ce": 0.32283724961598637, "rank": 0.0}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 2.909571647644043, "loss_scale": 4.0, "retries": 0, "seconds": 608.3, "max_memory_gb": 17.07, "ce": 0.27106787836237345, "rank": 0.0}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 1.7114171981811523, "loss_scale": 4.0, "retries": 0, "seconds": 610.2, "max_memory_gb": 17.07, "ce": 0.10103133914526552, "rank": 0.0}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 1.2813973426818848, "loss_scale": 4.0, "retries": 0, "seconds": 611.9, "max_memory_gb": 17.07, "ce": 0.19671695585202542, "rank": 0.0}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 1.121322512626648, "loss_scale": 4.0, "retries": 0, "seconds": 613.8, "max_memory_gb": 17.07, "ce": 0.09377096392563544, "rank": 0.0}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 4.404477119445801, "loss_scale": 4.0, "retries": 0, "seconds": 615.6, "max_memory_gb": 17.07, "ce": 0.35555448102240916, "rank": 0.0}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 2.9169607162475586, "loss_scale": 4.0, "retries": 0, "seconds": 617.4, "max_memory_gb": 17.07, "ce": 0.033682318124192534, "rank": 0.0}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 1.8644278049468994, "loss_scale": 4.0, "retries": 0, "seconds": 619.2, "max_memory_gb": 17.07, "ce": 0.13686623666217201, "rank": 0.0}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.1037485599517822, "loss_scale": 4.0, "retries": 0, "seconds": 621.0, "max_memory_gb": 17.07, "ce": 0.16998048721870873, "rank": 0.0}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 0.8952434659004211, "loss_scale": 4.0, "retries": 0, "seconds": 622.9, "max_memory_gb": 17.07, "ce": 0.09929678354455973, "rank": 0.0}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 1.492811679840088, "loss_scale": 4.0, "retries": 0, "seconds": 624.7, "max_memory_gb": 17.07, "ce": 0.22432842021225952, "rank": 0.0}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 3.5118789672851562, "loss_scale": 4.0, "retries": 0, "seconds": 626.5, "max_memory_gb": 17.07, "ce": 0.18314072635985212, "rank": 0.0}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 2.385190963745117, "loss_scale": 4.0, "retries": 0, "seconds": 628.4, "max_memory_gb": 17.07, "ce": 0.5240926992664754, "rank": 0.0}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 0.7259657382965088, "loss_scale": 4.0, "retries": 0, "seconds": 630.2, "max_memory_gb": 17.07, "ce": 0.0599660151056014, "rank": 0.0}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 0.5817398428916931, "loss_scale": 4.0, "retries": 0, "seconds": 632.0, "max_memory_gb": 17.07, "ce": 0.021431030068924883, "rank": 0.0}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 0.7391554713249207, "loss_scale": 4.0, "retries": 0, "seconds": 633.7, "max_memory_gb": 17.07, "ce": 0.15838755373988533, "rank": 0.0}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 1.8830976486206055, "loss_scale": 4.0, "retries": 0, "seconds": 635.6, "max_memory_gb": 17.07, "ce": 0.22593518062421936, "rank": 0.0}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 1.555491328239441, "loss_scale": 4.0, "retries": 0, "seconds": 637.3, "max_memory_gb": 17.07, "ce": 0.1452798682730645, "rank": 0.0}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 0.5640721321105957, "loss_scale": 4.0, "retries": 0, "seconds": 639.1, "max_memory_gb": 17.07, "ce": 0.010761227127659367, "rank": 0.0}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 0.3296874761581421, "loss_scale": 4.0, "retries": 0, "seconds": 641.0, "max_memory_gb": 17.07, "ce": 0.07659735067500151, "rank": 0.0}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 0.5727271437644958, "loss_scale": 4.0, "retries": 0, "seconds": 642.8, "max_memory_gb": 17.07, "ce": 0.05807900059153326, "rank": 0.0}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 2.834155559539795, "loss_scale": 4.0, "retries": 0, "seconds": 644.6, "max_memory_gb": 17.07, "ce": 0.08336102581961313, "rank": 0.0}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 2.8592257499694824, "loss_scale": 4.0, "retries": 0, "seconds": 646.4, "max_memory_gb": 17.07, "ce": 0.25858549023541855, "rank": 0.0}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 1.5912189483642578, "loss_scale": 4.0, "retries": 0, "seconds": 648.2, "max_memory_gb": 17.07, "ce": 0.1342949871905148, "rank": 0.0}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 1.1700838804244995, "loss_scale": 4.0, "retries": 0, "seconds": 650.0, "max_memory_gb": 17.07, "ce": 0.1440481521131005, "rank": 0.0}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 0.5560156106948853, "loss_scale": 4.0, "retries": 0, "seconds": 651.8, "max_memory_gb": 17.07, "ce": 0.06329603867561673, "rank": 0.0}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 2.3460693359375, "loss_scale": 4.0, "retries": 0, "seconds": 653.6, "max_memory_gb": 17.07, "ce": 0.09214880967192585, "rank": 0.0}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 1.0585005283355713, "loss_scale": 4.0, "retries": 0, "seconds": 655.5, "max_memory_gb": 17.07, "ce": 0.06412978527987434, "rank": 0.0}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 1.3235357999801636, "loss_scale": 4.0, "retries": 0, "seconds": 657.2, "max_memory_gb": 17.07, "ce": 0.1273672202587477, "rank": 0.0}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 1.8152165412902832, "loss_scale": 4.0, "retries": 0, "seconds": 659.0, "max_memory_gb": 17.07, "ce": 0.061949584109243006, "rank": 0.0}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 1.4361671209335327, "loss_scale": 4.0, "retries": 0, "seconds": 660.8, "max_memory_gb": 17.07, "ce": 0.0548218012390862, "rank": 0.09106852114200592}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 1.7425620555877686, "loss_scale": 4.0, "retries": 0, "seconds": 662.6, "max_memory_gb": 17.07, "ce": 0.19132329845888307, "rank": 0.0}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 3.169404983520508, "loss_scale": 4.0, "retries": 0, "seconds": 664.4, "max_memory_gb": 17.07, "ce": 0.1975895536943426, "rank": 0.0}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 0.44860103726387024, "loss_scale": 4.0, "retries": 0, "seconds": 666.2, "max_memory_gb": 17.07, "ce": 0.022088059717134456, "rank": 0.0}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 0.9096277952194214, "loss_scale": 4.0, "retries": 0, "seconds": 668.1, "max_memory_gb": 17.07, "ce": 0.0352019276797364, "rank": 0.0}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 1.1860151290893555, "loss_scale": 4.0, "retries": 0, "seconds": 669.9, "max_memory_gb": 17.07, "ce": 0.0909723480226603, "rank": 0.0}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 11.783330917358398, "loss_scale": 4.0, "retries": 0, "seconds": 671.7, "max_memory_gb": 17.07, "ce": 0.18876429425199603, "rank": 0.0}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 0.8285585641860962, "loss_scale": 4.0, "retries": 0, "seconds": 673.5, "max_memory_gb": 17.07, "ce": 0.0667541131242615, "rank": 0.0}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 0.3539080321788788, "loss_scale": 4.0, "retries": 0, "seconds": 675.3, "max_memory_gb": 17.07, "ce": 0.010033420734544052, "rank": 0.0}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 0.8152218461036682, "loss_scale": 4.0, "retries": 0, "seconds": 677.1, "max_memory_gb": 17.07, "ce": 0.028779318614397198, "rank": 0.1134939044713974}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 5.043531894683838, "loss_scale": 4.0, "retries": 0, "seconds": 678.9, "max_memory_gb": 17.07, "ce": 0.44636870357135194, "rank": 0.0}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 1.9784172773361206, "loss_scale": 4.0, "retries": 0, "seconds": 680.8, "max_memory_gb": 17.07, "ce": 0.08845300573102577, "rank": 0.0}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 0.5398197174072266, "loss_scale": 4.0, "retries": 0, "seconds": 682.6, "max_memory_gb": 17.07, "ce": 0.019825744933768874, "rank": 0.0}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 1.4779895544052124, "loss_scale": 4.0, "retries": 0, "seconds": 684.4, "max_memory_gb": 17.07, "ce": 0.08947414522299368, "rank": 0.0}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 1.094470500946045, "loss_scale": 4.0, "retries": 0, "seconds": 686.3, "max_memory_gb": 17.07, "ce": 0.14714987920160638, "rank": 0.0}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 0.9864087104797363, "loss_scale": 4.0, "retries": 0, "seconds": 688.1, "max_memory_gb": 17.07, "ce": 0.14969967745491886, "rank": 0.0}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 0.5909481644630432, "loss_scale": 4.0, "retries": 0, "seconds": 689.9, "max_memory_gb": 17.07, "ce": 0.03901681371007726, "rank": 0.0}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 0.7060020565986633, "loss_scale": 4.0, "retries": 0, "seconds": 691.6, "max_memory_gb": 17.07, "ce": 0.07379931965806463, "rank": 0.0}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 1.529030442237854, "loss_scale": 4.0, "retries": 0, "seconds": 693.4, "max_memory_gb": 17.07, "ce": 0.08481838408624753, "rank": 0.0}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 0.7604690790176392, "loss_scale": 4.0, "retries": 0, "seconds": 695.2, "max_memory_gb": 17.07, "ce": 0.10241743221558863, "rank": 0.0}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 9.233123779296875, "loss_scale": 4.0, "retries": 0, "seconds": 697.1, "max_memory_gb": 17.07, "ce": 0.22410575494450313, "rank": 0.0}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 1.193069338798523, "loss_scale": 4.0, "retries": 0, "seconds": 698.9, "max_memory_gb": 17.07, "ce": 0.21337099706215668, "rank": 0.0}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 2.52664852142334, "loss_scale": 4.0, "retries": 0, "seconds": 700.7, "max_memory_gb": 17.07, "ce": 0.3737221070095984, "rank": 0.0}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 0.4704552888870239, "loss_scale": 4.0, "retries": 0, "seconds": 702.7, "max_memory_gb": 17.07, "ce": 0.08494193402111705, "rank": 0.0}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 2.0477209091186523, "loss_scale": 4.0, "retries": 0, "seconds": 704.7, "max_memory_gb": 17.07, "ce": 0.04511750739402487, "rank": 0.0}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 0.7494965195655823, "loss_scale": 4.0, "retries": 0, "seconds": 706.7, "max_memory_gb": 17.07, "ce": 0.10933386861324834, "rank": 0.0}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 0.7021027207374573, "loss_scale": 4.0, "retries": 0, "seconds": 708.6, "max_memory_gb": 17.07, "ce": 0.04124580473580863, "rank": 0.0}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 0.41642898321151733, "loss_scale": 4.0, "retries": 0, "seconds": 710.7, "max_memory_gb": 17.07, "ce": 0.04252803751114698, "rank": 0.0}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 1.2018300294876099, "loss_scale": 4.0, "retries": 0, "seconds": 712.6, "max_memory_gb": 17.07, "ce": 0.037917776564427186, "rank": 0.0}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 0.4230819344520569, "loss_scale": 4.0, "retries": 0, "seconds": 714.6, "max_memory_gb": 17.07, "ce": 0.00920708458579611, "rank": 0.0}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 26.255252838134766, "loss_scale": 4.0, "retries": 0, "seconds": 716.4, "max_memory_gb": 17.07, "ce": 0.25619838587135746, "rank": 0.0}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 1.6634701490402222, "loss_scale": 4.0, "retries": 0, "seconds": 718.3, "max_memory_gb": 17.07, "ce": 0.23548005097836722, "rank": 0.0}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 0.038284510374069214, "loss_scale": 4.0, "retries": 0, "seconds": 720.1, "max_memory_gb": 17.07, "ce": 0.0013799599973935983, "rank": 0.12357497960329056}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 1.2120660543441772, "loss_scale": 4.0, "retries": 0, "seconds": 721.9, "max_memory_gb": 17.07, "ce": 0.10641674950602464, "rank": 0.0}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 1.4086698293685913, "loss_scale": 4.0, "retries": 0, "seconds": 723.7, "max_memory_gb": 17.07, "ce": 0.1023281363713977, "rank": 0.0}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 0.21180853247642517, "loss_scale": 4.0, "retries": 0, "seconds": 725.5, "max_memory_gb": 17.07, "ce": 0.0059830971749761375, "rank": 0.0}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 1.0071762800216675, "loss_scale": 4.0, "retries": 0, "seconds": 727.3, "max_memory_gb": 17.07, "ce": 0.043668870738201804, "rank": 0.0}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 0.5501851439476013, "loss_scale": 4.0, "retries": 0, "seconds": 729.1, "max_memory_gb": 17.07, "ce": 0.08653694298118353, "rank": 0.0}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 0.5763351321220398, "loss_scale": 4.0, "retries": 0, "seconds": 730.9, "max_memory_gb": 17.07, "ce": 0.058667441756369954, "rank": 0.0}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 2.583360433578491, "loss_scale": 4.0, "retries": 0, "seconds": 732.7, "max_memory_gb": 17.07, "ce": 0.19861409895929683, "rank": 0.0}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 13.158416748046875, "loss_scale": 4.0, "retries": 0, "seconds": 734.6, "max_memory_gb": 17.07, "ce": 0.0894062619986471, "rank": 0.0}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 2.815885305404663, "loss_scale": 4.0, "retries": 0, "seconds": 736.4, "max_memory_gb": 17.07, "ce": 0.1703111567421729, "rank": 0.0}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 1.0705885887145996, "loss_scale": 4.0, "retries": 0, "seconds": 738.2, "max_memory_gb": 17.07, "ce": 0.09370390299045539, "rank": 0.11385632306337357}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 0.18010492622852325, "loss_scale": 4.0, "retries": 0, "seconds": 740.0, "max_memory_gb": 17.07, "ce": 0.004503507983372401, "rank": 0.0}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 1.9644784927368164, "loss_scale": 4.0, "retries": 0, "seconds": 741.9, "max_memory_gb": 17.07, "ce": 0.15923804651811224, "rank": 0.0}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 0.2645432651042938, "loss_scale": 4.0, "retries": 0, "seconds": 743.6, "max_memory_gb": 17.07, "ce": 0.00860914092118037, "rank": 0.0}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 0.9108064770698547, "loss_scale": 4.0, "retries": 0, "seconds": 745.4, "max_memory_gb": 17.07, "ce": 0.0961341850647841, "rank": 0.0}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 3.9914767742156982, "loss_scale": 4.0, "retries": 0, "seconds": 747.2, "max_memory_gb": 17.07, "ce": 0.7934043400018709, "rank": 0.26582443714141846}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 2.100152015686035, "loss_scale": 4.0, "retries": 0, "seconds": 749.1, "max_memory_gb": 17.07, "ce": 0.3458219975600514, "rank": 0.0}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 0.5140936970710754, "loss_scale": 4.0, "retries": 0, "seconds": 750.9, "max_memory_gb": 17.07, "ce": 0.12971851189467998, "rank": 0.0}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 0.6014256477355957, "loss_scale": 4.0, "retries": 0, "seconds": 752.6, "max_memory_gb": 17.07, "ce": 0.015713756347395247, "rank": 0.0}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 4.582629203796387, "loss_scale": 4.0, "retries": 0, "seconds": 754.4, "max_memory_gb": 17.07, "ce": 0.18660980821368867, "rank": 0.0}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 0.41923099756240845, "loss_scale": 4.0, "retries": 0, "seconds": 756.2, "max_memory_gb": 17.07, "ce": 0.13411518688917567, "rank": 0.12494200468063354}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 37.57792663574219, "loss_scale": 4.0, "retries": 0, "seconds": 758.0, "max_memory_gb": 17.07, "ce": 0.12838644294834012, "rank": 0.0}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 46.84855651855469, "loss_scale": 4.0, "retries": 0, "seconds": 759.9, "max_memory_gb": 17.07, "ce": 1.0953660190152732, "rank": 0.0}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 3.1212525367736816, "loss_scale": 4.0, "retries": 0, "seconds": 761.7, "max_memory_gb": 17.07, "ce": 0.3671240887360909, "rank": 0.0}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 0.2585749626159668, "loss_scale": 4.0, "retries": 0, "seconds": 763.4, "max_memory_gb": 17.07, "ce": 0.01106073034588917, "rank": 0.0}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 2.6182937622070312, "loss_scale": 4.0, "retries": 0, "seconds": 765.2, "max_memory_gb": 17.07, "ce": 0.4173480944700714, "rank": 0.12703584134578705}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 0.5765938758850098, "loss_scale": 4.0, "retries": 0, "seconds": 767.1, "max_memory_gb": 17.07, "ce": 0.04368669150608184, "rank": 0.0}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 0.3604014813899994, "loss_scale": 4.0, "retries": 0, "seconds": 768.8, "max_memory_gb": 17.07, "ce": 0.04917530338298093, "rank": 0.0}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 2.34415864944458, "loss_scale": 4.0, "retries": 0, "seconds": 770.7, "max_memory_gb": 17.07, "ce": 0.2574216496814188, "rank": 0.0}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 0.3689505159854889, "loss_scale": 4.0, "retries": 0, "seconds": 772.7, "max_memory_gb": 17.07, "ce": 0.022059370305214543, "rank": 0.12283601611852646}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 3.319200277328491, "loss_scale": 4.0, "retries": 0, "seconds": 774.7, "max_memory_gb": 17.07, "ce": 0.24989836822032885, "rank": 0.0}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 2.886554002761841, "loss_scale": 4.0, "retries": 0, "seconds": 776.7, "max_memory_gb": 17.07, "ce": 0.34270689026016043, "rank": 0.0}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 0.8400687575340271, "loss_scale": 4.0, "retries": 0, "seconds": 778.7, "max_memory_gb": 17.07, "ce": 0.03323235437255789, "rank": 0.0}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 0.7354440689086914, "loss_scale": 4.0, "retries": 0, "seconds": 780.7, "max_memory_gb": 17.07, "ce": 0.031023762370750774, "rank": 0.0}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 1.697222113609314, "loss_scale": 4.0, "retries": 0, "seconds": 782.8, "max_memory_gb": 17.07, "ce": 0.1520396564665134, "rank": 0.0}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 0.5404331088066101, "loss_scale": 4.0, "retries": 0, "seconds": 784.8, "max_memory_gb": 17.07, "ce": 0.03423290415594238, "rank": 0.0}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 0.6819256544113159, "loss_scale": 4.0, "retries": 0, "seconds": 786.9, "max_memory_gb": 17.07, "ce": 0.0231154969878844, "rank": 0.0}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 0.23510095477104187, "loss_scale": 4.0, "retries": 0, "seconds": 788.8, "max_memory_gb": 17.07, "ce": 0.009597711679816712, "rank": 0.0}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 0.7846890091896057, "loss_scale": 4.0, "retries": 0, "seconds": 790.8, "max_memory_gb": 17.07, "ce": 0.08374249450662319, "rank": 0.0}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.106893062591553, "loss_scale": 4.0, "retries": 0, "seconds": 792.8, "max_memory_gb": 17.07, "ce": 0.3378004596552273, "rank": 0.0}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 0.8292356729507446, "loss_scale": 4.0, "retries": 0, "seconds": 794.8, "max_memory_gb": 17.07, "ce": 0.05237989705710788, "rank": 0.0}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 5.145674705505371, "loss_scale": 4.0, "retries": 0, "seconds": 796.7, "max_memory_gb": 17.07, "ce": 0.09289197684029205, "rank": 0.0}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 0.3372342586517334, "loss_scale": 4.0, "retries": 0, "seconds": 798.5, "max_memory_gb": 17.07, "ce": 0.005705777704861248, "rank": 0.0}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 1.9027626514434814, "loss_scale": 4.0, "retries": 0, "seconds": 800.2, "max_memory_gb": 17.07, "ce": 0.16114760867094446, "rank": 0.0}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 1.7734777927398682, "loss_scale": 4.0, "retries": 0, "seconds": 802.0, "max_memory_gb": 17.07, "ce": 0.1483695980741686, "rank": 0.0}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 3.6587629318237305, "loss_scale": 4.0, "retries": 0, "seconds": 803.8, "max_memory_gb": 17.07, "ce": 0.20620235965498068, "rank": 0.11472561210393906}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 0.9274665117263794, "loss_scale": 4.0, "retries": 0, "seconds": 805.8, "max_memory_gb": 17.07, "ce": 0.04244358319192543, "rank": 0.0}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 1.0503536462783813, "loss_scale": 4.0, "retries": 0, "seconds": 807.7, "max_memory_gb": 17.07, "ce": 0.10445101263758261, "rank": 0.0}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 0.6089009046554565, "loss_scale": 4.0, "retries": 0, "seconds": 809.7, "max_memory_gb": 17.07, "ce": 0.03384561899656546, "rank": 0.0}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 5.193615913391113, "loss_scale": 4.0, "retries": 0, "seconds": 811.6, "max_memory_gb": 17.07, "ce": 0.4015672987425205, "rank": 0.0}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 0.4811243712902069, "loss_scale": 4.0, "retries": 0, "seconds": 813.5, "max_memory_gb": 17.07, "ce": 0.043591915845809126, "rank": 0.0}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 0.6791467666625977, "loss_scale": 4.0, "retries": 0, "seconds": 815.6, "max_memory_gb": 17.07, "ce": 0.04577762757708115, "rank": 0.0}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 0.46420904994010925, "loss_scale": 4.0, "retries": 0, "seconds": 817.5, "max_memory_gb": 17.07, "ce": 0.043870072889149014, "rank": 0.0}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 1.9517804384231567, "loss_scale": 4.0, "retries": 0, "seconds": 819.4, "max_memory_gb": 17.07, "ce": 0.35235829099337934, "rank": 0.08150660991668701}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 9.789992332458496, "loss_scale": 4.0, "retries": 0, "seconds": 821.2, "max_memory_gb": 17.07, "ce": 0.09337835057112898, "rank": 0.0}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 0.5306103825569153, "loss_scale": 4.0, "retries": 0, "seconds": 823.0, "max_memory_gb": 17.07, "ce": 0.12151554825823041, "rank": 0.0}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 3.0144312381744385, "loss_scale": 4.0, "retries": 0, "seconds": 824.8, "max_memory_gb": 17.07, "ce": 0.1495944716843951, "rank": 0.0}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 0.36113861203193665, "loss_scale": 4.0, "retries": 0, "seconds": 826.5, "max_memory_gb": 17.07, "ce": 0.007375790482456068, "rank": 0.0}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 1.9390101432800293, "loss_scale": 4.0, "retries": 0, "seconds": 828.4, "max_memory_gb": 17.07, "ce": 0.07605762491311907, "rank": 0.0}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 2.60428524017334, "loss_scale": 4.0, "retries": 0, "seconds": 830.2, "max_memory_gb": 17.07, "ce": 0.06502909108166932, "rank": 0.0}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 1.1738529205322266, "loss_scale": 4.0, "retries": 0, "seconds": 832.0, "max_memory_gb": 17.07, "ce": 0.15241625948237925, "rank": 0.0}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 1.9409841299057007, "loss_scale": 4.0, "retries": 0, "seconds": 833.8, "max_memory_gb": 17.07, "ce": 0.0578440132585456, "rank": 0.0}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 18.207061767578125, "loss_scale": 4.0, "retries": 0, "seconds": 835.6, "max_memory_gb": 17.07, "ce": 0.2971640246714742, "rank": 0.0}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 3.6654105186462402, "loss_scale": 4.0, "retries": 0, "seconds": 837.5, "max_memory_gb": 17.07, "ce": 0.1011205058653104, "rank": 0.0}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 0.6985408067703247, "loss_scale": 4.0, "retries": 0, "seconds": 839.3, "max_memory_gb": 17.07, "ce": 0.016042737724546896, "rank": 0.0}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 0.45658233761787415, "loss_scale": 4.0, "retries": 0, "seconds": 841.1, "max_memory_gb": 17.07, "ce": 0.02735866485909355, "rank": 0.0}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 1.9658384323120117, "loss_scale": 4.0, "retries": 0, "seconds": 842.9, "max_memory_gb": 17.07, "ce": 0.20057556172741897, "rank": 0.0}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 2.697204351425171, "loss_scale": 4.0, "retries": 0, "seconds": 844.7, "max_memory_gb": 17.07, "ce": 0.11730930613975943, "rank": 0.0}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 0.14157050848007202, "loss_scale": 4.0, "retries": 0, "seconds": 846.5, "max_memory_gb": 17.07, "ce": 0.005691257028047403, "rank": 0.0}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 3.169820547103882, "loss_scale": 4.0, "retries": 0, "seconds": 848.4, "max_memory_gb": 17.07, "ce": 0.32909538338344646, "rank": 0.0}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 0.6036567091941833, "loss_scale": 4.0, "retries": 0, "seconds": 850.2, "max_memory_gb": 17.07, "ce": 0.021676837135146343, "rank": 0.0}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 0.171522319316864, "loss_scale": 4.0, "retries": 0, "seconds": 852.0, "max_memory_gb": 17.07, "ce": 0.0050659929634093714, "rank": 0.0}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 0.05187675729393959, "loss_scale": 4.0, "retries": 0, "seconds": 853.8, "max_memory_gb": 17.07, "ce": 0.00224459679066058, "rank": 0.0}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 2.495830774307251, "loss_scale": 4.0, "retries": 0, "seconds": 855.5, "max_memory_gb": 17.07, "ce": 0.13217426855226222, "rank": 0.0}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 1.4948444366455078, "loss_scale": 4.0, "retries": 0, "seconds": 857.4, "max_memory_gb": 17.07, "ce": 0.08219582773926959, "rank": 0.0}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 22.735654830932617, "loss_scale": 4.0, "retries": 0, "seconds": 859.2, "max_memory_gb": 17.07, "ce": 0.17132001294612564, "rank": 0.0}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 12.477293014526367, "loss_scale": 4.0, "retries": 0, "seconds": 861.0, "max_memory_gb": 17.07, "ce": 0.08297931605648046, "rank": 0.0}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 1.590257167816162, "loss_scale": 4.0, "retries": 0, "seconds": 862.8, "max_memory_gb": 17.07, "ce": 0.10943547349552318, "rank": 0.0}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 1.8161860704421997, "loss_scale": 4.0, "retries": 0, "seconds": 864.7, "max_memory_gb": 17.07, "ce": 0.07286103317255765, "rank": 0.0}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 2.3037586212158203, "loss_scale": 4.0, "retries": 0, "seconds": 866.5, "max_memory_gb": 17.07, "ce": 0.21536048196230695, "rank": 0.0}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 2.3302812576293945, "loss_scale": 4.0, "retries": 0, "seconds": 868.3, "max_memory_gb": 17.07, "ce": 0.04709954548798123, "rank": 0.0}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 0.3465645909309387, "loss_scale": 4.0, "retries": 0, "seconds": 870.0, "max_memory_gb": 17.07, "ce": 0.01326375893313525, "rank": 0.12434838712215424}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 1.0027607679367065, "loss_scale": 4.0, "retries": 0, "seconds": 871.9, "max_memory_gb": 17.07, "ce": 0.11332754966315406, "rank": 0.0}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 1.9046868085861206, "loss_scale": 4.0, "retries": 0, "seconds": 873.6, "max_memory_gb": 17.07, "ce": 0.18260013845429057, "rank": 0.0}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 0.21993429958820343, "loss_scale": 4.0, "retries": 0, "seconds": 875.5, "max_memory_gb": 17.07, "ce": 0.007899392839135544, "rank": 0.0}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 0.8609859943389893, "loss_scale": 4.0, "retries": 0, "seconds": 877.3, "max_memory_gb": 17.07, "ce": 0.03201362054414858, "rank": 0.0}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 1.2932738065719604, "loss_scale": 4.0, "retries": 0, "seconds": 879.1, "max_memory_gb": 17.07, "ce": 0.08712349830511812, "rank": 0.0}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 3.5237410068511963, "loss_scale": 4.0, "retries": 0, "seconds": 880.9, "max_memory_gb": 17.07, "ce": 0.3628507033849928, "rank": 0.0}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 0.5804839134216309, "loss_scale": 4.0, "retries": 0, "seconds": 882.6, "max_memory_gb": 17.07, "ce": 0.09859475244593341, "rank": 0.0}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 1.7507072687149048, "loss_scale": 4.0, "retries": 0, "seconds": 884.5, "max_memory_gb": 17.07, "ce": 0.04988055549938508, "rank": 0.0}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 3.2354159355163574, "loss_scale": 4.0, "retries": 0, "seconds": 886.2, "max_memory_gb": 17.07, "ce": 0.05846782234675629, "rank": 0.0}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 8.160881996154785, "loss_scale": 4.0, "retries": 0, "seconds": 888.0, "max_memory_gb": 17.07, "ce": 0.4901666967998608, "rank": 0.0}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 2.2569077014923096, "loss_scale": 4.0, "retries": 0, "seconds": 889.8, "max_memory_gb": 17.07, "ce": 0.4564782838760948, "rank": 0.0}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 0.9625498652458191, "loss_scale": 4.0, "retries": 0, "seconds": 891.6, "max_memory_gb": 17.07, "ce": 0.04622445131553832, "rank": 0.0}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 2.2782251834869385, "loss_scale": 4.0, "retries": 0, "seconds": 893.4, "max_memory_gb": 17.07, "ce": 0.17797823516571043, "rank": 0.0}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 17.91301918029785, "loss_scale": 4.0, "retries": 0, "seconds": 895.2, "max_memory_gb": 17.07, "ce": 0.2873985742030527, "rank": 0.0}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 1.0695672035217285, "loss_scale": 4.0, "retries": 0, "seconds": 897.0, "max_memory_gb": 17.07, "ce": 0.06901232023528792, "rank": 0.0}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 2.850689172744751, "loss_scale": 4.0, "retries": 0, "seconds": 898.8, "max_memory_gb": 17.07, "ce": 0.19351065369664866, "rank": 0.0}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 0.3209730088710785, "loss_scale": 4.0, "retries": 0, "seconds": 900.6, "max_memory_gb": 17.07, "ce": 0.01581470083874592, "rank": 0.0}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 0.8218419551849365, "loss_scale": 4.0, "retries": 0, "seconds": 902.4, "max_memory_gb": 17.07, "ce": 0.12537065009746584, "rank": 0.0}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 0.3429209291934967, "loss_scale": 4.0, "retries": 0, "seconds": 904.2, "max_memory_gb": 17.07, "ce": 0.009044051716045942, "rank": 0.0}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 0.86928391456604, "loss_scale": 4.0, "retries": 0, "seconds": 906.1, "max_memory_gb": 17.07, "ce": 0.04164593260247784, "rank": 0.0}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 1.4742517471313477, "loss_scale": 4.0, "retries": 0, "seconds": 907.9, "max_memory_gb": 17.07, "ce": 0.10186970197764822, "rank": 0.0}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 0.9923144578933716, "loss_scale": 4.0, "retries": 0, "seconds": 909.7, "max_memory_gb": 17.07, "ce": 0.04583333553364355, "rank": 0.0}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 0.29576727747917175, "loss_scale": 4.0, "retries": 0, "seconds": 911.5, "max_memory_gb": 17.07, "ce": 0.010443472214774374, "rank": 0.0}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 2.109375, "loss_scale": 4.0, "retries": 0, "seconds": 913.3, "max_memory_gb": 17.07, "ce": 0.15028985638355152, "rank": 0.0}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 0.9160761833190918, "loss_scale": 4.0, "retries": 0, "seconds": 915.0, "max_memory_gb": 17.07, "ce": 0.10262638111089473, "rank": 0.0}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 2.727335214614868, "loss_scale": 4.0, "retries": 0, "seconds": 916.8, "max_memory_gb": 17.07, "ce": 0.18307064038845056, "rank": 0.0}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 2.383162260055542, "loss_scale": 4.0, "retries": 0, "seconds": 918.6, "max_memory_gb": 17.07, "ce": 0.2101661121523648, "rank": 0.0}
diff --git a/runs/single_llava/replacement/L01_s13/adapter/README.md b/runs/single_llava/replacement/L01_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L01_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L01_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L01_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..edaf9930888a0efb104c2e3bd26f3d3f9546254f
--- /dev/null
+++ b/runs/single_llava/replacement/L01_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.1.mlp.down_proj",
+    "model.language_model.layers.1.mlp.up_proj",
+    "model.language_model.layers.1.self_attn.k_proj",
+    "model.language_model.layers.1.self_attn.o_proj",
+    "model.language_model.layers.1.mlp.gate_proj",
+    "model.language_model.layers.1.self_attn.v_proj",
+    "model.language_model.layers.1.self_attn.q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L01_s13/config.json b/runs/single_llava/replacement/L01_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..310c3c4eaf7b9538f971d5f5a4d78d2842bc0898
--- /dev/null
+++ b/runs/single_llava/replacement/L01_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 1,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L01_s13/steps.jsonl b/runs/single_llava/replacement/L01_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..31de761db9a42368beb6b7127da8e17da6e4d49b
--- /dev/null
+++ b/runs/single_llava/replacement/L01_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 1.2064614295959473, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 28.1, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 1.9203506708145142, "loss_scale": 4.0, "retries": 0, "seconds": 3.2, "max_memory_gb": 28.1, "pair": 1.0935343503952026, "sign": 0.5360194481909275, "keep": 5.372136370278895, "leak": 3.286371175192926e-05, "replacement": 4.047620445489883}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 2.9909889698028564, "loss_scale": 4.0, "retries": 0, "seconds": 4.7, "max_memory_gb": 28.21, "pair": 1.0886357426643372, "sign": 0.5389106776565313, "keep": 10.793509721755981, "leak": 8.720515489812897e-05, "replacement": 6.754867672920227}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 2.438544511795044, "loss_scale": 4.0, "retries": 0, "seconds": 6.2, "max_memory_gb": 28.23, "pair": 1.0868645012378693, "sign": 0.4686596356332302, "keep": 7.6591052347794175, "leak": 0.0002986343824886717, "replacement": 5.150821655988693}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 1.2520710229873657, "loss_scale": 4.0, "retries": 0, "seconds": 7.7, "max_memory_gb": 28.23, "pair": 1.0952904224395752, "sign": 0.7755317389965057, "keep": 0.15825707837939262, "leak": 9.664191480851514e-06, "replacement": 1.5621872544288635}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 1.366557002067566, "loss_scale": 4.0, "retries": 0, "seconds": 9.1, "max_memory_gb": 28.23, "pair": 1.090243399143219, "sign": 0.7347011715173721, "keep": 2.9229056499898434, "leak": 0.00024990429642457457, "replacement": 2.9191092550754547}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 2.9231042861938477, "loss_scale": 4.0, "retries": 0, "seconds": 10.6, "max_memory_gb": 28.23, "pair": 1.0588066279888153, "sign": 0.5443471390753984, "keep": 6.856645606458187, "leak": 0.0012899059238407062, "replacement": 4.7596255242824554}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 2.8990399837493896, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 28.23, "pair": 1.0666221976280212, "sign": 0.5811258107423782, "keep": 7.358190163969994, "leak": 0.0009478228332682193, "replacement": 5.036517083644867}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 2.7322309017181396, "loss_scale": 4.0, "retries": 0, "seconds": 13.5, "max_memory_gb": 28.23, "pair": 1.0557132065296173, "sign": 0.6164673883467913, "keep": 5.685506410896778, "leak": 0.0025880115633185596, "replacement": 4.207347273826599}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 1.9731487035751343, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 28.23, "pair": 1.0676902830600739, "sign": 0.777082059532404, "keep": 3.260420948266983, "leak": 0.0012309863577684155, "replacement": 3.0867496132850647}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 3.235236644744873, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 28.23, "pair": 1.0381332337856293, "sign": 0.45064304023981094, "keep": 5.23762709274888, "leak": 0.0040979551422424265, "replacement": 3.8832928240299225}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 2.8394651412963867, "loss_scale": 4.0, "retries": 0, "seconds": 17.9, "max_memory_gb": 28.23, "pair": 1.0478374361991882, "sign": 0.6042537540197372, "keep": 3.0315904188901186, "leak": 0.005192876700675697, "replacement": 2.867057740688324}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 2.8243892192840576, "loss_scale": 4.0, "retries": 0, "seconds": 19.4, "max_memory_gb": 28.23, "pair": 1.0552848875522614, "sign": 0.5020328983664513, "keep": 5.323097825981677, "leak": 0.0037944295092415814, "replacement": 3.9687987864017487}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 4.51459264755249, "loss_scale": 4.0, "retries": 0, "seconds": 20.8, "max_memory_gb": 28.23, "pair": 1.0073942989110947, "sign": 0.3911040537059307, "keep": 7.751185692846775, "leak": 0.006952615825071007, "replacement": 5.080277353525162}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 2.893348455429077, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 28.23, "pair": 1.039753645658493, "sign": 0.6872461810708046, "keep": 5.145512845134363, "leak": 0.005408968034316786, "replacement": 3.9574853479862213}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 2.510877847671509, "loss_scale": 4.0, "retries": 0, "seconds": 23.7, "max_memory_gb": 28.23, "pair": 1.0542737245559692, "sign": 0.7778775170445442, "keep": 2.7359928507357836, "leak": 0.002329099690541625, "replacement": 2.811791181564331}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 2.559412717819214, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 28.23, "pair": 1.0482424199581146, "sign": 0.7022040039300919, "keep": 3.1616072822362185, "leak": 0.008328939576131233, "replacement": 2.982230305671692}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 3.384047031402588, "loss_scale": 4.0, "retries": 0, "seconds": 26.6, "max_memory_gb": 28.31, "pair": 1.0372178852558136, "sign": 0.5310962684452534, "keep": 7.147120267152786, "leak": 0.004484221164602786, "replacement": 4.877447158098221}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 6.047946929931641, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 28.31, "pair": 0.9597442001104355, "sign": 0.5151579175144434, "keep": 10.643661499023438, "leak": 0.017065538791939616, "replacement": 6.543420195579529}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 1.8754479885101318, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 28.31, "pair": 1.0627508461475372, "sign": 0.4235028326511383, "keep": 0.022041441407054663, "leak": 0.0010271946011926048, "replacement": 1.2857797741889954}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 4.080712795257568, "loss_scale": 4.0, "retries": 0, "seconds": 31.0, "max_memory_gb": 28.31, "pair": 1.0108721554279327, "sign": 0.36516197957098484, "keep": 5.511942058801651, "leak": 0.004931041024974547, "replacement": 3.9506568908691406}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 3.9673988819122314, "loss_scale": 4.0, "retries": 0, "seconds": 32.5, "max_memory_gb": 28.31, "pair": 0.99996517598629, "sign": 0.5640296265482903, "keep": 5.889913931488991, "leak": 0.02488294823706383, "replacement": 4.233157724142075}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 5.116759777069092, "loss_scale": 4.0, "retries": 0, "seconds": 34.0, "max_memory_gb": 28.31, "pair": 0.9483979493379593, "sign": 0.23783420398831367, "keep": 5.285461414139718, "leak": 0.0242135941516608, "replacement": 3.716099262237549}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 6.650275230407715, "loss_scale": 4.0, "retries": 0, "seconds": 35.4, "max_memory_gb": 28.31, "pair": 0.9303543865680695, "sign": 0.24166646227240562, "keep": 10.002965450286865, "leak": 0.017152071348391473, "replacement": 6.056958317756653}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 5.9953436851501465, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 28.31, "pair": 0.9403261840343475, "sign": 0.3796186149120331, "keep": 8.417151860892773, "leak": 0.027077193321019877, "replacement": 5.345480740070343}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 7.337827682495117, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 28.31, "pair": 0.898433193564415, "sign": 0.6513789165765047, "keep": 10.45899248123169, "leak": 0.04399687459226698, "replacement": 6.464617967605591}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 5.106832027435303, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 28.31, "pair": 0.9555302113294601, "sign": 0.5836154073476791, "keep": 8.379963874816895, "leak": 0.0273619813087862, "replacement": 5.444160342216492}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 4.508091449737549, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 28.31, "pair": 1.0106732249259949, "sign": 0.7647982072085142, "keep": 5.601793512701988, "leak": 0.019592367548739276, "replacement": 4.19886714220047}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 3.3029260635375977, "loss_scale": 4.0, "retries": 0, "seconds": 42.7, "max_memory_gb": 28.31, "pair": 1.019159808754921, "sign": 0.837619261816144, "keep": 5.203643860295415, "leak": 0.013594830680176528, "replacement": 4.043190062046051}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 5.527347564697266, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 28.31, "pair": 0.9094822555780411, "sign": 0.3379298970103264, "keep": 5.4809863064438105, "leak": 0.05546025665535126, "replacement": 3.8328053951263428}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 4.521984577178955, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 28.31, "pair": 0.9640375673770905, "sign": 0.7963526584208012, "keep": 5.667712330818176, "leak": 0.04623612667637644, "replacement": 4.207629203796387}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 5.89735746383667, "loss_scale": 4.0, "retries": 0, "seconds": 47.2, "max_memory_gb": 28.31, "pair": 0.9225685000419617, "sign": 0.3705635294318199, "keep": 10.75143575668335, "leak": 0.050408539827913046, "replacement": 6.496170163154602}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 4.402333736419678, "loss_scale": 4.0, "retries": 0, "seconds": 48.7, "max_memory_gb": 28.31, "pair": 0.9733958095312119, "sign": 0.6091719977557659, "keep": 4.869804764166474, "leak": 0.0645468303191592, "replacement": 3.729020833969116}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 5.334419250488281, "loss_scale": 4.0, "retries": 0, "seconds": 50.2, "max_memory_gb": 28.31, "pair": 0.9623083025217056, "sign": 0.5099168978631496, "keep": 5.27544123865664, "leak": 0.019477405890938826, "replacement": 3.8598566949367523}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 6.531670570373535, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 28.31, "pair": 0.88686104118824, "sign": 0.4187486879527569, "keep": 5.446292300708592, "leak": 0.056503811036236584, "replacement": 3.8335074484348297}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 6.247174263000488, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 28.31, "pair": 0.8850264847278595, "sign": 0.41849714517593384, "keep": 5.151389390230179, "leak": 0.05123818043648498, "replacement": 3.6827792823314667}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 6.451098442077637, "loss_scale": 4.0, "retries": 0, "seconds": 54.5, "max_memory_gb": 28.31, "pair": 0.903120219707489, "sign": 0.5706776920706034, "keep": 4.909765258431435, "leak": 0.07286004293389325, "replacement": 3.66155669093132}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 10.223501205444336, "loss_scale": 4.0, "retries": 0, "seconds": 56.0, "max_memory_gb": 28.31, "pair": 0.8395250737667084, "sign": 0.7563560251146555, "keep": 9.42491364479065, "leak": 0.06878221756778657, "replacement": 5.947355389595032}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 4.182717800140381, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 28.31, "pair": 0.9851019084453583, "sign": 0.5539913102984428, "keep": 2.8585397116839886, "leak": 0.06307342604850419, "replacement": 2.707135856151581}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 5.825589179992676, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 28.31, "pair": 0.917019784450531, "sign": 0.47336290031671524, "keep": 5.536472365260124, "leak": 0.03770790284033865, "replacement": 3.9313643872737885}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 2.604853868484497, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 28.31, "pair": 1.0333302468061447, "sign": 1.0114381909370422, "keep": 2.9738420099020004, "leak": 0.020052772808412556, "replacement": 3.030983626842499}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 5.365666389465332, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 28.31, "pair": 0.9597094058990479, "sign": 0.3887745011597872, "keep": 4.6935293562710285, "leak": 0.024205413181334734, "replacement": 3.506912589073181}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 6.701169490814209, "loss_scale": 4.0, "retries": 0, "seconds": 63.1, "max_memory_gb": 28.31, "pair": 0.9313724637031555, "sign": 0.5008247345685959, "keep": 4.843857611529529, "leak": 0.024113090708851814, "replacement": 3.6097419559955597}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 4.614229202270508, "loss_scale": 4.0, "retries": 0, "seconds": 64.6, "max_memory_gb": 28.31, "pair": 0.9423378556966782, "sign": 0.5577771365642548, "keep": 2.7648770338855684, "leak": 0.03726099920459092, "replacement": 2.6129801869392395}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 3.58113694190979, "loss_scale": 4.0, "retries": 0, "seconds": 66.1, "max_memory_gb": 28.31, "pair": 1.0264026820659637, "sign": 0.8817729502916336, "keep": 2.8161230087280273, "leak": 0.008326309727067382, "replacement": 2.877432346343994}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 6.231398582458496, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 28.31, "pair": 0.9676700830459595, "sign": 0.7097924537956715, "keep": 4.571367770433426, "leak": 0.041740693342944724, "replacement": 3.618685483932495}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 7.241882801055908, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 28.31, "pair": 0.8455816060304642, "sign": 0.6343348901718855, "keep": 4.109103922732174, "leak": 0.059944244276266545, "replacement": 3.2322870194911957}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 9.410306930541992, "loss_scale": 4.0, "retries": 0, "seconds": 70.4, "max_memory_gb": 28.31, "pair": 0.7878901213407516, "sign": 0.19086104491725564, "keep": 7.605477266013622, "leak": 0.08362184837460518, "replacement": 4.706964671611786}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 9.89123821258545, "loss_scale": 4.0, "retries": 0, "seconds": 71.8, "max_memory_gb": 28.31, "pair": 0.8026157021522522, "sign": 0.5816869921982288, "keep": 7.1153698693960905, "leak": 0.10952712688595057, "replacement": 4.678525924682617}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 10.424017906188965, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 28.31, "pair": 0.8835199922323227, "sign": 0.4224504828453064, "keep": 7.23925921972841, "leak": 0.05066790886485251, "replacement": 4.727041661739349}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 5.316908836364746, "loss_scale": 4.0, "retries": 0, "seconds": 74.7, "max_memory_gb": 28.31, "pair": 0.9373003393411636, "sign": 0.5645432993769646, "keep": 2.671745039988309, "leak": 0.07095145968196448, "replacement": 2.5731823444366455}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 7.44608736038208, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 28.31, "pair": 0.8549740314483643, "sign": 0.6733440477401018, "keep": 4.702118143439293, "leak": 0.0760664144763723, "replacement": 3.5617216527462006}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 8.79843521118164, "loss_scale": 4.0, "retries": 0, "seconds": 77.6, "max_memory_gb": 28.31, "pair": 0.9279198050498962, "sign": 0.7350089699029922, "keep": 7.48223090171814, "leak": 0.07041453033861345, "replacement": 5.054143339395523}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 12.110786437988281, "loss_scale": 4.0, "retries": 0, "seconds": 79.1, "max_memory_gb": 28.31, "pair": 0.7972243279218674, "sign": 0.60348716750741, "keep": 7.74716754257679, "leak": 0.12135211029089987, "replacement": 5.002889811992645}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 9.80186653137207, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 28.31, "pair": 0.844013124704361, "sign": 0.41380082443356514, "keep": 4.709454322699457, "leak": 0.08278460911242291, "replacement": 3.426336884498596}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 19.526823043823242, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 28.31, "pair": 0.6257463693618774, "sign": 0.15947477333247662, "keep": 9.068302392959595, "leak": 0.1815615687519312, "replacement": 5.285025238990784}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 6.110527038574219, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 28.31, "pair": 0.9655852764844894, "sign": 0.7901964224874973, "keep": 2.3593820091336966, "leak": 0.022834991119452752, "replacement": 2.546083301305771}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 10.435531616210938, "loss_scale": 4.0, "retries": 0, "seconds": 84.8, "max_memory_gb": 28.31, "pair": 0.8633890897035599, "sign": 0.26485046558082104, "keep": 4.51968799775932, "leak": 0.16254023421788588, "replacement": 3.2962935864925385}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 5.365446090698242, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 28.31, "pair": 0.9560921043157578, "sign": 0.6488639712333679, "keep": 2.764400940388441, "leak": 0.005607108312688069, "replacement": 2.664126366376877}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 11.766348838806152, "loss_scale": 4.0, "retries": 0, "seconds": 87.7, "max_memory_gb": 28.31, "pair": 0.8328937143087387, "sign": 0.6171442661434412, "keep": 4.5465568620711565, "leak": 0.07412441039923579, "replacement": 3.4332753717899323}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 16.359935760498047, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 28.31, "pair": 0.8503867834806442, "sign": 0.6532436721026897, "keep": 6.202388972043991, "leak": 0.09309256065898808, "replacement": 4.301476359367371}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 6.462647914886475, "loss_scale": 4.0, "retries": 0, "seconds": 90.6, "max_memory_gb": 28.31, "pair": 0.9544018656015396, "sign": 0.6955659314990044, "keep": 2.518391467630863, "leak": 0.03940991705894703, "replacement": 2.5712329745292664}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 14.245582580566406, "loss_scale": 4.0, "retries": 0, "seconds": 92.1, "max_memory_gb": 28.31, "pair": 0.9532579779624939, "sign": 0.5771984085440636, "keep": 6.511389487423003, "leak": 0.052598995505832136, "replacement": 4.510701537132263}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 20.58387565612793, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 28.37, "pair": 0.7771196812391281, "sign": 0.4422048646956682, "keep": 6.237732566893101, "leak": 0.13227418216411024, "replacement": 4.150156855583191}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 19.97171974182129, "loss_scale": 4.0, "retries": 0, "seconds": 95.1, "max_memory_gb": 28.37, "pair": 0.923088937997818, "sign": 0.5652844682335854, "keep": 6.242365300655365, "leak": 0.06697325129061937, "replacement": 4.343657076358795}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 3.5378262996673584, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 28.37, "pair": 1.1147570312023163, "sign": 0.6703606322407722, "keep": 1.7618393530137837, "leak": 0.03847339870117139, "replacement": 2.340475380420685}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 16.963272094726562, "loss_scale": 4.0, "retries": 0, "seconds": 98.1, "max_memory_gb": 28.37, "pair": 0.9311744868755341, "sign": 0.39207719825208187, "keep": 4.842343777418137, "leak": 0.22833874000934884, "replacement": 3.6054696440696716}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 7.541445255279541, "loss_scale": 4.0, "retries": 0, "seconds": 99.6, "max_memory_gb": 28.37, "pair": 0.9647309482097626, "sign": 0.496286753565073, "keep": 2.0997156309895217, "leak": 0.03382507737114793, "replacement": 2.2711883783340454}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 14.613987922668457, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 28.37, "pair": 0.8811893314123154, "sign": 0.6739459335803986, "keep": 3.3208134612068534, "leak": 0.07590352813713253, "replacement": 2.897544890642166}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 10.25329875946045, "loss_scale": 4.0, "retries": 0, "seconds": 102.5, "max_memory_gb": 28.37, "pair": 1.1159453094005585, "sign": 0.6143811494112015, "keep": 3.005822015926242, "leak": 0.3001386791729601, "replacement": 3.001081496477127}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 17.216718673706055, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 28.37, "pair": 0.9714463353157043, "sign": 0.17833435907959938, "keep": 4.6677679708227515, "leak": 0.27742443114402704, "replacement": 3.4638533890247345}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 22.286243438720703, "loss_scale": 4.0, "retries": 0, "seconds": 105.4, "max_memory_gb": 28.37, "pair": 0.8828008770942688, "sign": 0.27820827066898346, "keep": 5.417152285575867, "leak": 0.211393466219306, "replacement": 3.7833295464515686}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 2.3297510147094727, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 28.37, "pair": 0.988838791847229, "sign": 0.7876889407634735, "keep": 0.42911104252561927, "leak": 0.022489242583105806, "replacement": 1.6028611660003662}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 11.168275833129883, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 28.37, "pair": 1.0671458691358566, "sign": 0.5053708935156465, "keep": 2.564120741561055, "leak": 0.08810874349933329, "replacement": 2.6239189505577087}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 5.025659084320068, "loss_scale": 4.0, "retries": 0, "seconds": 109.8, "max_memory_gb": 28.37, "pair": 0.9497493207454681, "sign": 0.8885030541568995, "keep": 2.566627651453018, "leak": 0.07997934365994297, "replacement": 2.697309583425522}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 11.138349533081055, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 28.37, "pair": 0.943517416715622, "sign": 0.4489050544798374, "keep": 2.210285283625126, "leak": 0.049741662689484656, "replacement": 2.2855480313301086}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 12.5155611038208, "loss_scale": 4.0, "retries": 0, "seconds": 112.7, "max_memory_gb": 28.37, "pair": 1.062734067440033, "sign": 0.517765874043107, "keep": 3.995447427034378, "leak": 0.2835976657224819, "replacement": 3.3902401328086853}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 5.532680988311768, "loss_scale": 4.0, "retries": 0, "seconds": 114.1, "max_memory_gb": 28.37, "pair": 1.0951288938522339, "sign": 0.7577540278434753, "keep": 2.1622280813753605, "leak": 0.3546471244189888, "replacement": 2.643781751394272}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 6.455410480499268, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 28.37, "pair": 0.9913413822650909, "sign": 0.5863281525671482, "keep": 1.1417318261228502, "leak": 0.0239617737242952, "replacement": 1.8613618314266205}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 11.08266830444336, "loss_scale": 4.0, "retries": 0, "seconds": 117.0, "max_memory_gb": 28.37, "pair": 1.2179457247257233, "sign": 0.42889972031116486, "keep": 1.784160296432674, "leak": 0.06939914176473394, "replacement": 2.3418256044387817}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 16.560224533081055, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 28.37, "pair": 0.8921767473220825, "sign": 0.4669102653861046, "keep": 2.9033939838409424, "leak": 0.5957418093457818, "replacement": 2.726264327764511}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 14.774846076965332, "loss_scale": 4.0, "retries": 0, "seconds": 119.8, "max_memory_gb": 28.37, "pair": 1.1758395433425903, "sign": 0.41806771233677864, "keep": 2.5323125272989273, "leak": 0.027268432325399772, "replacement": 2.6578467786312103}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 15.407487869262695, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 28.37, "pair": 0.9721937924623489, "sign": 0.44487715885043144, "keep": 2.3490410186350346, "leak": 0.601602675858885, "replacement": 2.5195535123348236}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 6.106161117553711, "loss_scale": 4.0, "retries": 0, "seconds": 122.8, "max_memory_gb": 28.37, "pair": 1.0420345813035965, "sign": 0.775519035756588, "keep": 1.10721190366894, "leak": 0.32737258228007704, "replacement": 2.06524321436882}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 5.624213218688965, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 28.37, "pair": 1.0678489357233047, "sign": 0.5867315828800201, "keep": 0.14577221870422363, "leak": 0.026043539233796764, "replacement": 1.440611720085144}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 15.074921607971191, "loss_scale": 4.0, "retries": 0, "seconds": 125.7, "max_memory_gb": 28.37, "pair": 1.0769954919815063, "sign": 0.6749863401055336, "keep": 2.4663716703653336, "leak": 0.05023262253962457, "replacement": 2.6602326035499573}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 15.702588081359863, "loss_scale": 4.0, "retries": 0, "seconds": 127.1, "max_memory_gb": 28.37, "pair": 0.9768049269914627, "sign": 0.46845003217458725, "keep": 1.8067618869245052, "leak": 0.5912908678874373, "replacement": 2.262233555316925}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 9.381375312805176, "loss_scale": 4.0, "retries": 0, "seconds": 128.6, "max_memory_gb": 28.37, "pair": 1.1561227291822433, "sign": 0.6378518119454384, "keep": 1.3802634067833424, "leak": 0.04426308919209987, "replacement": 2.176246076822281}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 6.296018600463867, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 28.37, "pair": 1.1368612349033356, "sign": 0.7031178697943687, "keep": 0.22903459891676903, "leak": 0.0473584871579078, "replacement": 1.6147770881652832}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 10.985401153564453, "loss_scale": 4.0, "retries": 0, "seconds": 131.4, "max_memory_gb": 28.37, "pair": 0.8929917216300964, "sign": 0.5917671173810959, "keep": 1.2370345955714583, "leak": 0.43639065977185965, "replacement": 1.9164902865886688}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 3.6315064430236816, "loss_scale": 4.0, "retries": 0, "seconds": 132.9, "max_memory_gb": 28.37, "pair": 0.9918826222419739, "sign": 0.9899481534957886, "keep": 0.6525968760251999, "leak": 0.022871337656397372, "replacement": 1.818873018026352}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 5.612170219421387, "loss_scale": 4.0, "retries": 0, "seconds": 134.4, "max_memory_gb": 28.37, "pair": 1.0547542870044708, "sign": 0.5568342283368111, "keep": 0.6405666349455714, "leak": 0.9313061884604394, "replacement": 1.8862813115119934}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 11.988249778747559, "loss_scale": 4.0, "retries": 0, "seconds": 135.7, "max_memory_gb": 28.37, "pair": 1.1144708544015884, "sign": 0.3403938114643097, "keep": 1.1272998377680779, "leak": 0.13851716835051775, "replacement": 1.882946938276291}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 15.110974311828613, "loss_scale": 4.0, "retries": 0, "seconds": 137.2, "max_memory_gb": 28.37, "pair": 0.8546710908412933, "sign": 0.39455367252230644, "keep": 1.5598140209913254, "leak": 0.27677516732364893, "replacement": 1.901048719882965}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 9.130464553833008, "loss_scale": 4.0, "retries": 0, "seconds": 138.7, "max_memory_gb": 28.37, "pair": 0.8504581898450851, "sign": 0.48224484361708164, "keep": 1.2989202812314034, "leak": 0.1274547764332965, "replacement": 1.772904485464096}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 12.330193519592285, "loss_scale": 4.0, "retries": 0, "seconds": 140.2, "max_memory_gb": 28.37, "pair": 1.0408172607421875, "sign": 0.46680130623281, "keep": 1.3434145748615265, "leak": 2.123078867800359, "replacement": 2.476694881916046}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 7.606297016143799, "loss_scale": 4.0, "retries": 0, "seconds": 141.6, "max_memory_gb": 28.37, "pair": 0.9540564566850662, "sign": 0.512284904718399, "keep": 0.6606472358107567, "leak": 0.24877967778593302, "replacement": 1.6027174293994904}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 8.467841148376465, "loss_scale": 4.0, "retries": 0, "seconds": 143.0, "max_memory_gb": 28.37, "pair": 0.8653485924005508, "sign": 0.5321787558495998, "keep": 0.6390587640926242, "leak": 0.30802788512664847, "replacement": 1.5279743373394012}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 5.6578216552734375, "loss_scale": 4.0, "retries": 0, "seconds": 144.5, "max_memory_gb": 28.37, "pair": 1.0315280854701996, "sign": 0.891327902674675, "keep": 0.356582997366786, "leak": 0.049141098745167255, "replacement": 1.6677687764167786}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 11.053279876708984, "loss_scale": 4.0, "retries": 0, "seconds": 146.0, "max_memory_gb": 28.37, "pair": 0.9185400158166885, "sign": 0.6625103950500488, "keep": 1.2019813060760498, "leak": 0.9717626213096082, "replacement": 2.0937265157699585}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 15.177087783813477, "loss_scale": 4.0, "retries": 0, "seconds": 147.4, "max_memory_gb": 28.37, "pair": 0.9333577305078506, "sign": 0.19265197403728962, "keep": 0.8349497839808464, "leak": 0.09968166111502796, "replacement": 1.4720790684223175}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 9.174373626708984, "loss_scale": 4.0, "retries": 0, "seconds": 148.9, "max_memory_gb": 28.37, "pair": 0.7286864668130875, "sign": 0.328602597117424, "keep": 1.3477341458201408, "leak": 0.6372322551906109, "replacement": 1.7261628806591034}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 8.336980819702148, "loss_scale": 4.0, "retries": 0, "seconds": 150.4, "max_memory_gb": 28.37, "pair": 0.7036226391792297, "sign": 0.35051702708005905, "keep": 0.5147724952548742, "leak": 1.0086488705128431, "replacement": 1.3884296119213104}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 8.689103126525879, "loss_scale": 4.0, "retries": 0, "seconds": 151.8, "max_memory_gb": 28.37, "pair": 0.8858087062835693, "sign": 0.8330366052687168, "keep": 0.5496240751817822, "leak": 0.07433482725173235, "replacement": 1.5957227647304535}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 8.106263160705566, "loss_scale": 4.0, "retries": 0, "seconds": 153.3, "max_memory_gb": 28.37, "pair": 0.7768567502498627, "sign": 0.4051705449819565, "keep": 0.6319517530500889, "leak": 1.8427422028034925, "replacement": 1.7561034858226776}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 6.615788459777832, "loss_scale": 4.0, "retries": 0, "seconds": 154.7, "max_memory_gb": 28.37, "pair": 0.7040611505508423, "sign": 0.4592324532568455, "keep": 0.6767781972885132, "leak": 1.9309251342201605, "replacement": 1.7547977268695831}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 5.144339561462402, "loss_scale": 4.0, "retries": 0, "seconds": 156.2, "max_memory_gb": 28.37, "pair": 0.9108491241931915, "sign": 0.7485629171133041, "keep": 0.3802998084574938, "leak": 0.045795324454957154, "replacement": 1.4867292940616608}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 3.981196641921997, "loss_scale": 4.0, "retries": 0, "seconds": 157.7, "max_memory_gb": 28.37, "pair": 0.8834050893783569, "sign": 0.6996964514255524, "keep": 0.3648503515869379, "leak": 1.362419652636163, "replacement": 1.756283462047577}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 7.042347431182861, "loss_scale": 4.0, "retries": 0, "seconds": 159.1, "max_memory_gb": 28.37, "pair": 0.7081726863980293, "sign": 0.47656952776014805, "keep": 0.7585049420595169, "leak": 1.593768632854335, "replacement": 1.7241520881652832}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 5.907708644866943, "loss_scale": 4.0, "retries": 0, "seconds": 160.5, "max_memory_gb": 28.37, "pair": 0.7681246437132359, "sign": 0.5243641771376133, "keep": 0.6062935963273048, "leak": 0.8072714455192909, "replacement": 1.535271406173706}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 7.660866737365723, "loss_scale": 4.0, "retries": 0, "seconds": 161.9, "max_memory_gb": 28.37, "pair": 0.5970887690782547, "sign": 0.4204434547573328, "keep": 0.37618338502943516, "leak": 0.03972418757621199, "replacement": 1.0053332149982452}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 5.743816375732422, "loss_scale": 4.0, "retries": 0, "seconds": 163.3, "max_memory_gb": 28.37, "pair": 0.697915330529213, "sign": 0.43632886931300163, "keep": 0.3985577542334795, "leak": 0.395101404748857, "replacement": 1.2141339778900146}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 7.164061546325684, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 28.37, "pair": 0.7112624868750572, "sign": 0.5147732272744179, "keep": 0.4042886421084404, "leak": 0.26610171736683697, "replacement": 1.237318828701973}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 9.828256607055664, "loss_scale": 4.0, "retries": 0, "seconds": 166.2, "max_memory_gb": 28.37, "pair": 0.6056149899959564, "sign": 0.49214974138885736, "keep": 1.0951697304844856, "leak": 1.017520054941997, "replacement": 1.6536547094583511}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 4.644096374511719, "loss_scale": 4.0, "retries": 0, "seconds": 167.7, "max_memory_gb": 28.37, "pair": 0.7507965639233589, "sign": 0.5649907328188419, "keep": 0.5101118013262749, "leak": 0.8727857887934078, "replacement": 1.5065442621707916}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 4.776440143585205, "loss_scale": 4.0, "retries": 0, "seconds": 169.1, "max_memory_gb": 28.37, "pair": 0.7490077018737793, "sign": 0.6245369650423527, "keep": 0.20865621487610042, "leak": 0.11678361904341727, "replacement": 1.1948001980781555}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 10.857047080993652, "loss_scale": 4.0, "retries": 0, "seconds": 170.4, "max_memory_gb": 28.37, "pair": 0.579217791557312, "sign": 0.38789328560233116, "keep": 0.8804319649934769, "leak": 0.7301255762577057, "replacement": 1.3959118276834488}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 5.747483253479004, "loss_scale": 4.0, "retries": 0, "seconds": 171.8, "max_memory_gb": 28.37, "pair": 0.8310734927654266, "sign": 0.5161832310259342, "keep": 0.34796499903313816, "leak": 0.15633056254591793, "replacement": 1.3022301644086838}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 4.864324569702148, "loss_scale": 4.0, "retries": 0, "seconds": 173.2, "max_memory_gb": 28.37, "pair": 0.8670882433652878, "sign": 0.792520746588707, "keep": 0.3393040783703327, "leak": 0.18137371633201838, "replacement": 1.478344053030014}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 5.145506381988525, "loss_scale": 4.0, "retries": 0, "seconds": 174.6, "max_memory_gb": 28.37, "pair": 0.5604753196239471, "sign": 0.4938345425762236, "keep": 0.8421669183298945, "leak": 0.8248657810036093, "replacement": 1.4346925541758537}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 4.248329162597656, "loss_scale": 4.0, "retries": 0, "seconds": 176.1, "max_memory_gb": 28.37, "pair": 0.7553316056728363, "sign": 0.6709622964262962, "keep": 0.602683583740145, "leak": 1.8804342211224139, "replacement": 1.862263098359108}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 7.697309494018555, "loss_scale": 4.0, "retries": 0, "seconds": 177.6, "max_memory_gb": 28.37, "pair": 0.6888048052787781, "sign": 0.6061166524887085, "keep": 0.27108308719471097, "leak": 0.36452262991224416, "replacement": 1.2185353338718414}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 3.136570930480957, "loss_scale": 4.0, "retries": 0, "seconds": 179.0, "max_memory_gb": 28.37, "pair": 0.723976943641901, "sign": 0.7564045367762446, "keep": 0.8270012894645333, "leak": 0.9050218230113387, "replacement": 1.7419352680444717}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 3.5897140502929688, "loss_scale": 4.0, "retries": 0, "seconds": 180.3, "max_memory_gb": 28.37, "pair": 0.8957079350948334, "sign": 1.0059861838817596, "keep": 0.5511905383318663, "leak": 0.02323243688390164, "replacement": 1.6801044046878815}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 4.3902363777160645, "loss_scale": 4.0, "retries": 0, "seconds": 181.7, "max_memory_gb": 28.37, "pair": 0.8222803920507431, "sign": 0.7768445312976837, "keep": 0.2910741874948144, "leak": 0.07394256559200585, "replacement": 1.3747253715991974}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 4.641404151916504, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 28.37, "pair": 0.6463063582777977, "sign": 0.6054981765337288, "keep": 0.6973649337887764, "leak": 0.32820528091269807, "replacement": 1.3797892928123474}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 4.7890753746032715, "loss_scale": 4.0, "retries": 0, "seconds": 184.5, "max_memory_gb": 28.37, "pair": 0.5791436955332756, "sign": 0.3330869833007455, "keep": 0.13357738102786243, "leak": 0.33061130717396736, "replacement": 0.8951286971569061}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 4.378943920135498, "loss_scale": 4.0, "retries": 0, "seconds": 185.9, "max_memory_gb": 28.37, "pair": 0.4690816216170788, "sign": 0.28581067360937595, "keep": 0.2886233264580369, "leak": 0.9791077524423599, "replacement": 1.0010755360126495}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 4.752859592437744, "loss_scale": 4.0, "retries": 0, "seconds": 187.4, "max_memory_gb": 28.37, "pair": 0.35690736025571823, "sign": 0.19022171385586262, "keep": 0.3063071519136429, "leak": 2.067562637384981, "replacement": 1.1220624297857285}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 5.418487071990967, "loss_scale": 4.0, "retries": 0, "seconds": 188.9, "max_memory_gb": 28.37, "pair": 0.6529441177845001, "sign": 0.834193080663681, "keep": 0.6452407483011484, "leak": 0.8119224179536104, "replacement": 1.595641627907753}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 4.005179405212402, "loss_scale": 4.0, "retries": 0, "seconds": 190.3, "max_memory_gb": 28.37, "pair": 0.6764478906989098, "sign": 0.6605062484741211, "keep": 0.5882201120257378, "leak": 0.28139239468146116, "replacement": 1.3711591809988022}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 5.614312648773193, "loss_scale": 4.0, "retries": 0, "seconds": 191.8, "max_memory_gb": 28.37, "pair": 0.3702457994222641, "sign": 0.1125751193612814, "keep": 0.850623887963593, "leak": 1.4828805401921272, "replacement": 1.2225654423236847}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 5.006680488586426, "loss_scale": 4.0, "retries": 0, "seconds": 193.2, "max_memory_gb": 28.37, "pair": 0.41717177256941795, "sign": 0.3881170470267534, "keep": 0.5529682710766792, "leak": 2.4784541111439466, "replacement": 1.5073279589414597}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 5.184145450592041, "loss_scale": 4.0, "retries": 0, "seconds": 194.7, "max_memory_gb": 28.37, "pair": 0.7154359519481659, "sign": 0.6431562900543213, "keep": 0.33533813850954175, "leak": 0.21516123600304127, "replacement": 1.2584735006093979}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 3.54247784614563, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 28.37, "pair": 0.4037248119711876, "sign": 0.38316295575350523, "keep": 0.2829528823494911, "leak": 3.2561163706704974, "replacement": 1.550811842083931}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 5.7335124015808105, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 28.37, "pair": 0.5402147397398949, "sign": 0.4449589643627405, "keep": 0.4867118541151285, "leak": 0.21067606942784778, "replacement": 1.0587191358208656}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 7.449153423309326, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 28.37, "pair": 0.5239425003528595, "sign": 0.4307292839512229, "keep": 0.26367769949138165, "leak": 2.6376352440565825, "replacement": 1.5305548310279846}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 7.345263481140137, "loss_scale": 4.0, "retries": 0, "seconds": 200.6, "max_memory_gb": 28.37, "pair": 0.4512815363705158, "sign": 0.25360027328133583, "keep": 0.1845940495841205, "leak": 0.20658119162544608, "replacement": 0.7220240160822868}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 9.8827543258667, "loss_scale": 4.0, "retries": 0, "seconds": 202.0, "max_memory_gb": 28.37, "pair": 0.4987157806754112, "sign": 0.548891281709075, "keep": 0.4806877374649048, "leak": 1.0565679110586643, "replacement": 1.277647241950035}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 4.465818405151367, "loss_scale": 4.0, "retries": 0, "seconds": 203.4, "max_memory_gb": 28.37, "pair": 0.47238457947969437, "sign": 0.5003903079777956, "keep": 0.17401392944157124, "leak": 0.809846930205822, "replacement": 1.0120484456419945}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 3.1517763137817383, "loss_scale": 4.0, "retries": 0, "seconds": 204.9, "max_memory_gb": 28.37, "pair": 0.5072402879595757, "sign": 0.5148539505898952, "keep": 0.45215779077261686, "leak": 1.2493646182119846, "replacement": 1.3030873239040375}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 3.175367832183838, "loss_scale": 4.0, "retries": 0, "seconds": 206.4, "max_memory_gb": 28.37, "pair": 0.4488692879676819, "sign": 0.37630844209343195, "keep": 0.2447611801326275, "leak": 0.42271711135981604, "replacement": 0.8650833666324615}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 4.228688716888428, "loss_scale": 4.0, "retries": 0, "seconds": 207.8, "max_memory_gb": 28.37, "pair": 0.7865723371505737, "sign": 0.9399761408567429, "keep": 0.23336386180017143, "leak": 0.25849513756111264, "replacement": 1.4378661066293716}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 4.082367897033691, "loss_scale": 4.0, "retries": 0, "seconds": 209.3, "max_memory_gb": 28.37, "pair": 0.4332694336771965, "sign": 0.33444237895309925, "keep": 0.2659348286688328, "leak": 1.7757867397740483, "replacement": 1.1774047315120697}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 6.2709832191467285, "loss_scale": 4.0, "retries": 0, "seconds": 210.7, "max_memory_gb": 28.37, "pair": 0.5381206423044205, "sign": 0.29237081250175834, "keep": 0.08482456114143133, "leak": 0.8658636659383774, "replacement": 0.9431842118501663}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 3.125648021697998, "loss_scale": 4.0, "retries": 0, "seconds": 212.1, "max_memory_gb": 28.37, "pair": 0.5428645629435778, "sign": 0.5345993535593152, "keep": 0.23795153014361858, "leak": 1.274527987698093, "replacement": 1.247771978378296}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 4.621383190155029, "loss_scale": 4.0, "retries": 0, "seconds": 213.6, "max_memory_gb": 28.37, "pair": 0.5286695510149002, "sign": 0.5050213858485222, "keep": 0.4798211334273219, "leak": 1.2924079662188888, "replacement": 1.3441928178071976}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 6.524885654449463, "loss_scale": 4.0, "retries": 0, "seconds": 215.0, "max_memory_gb": 28.37, "pair": 0.518331877887249, "sign": 0.27712012827396393, "keep": 0.06122398842126131, "leak": 0.3337036914890632, "replacement": 0.7709298580884933}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 5.391860485076904, "loss_scale": 4.0, "retries": 0, "seconds": 216.5, "max_memory_gb": 28.37, "pair": 0.54136161506176, "sign": 0.4415066298097372, "keep": 0.1605128338560462, "leak": 0.18497522082179785, "replacement": 0.8886151909828186}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 5.831388473510742, "loss_scale": 4.0, "retries": 0, "seconds": 217.9, "max_memory_gb": 28.37, "pair": 0.4061109125614166, "sign": 0.33206631802022457, "keep": 0.4589902423322201, "leak": 2.2450577234849334, "replacement": 1.3629036247730255}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 5.7932963371276855, "loss_scale": 4.0, "retries": 0, "seconds": 219.3, "max_memory_gb": 28.37, "pair": 0.4010007679462433, "sign": 0.3595801815390587, "keep": 0.33017016761004925, "leak": 0.516539566218853, "replacement": 0.8750108480453491}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 5.066405773162842, "loss_scale": 4.0, "retries": 0, "seconds": 220.7, "max_memory_gb": 28.37, "pair": 0.3030494190752506, "sign": 0.19920210260897875, "keep": 0.16656242311000824, "leak": 2.795734018087387, "replacement": 1.1848651617765427}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 4.86459493637085, "loss_scale": 4.0, "retries": 0, "seconds": 222.1, "max_memory_gb": 28.37, "pair": 0.3624921888113022, "sign": 0.2346968613564968, "keep": 0.1055678278207779, "leak": 1.7160264924168587, "replacement": 0.9616311639547348}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 6.523874282836914, "loss_scale": 4.0, "retries": 0, "seconds": 223.5, "max_memory_gb": 28.37, "pair": 0.2817970737814903, "sign": 0.18283217772841454, "keep": 0.7958383746445179, "leak": 1.0436024516820908, "replacement": 1.0320329517126083}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 4.463070869445801, "loss_scale": 4.0, "retries": 0, "seconds": 224.9, "max_memory_gb": 28.37, "pair": 0.49815602600574493, "sign": 0.5006791837513447, "keep": 1.1978049741592258, "leak": 1.9477633275091648, "replacement": 1.8343389630317688}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 3.7470779418945312, "loss_scale": 4.0, "retries": 0, "seconds": 226.3, "max_memory_gb": 28.37, "pair": 0.5522435382008553, "sign": 0.4972009602934122, "keep": 0.252041946281679, "leak": 0.42666900297626853, "replacement": 1.0335322171449661}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 4.677918910980225, "loss_scale": 4.0, "retries": 0, "seconds": 227.8, "max_memory_gb": 28.37, "pair": 0.737963117659092, "sign": 1.1165216490626335, "keep": 0.2267762660048902, "leak": 0.2226626550545916, "replacement": 1.4652777016162872}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 4.544271469116211, "loss_scale": 4.0, "retries": 0, "seconds": 229.2, "max_memory_gb": 28.37, "pair": 0.598640650510788, "sign": 0.6400938853621483, "keep": 0.4150947667658329, "leak": 0.3693064791150391, "replacement": 1.2185616195201874}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 4.739049434661865, "loss_scale": 4.0, "retries": 0, "seconds": 230.7, "max_memory_gb": 28.37, "pair": 0.27703687362372875, "sign": 0.16500114183872938, "keep": 0.09370364714413881, "leak": 2.353004202246666, "replacement": 0.9946403205394745}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 10.680992126464844, "loss_scale": 4.0, "retries": 0, "seconds": 232.2, "max_memory_gb": 28.37, "pair": 0.685407355427742, "sign": 0.3540457640774548, "keep": 0.27455572597682476, "leak": 1.341021133652248, "replacement": 1.3349633812904358}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 3.30143666267395, "loss_scale": 4.0, "retries": 0, "seconds": 233.6, "max_memory_gb": 28.37, "pair": 0.6009110100567341, "sign": 0.5805776854977012, "keep": 0.40272400225512683, "leak": 0.6124305557459593, "replacement": 1.2456694692373276}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 3.9432036876678467, "loss_scale": 4.0, "retries": 0, "seconds": 235.0, "max_memory_gb": 28.37, "pair": 0.574364660307765, "sign": 0.5757326297461987, "keep": 0.11112676979973912, "leak": 0.2839058329409454, "replacement": 0.9887707978487015}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 3.320655107498169, "loss_scale": 4.0, "retries": 0, "seconds": 236.4, "max_memory_gb": 28.37, "pair": 0.36381416767835617, "sign": 0.2741980985738337, "keep": 0.08671162370592356, "leak": 0.8356683179736137, "replacement": 0.7531861215829849}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 4.930375099182129, "loss_scale": 4.0, "retries": 0, "seconds": 237.8, "max_memory_gb": 28.37, "pair": 0.6290194056928158, "sign": 0.8391050356440246, "keep": 0.3828345462679863, "leak": 0.9189795014099218, "replacement": 1.469734013080597}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 4.800793647766113, "loss_scale": 4.0, "retries": 0, "seconds": 239.3, "max_memory_gb": 28.37, "pair": 0.4535849317908287, "sign": 0.525258420035243, "keep": 0.18147198297083378, "leak": 1.9764194209128618, "replacement": 1.301054984331131}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 5.839531421661377, "loss_scale": 4.0, "retries": 0, "seconds": 240.8, "max_memory_gb": 28.37, "pair": 0.415402315557003, "sign": 0.4473109245300293, "keep": 0.2214446454308927, "leak": 1.479854752193205, "replacement": 1.1197437793016434}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 5.068482398986816, "loss_scale": 4.0, "retries": 0, "seconds": 242.3, "max_memory_gb": 28.37, "pair": 0.39783274102956057, "sign": 0.3818244063295424, "keep": 0.553196020424366, "leak": 1.4563261616567615, "replacement": 1.2294244542717934}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 5.434572219848633, "loss_scale": 4.0, "retries": 0, "seconds": 243.7, "max_memory_gb": 28.37, "pair": 0.4478835165500641, "sign": 0.4127387683838606, "keep": 0.14265488181263208, "leak": 1.086575798690319, "replacement": 0.9972243010997772}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 6.11805534362793, "loss_scale": 4.0, "retries": 0, "seconds": 245.2, "max_memory_gb": 28.37, "pair": 0.38257088139653206, "sign": 0.4209011010825634, "keep": 0.8438295498490334, "leak": 1.8407092541456223, "replacement": 1.4751135110855103}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 6.169731616973877, "loss_scale": 4.0, "retries": 0, "seconds": 246.6, "max_memory_gb": 28.37, "pair": 0.2813180312514305, "sign": 0.08861045353114605, "keep": 0.16519320290535688, "leak": 1.5090473480522633, "replacement": 0.7854816988110542}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 7.838569164276123, "loss_scale": 4.0, "retries": 0, "seconds": 248.1, "max_memory_gb": 28.37, "pair": 0.4476832374930382, "sign": 0.38884202018380165, "keep": 0.1265370100736618, "leak": 1.9998406246304512, "replacement": 1.2053329646587372}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 9.330373764038086, "loss_scale": 4.0, "retries": 0, "seconds": 249.6, "max_memory_gb": 28.37, "pair": 0.3904432952404022, "sign": 0.07791568757966161, "keep": 0.18973711133003235, "leak": 4.556745816022158, "replacement": 1.6634561419487}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 8.297991752624512, "loss_scale": 4.0, "retries": 0, "seconds": 251.0, "max_memory_gb": 28.37, "pair": 0.24621276184916496, "sign": 0.37306540831923485, "keep": 0.33154806680977345, "leak": 2.5551589727401733, "replacement": 1.2373092621564865}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 7.991055011749268, "loss_scale": 4.0, "retries": 0, "seconds": 252.4, "max_memory_gb": 28.37, "pair": 0.40546615421772003, "sign": 0.31251533702015877, "keep": 0.253039738163352, "leak": 1.0250388979911804, "replacement": 0.9445034265518188}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 5.226135730743408, "loss_scale": 4.0, "retries": 0, "seconds": 253.9, "max_memory_gb": 28.37, "pair": 0.4673825837671757, "sign": 0.5092608444392681, "keep": 0.252448208630085, "leak": 1.6284963702782989, "replacement": 1.2553612142801285}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 7.616530895233154, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 28.37, "pair": 0.29641886427998543, "sign": 0.12640626076608896, "keep": 0.17589660361409187, "leak": 1.9340675808489323, "replacement": 0.9310871884226799}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 6.771091461181641, "loss_scale": 4.0, "retries": 0, "seconds": 256.7, "max_memory_gb": 28.37, "pair": 0.48060116171836853, "sign": 0.3201225381344557, "keep": 0.28329819324426353, "leak": 0.8336734771728516, "replacement": 0.990729883313179}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 3.9901041984558105, "loss_scale": 4.0, "retries": 0, "seconds": 258.1, "max_memory_gb": 28.37, "pair": 0.5512789189815521, "sign": 0.5306862331926823, "keep": 0.21984605956822634, "leak": 1.4582030882593244, "replacement": 1.2910958528518677}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 3.975703716278076, "loss_scale": 4.0, "retries": 0, "seconds": 259.5, "max_memory_gb": 28.37, "pair": 0.648332342505455, "sign": 0.7402705317363143, "keep": 0.5905326653737575, "leak": 0.9782891708891839, "replacement": 1.5583062171936035}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 10.072260856628418, "loss_scale": 4.0, "retries": 0, "seconds": 260.8, "max_memory_gb": 28.37, "pair": 0.7080506086349487, "sign": 0.6264525800943375, "keep": 0.8792223734781146, "leak": 0.4330941718071699, "replacement": 1.5691616609692574}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 6.793501853942871, "loss_scale": 4.0, "retries": 0, "seconds": 262.3, "max_memory_gb": 28.37, "pair": 0.21886850520968437, "sign": 0.10270883236080408, "keep": 0.12540677096694708, "leak": 2.8836709885363234, "replacement": 1.053844042122364}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 4.9153313636779785, "loss_scale": 4.0, "retries": 0, "seconds": 263.7, "max_memory_gb": 28.37, "pair": 0.5942527391016483, "sign": 0.5861824443563819, "keep": 0.46062790090218186, "leak": 0.8713158098980784, "replacement": 1.3354868590831757}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 4.79774284362793, "loss_scale": 4.0, "retries": 0, "seconds": 265.1, "max_memory_gb": 28.37, "pair": 0.22591155767440796, "sign": 0.12445098115131259, "keep": 0.028069284569937736, "leak": 1.3770036101341248, "replacement": 0.6464225947856903}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 8.38791561126709, "loss_scale": 4.0, "retries": 0, "seconds": 266.5, "max_memory_gb": 28.37, "pair": 0.4561934098601341, "sign": 0.35621856432408094, "keep": 0.21710483776405454, "leak": 0.8867325615137815, "replacement": 0.9645382463932037}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 9.403820037841797, "loss_scale": 4.0, "retries": 0, "seconds": 268.0, "max_memory_gb": 28.37, "pair": 0.3372543714940548, "sign": 0.24477084400132298, "keep": 0.05420290119946003, "leak": 3.7156813219189644, "replacement": 1.4156616032123566}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 3.7118513584136963, "loss_scale": 4.0, "retries": 0, "seconds": 269.3, "max_memory_gb": 28.37, "pair": 0.3617665283381939, "sign": 0.2255651392042637, "keep": 0.11599834228400141, "leak": 0.6699912315234542, "replacement": 0.7000460773706436}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 4.742480278015137, "loss_scale": 4.0, "retries": 0, "seconds": 270.8, "max_memory_gb": 28.37, "pair": 0.3358227536082268, "sign": 0.06830672174692154, "keep": 0.13254398107528687, "leak": 1.7530948693165556, "replacement": 0.8745218366384506}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 5.017386436462402, "loss_scale": 4.0, "retries": 0, "seconds": 272.2, "max_memory_gb": 28.37, "pair": 0.5135142058134079, "sign": 0.5115519352257252, "keep": 0.2424763438757509, "leak": 0.3757174756610766, "replacement": 0.9844576865434647}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 5.736323833465576, "loss_scale": 4.0, "retries": 0, "seconds": 273.6, "max_memory_gb": 28.37, "pair": 0.34611982107162476, "sign": 0.2238410823047161, "keep": 0.10087718814611435, "leak": 1.4767992397537455, "replacement": 0.8776787593960762}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 4.511994361877441, "loss_scale": 4.0, "retries": 0, "seconds": 275.0, "max_memory_gb": 28.37, "pair": 0.2481253333389759, "sign": 0.1903011528775096, "keep": 0.05332311987876892, "leak": 1.5964491814374924, "replacement": 0.7690497636795044}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 4.958842754364014, "loss_scale": 4.0, "retries": 0, "seconds": 276.3, "max_memory_gb": 28.37, "pair": 0.5776583477854729, "sign": 0.6279083220288157, "keep": 0.715209512040019, "leak": 0.5288543999195099, "replacement": 1.3814308494329453}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 4.3271379470825195, "loss_scale": 4.0, "retries": 0, "seconds": 277.7, "max_memory_gb": 28.37, "pair": 0.5973517745733261, "sign": 0.4796930570155382, "keep": 0.3482802421785891, "leak": 0.5502458377022776, "replacement": 1.1488998234272003}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 5.495176792144775, "loss_scale": 4.0, "retries": 0, "seconds": 279.2, "max_memory_gb": 28.37, "pair": 0.35882798209786415, "sign": 0.16392286121845245, "keep": 0.10787041811272502, "leak": 0.3794635608792305, "replacement": 0.5895905047655106}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 10.78864574432373, "loss_scale": 4.0, "retries": 0, "seconds": 280.6, "max_memory_gb": 28.37, "pair": 0.7034853249788284, "sign": 0.8237708397209644, "keep": 0.9338393567595631, "leak": 0.9409442991018295, "replacement": 1.8175264596939087}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 8.223188400268555, "loss_scale": 4.0, "retries": 0, "seconds": 282.2, "max_memory_gb": 28.37, "pair": 0.5831897258758545, "sign": 0.3800864741206169, "keep": 0.25006465055048466, "leak": 2.222648180089891, "replacement": 1.4539273381233215}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 3.6813836097717285, "loss_scale": 4.0, "retries": 0, "seconds": 283.7, "max_memory_gb": 28.37, "pair": 0.23669717460870743, "sign": 0.10015294654294848, "keep": 0.060826333588920534, "leak": 0.7609100267291069, "replacement": 0.5074142962694168}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 5.563685417175293, "loss_scale": 4.0, "retries": 0, "seconds": 285.1, "max_memory_gb": 28.37, "pair": 0.5552707388997078, "sign": 0.5240435320883989, "keep": 0.15690070390701294, "leak": 0.4856960829347372, "replacement": 1.017166867852211}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 4.1245503425598145, "loss_scale": 4.0, "retries": 0, "seconds": 286.4, "max_memory_gb": 28.37, "pair": 0.5710383951663971, "sign": 0.5923140719532967, "keep": 0.24655587784945965, "leak": 0.9687577579170465, "replacement": 1.2326628118753433}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 7.924482345581055, "loss_scale": 4.0, "retries": 0, "seconds": 287.9, "max_memory_gb": 28.37, "pair": 0.2650460824370384, "sign": 0.1289368811994791, "keep": 0.15777742979116738, "leak": 2.4827843010425568, "replacement": 1.0290993005037308}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 4.146513938903809, "loss_scale": 4.0, "retries": 0, "seconds": 289.3, "max_memory_gb": 28.37, "pair": 0.546274445950985, "sign": 0.6196282524615526, "keep": 0.4205857226625085, "leak": 0.538646197412163, "replacement": 1.2010429799556732}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 4.595860004425049, "loss_scale": 4.0, "retries": 0, "seconds": 290.7, "max_memory_gb": 28.37, "pair": 0.5279606394469738, "sign": 0.6303436774760485, "keep": 0.2480868622660637, "leak": 0.7842610403895378, "replacement": 1.1632412001490593}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 6.7563958168029785, "loss_scale": 4.0, "retries": 0, "seconds": 292.0, "max_memory_gb": 28.37, "pair": 0.4542549401521683, "sign": 0.32808955758810043, "keep": 0.018171058618463576, "leak": 0.5940102711319923, "replacement": 0.7758878469467163}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 4.9545369148254395, "loss_scale": 4.0, "retries": 0, "seconds": 293.5, "max_memory_gb": 28.37, "pair": 0.49368905276060104, "sign": 0.4597696904093027, "keep": 0.26216272776946425, "leak": 1.2126049024591339, "replacement": 1.157806470990181}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 6.091945171356201, "loss_scale": 4.0, "retries": 0, "seconds": 294.9, "max_memory_gb": 28.37, "pair": 0.48040133714675903, "sign": 0.5627340115606785, "keep": 0.24551920965313911, "leak": 1.3973497226834297, "replacement": 1.2338653951883316}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 7.203954219818115, "loss_scale": 4.0, "retries": 0, "seconds": 296.4, "max_memory_gb": 28.37, "pair": 0.2706631124019623, "sign": 0.30877041444182396, "keep": 0.12675643974216655, "leak": 1.729524977505207, "replacement": 0.9208077788352966}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 6.154674530029297, "loss_scale": 4.0, "retries": 0, "seconds": 297.9, "max_memory_gb": 28.37, "pair": 0.3857118524610996, "sign": 0.389792338013649, "keep": 0.1328056287020445, "leak": 2.308106532524107, "replacement": 1.2240374982357025}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 9.391688346862793, "loss_scale": 4.0, "retries": 0, "seconds": 299.4, "max_memory_gb": 28.37, "pair": 0.9188861735165119, "sign": 1.0202591679990292, "keep": 1.4949223657604307, "leak": 0.29632644042385436, "replacement": 2.2505586743354797}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 4.969397068023682, "loss_scale": 4.0, "retries": 0, "seconds": 300.9, "max_memory_gb": 28.37, "pair": 0.2997829131782055, "sign": 0.08720983657985926, "keep": 0.04880082979798317, "leak": 1.6567281307652593, "replacement": 0.7819702848792076}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 4.717320442199707, "loss_scale": 4.0, "retries": 0, "seconds": 302.4, "max_memory_gb": 28.37, "pair": 0.5511952433735132, "sign": 0.586248379200697, "keep": 0.3901429884135723, "leak": 1.069607914370863, "replacement": 1.3067929074168205}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 4.093661785125732, "loss_scale": 4.0, "retries": 0, "seconds": 303.9, "max_memory_gb": 28.37, "pair": 0.39527197927236557, "sign": 0.46596800442785025, "keep": 0.5607371403602883, "leak": 2.1039623096585274, "replacement": 1.4346150904893875}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 5.897989749908447, "loss_scale": 4.0, "retries": 0, "seconds": 305.2, "max_memory_gb": 28.37, "pair": 0.5841905754059553, "sign": 0.557175713358447, "keep": 0.3373924195766449, "leak": 0.719905948091764, "replacement": 1.2114511281251907}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 6.907708644866943, "loss_scale": 4.0, "retries": 0, "seconds": 306.7, "max_memory_gb": 28.37, "pair": 0.4945916198194027, "sign": 0.6093033952638507, "keep": 0.5321023166179657, "leak": 0.4840774276526645, "replacement": 1.1863138303160667}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 7.4113969802856445, "loss_scale": 4.0, "retries": 0, "seconds": 308.1, "max_memory_gb": 28.37, "pair": 0.5057410374283791, "sign": 0.47524852864444256, "keep": 0.863289772067219, "leak": 1.8753160312771797, "replacement": 1.64383926987648}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 4.337426662445068, "loss_scale": 4.0, "retries": 0, "seconds": 309.6, "max_memory_gb": 28.37, "pair": 0.6135958731174469, "sign": 0.6210207529366016, "keep": 0.5409692309331149, "leak": 0.35873313673073426, "replacement": 1.284274160861969}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 8.435514450073242, "loss_scale": 4.0, "retries": 0, "seconds": 311.1, "max_memory_gb": 28.37, "pair": 0.7561970576643944, "sign": 0.9396804571151733, "keep": 1.3030601993668824, "leak": 0.5317900404334068, "replacement": 2.0105149149894714}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 3.4153904914855957, "loss_scale": 4.0, "retries": 0, "seconds": 312.6, "max_memory_gb": 28.37, "pair": 0.3122050426900387, "sign": 0.2970293965190649, "keep": 0.30262327566742897, "leak": 0.8144367160275578, "replacement": 0.8156405836343765}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 9.783738136291504, "loss_scale": 4.0, "retries": 0, "seconds": 314.0, "max_memory_gb": 28.37, "pair": 0.2538519985973835, "sign": 0.21448253002017736, "keep": 0.06345473951660097, "leak": 2.083245113492012, "replacement": 0.9136319011449814}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 4.484867572784424, "loss_scale": 4.0, "retries": 0, "seconds": 315.5, "max_memory_gb": 28.37, "pair": 0.8053535372018814, "sign": 1.0128510892391205, "keep": 0.3547830779571086, "leak": 0.14084949088282883, "replacement": 1.5243829488754272}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 5.858963489532471, "loss_scale": 4.0, "retries": 0, "seconds": 316.9, "max_memory_gb": 28.37, "pair": 0.469449445605278, "sign": 0.5407277531921864, "keep": 0.17342666583135724, "leak": 0.5931502431631088, "replacement": 0.9748142212629318}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 4.109893798828125, "loss_scale": 4.0, "retries": 0, "seconds": 318.4, "max_memory_gb": 28.37, "pair": 0.31511469930410385, "sign": 0.41223520040512085, "keep": 0.20566506165778264, "leak": 0.9094764143228531, "replacement": 0.8514339476823807}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 3.4327759742736816, "loss_scale": 4.0, "retries": 0, "seconds": 319.8, "max_memory_gb": 28.37, "pair": 0.30437917076051235, "sign": 0.28574938559904695, "keep": 0.1209590844810009, "leak": 1.3484811009839177, "replacement": 0.8448536768555641}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 6.808250904083252, "loss_scale": 4.0, "retries": 0, "seconds": 321.3, "max_memory_gb": 28.37, "pair": 0.3523213490843773, "sign": 0.20944377034902573, "keep": 0.3950484739616513, "leak": 1.5717144524678588, "replacement": 1.0474960878491402}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 5.292453289031982, "loss_scale": 4.0, "retries": 0, "seconds": 322.8, "max_memory_gb": 28.37, "pair": 0.49933674558997154, "sign": 0.5977616086602211, "keep": 0.27748848526971415, "leak": 0.3404412309755571, "replacement": 1.0220720767974854}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 6.282459735870361, "loss_scale": 4.0, "retries": 0, "seconds": 324.3, "max_memory_gb": 28.37, "pair": 0.3265173379331827, "sign": 0.3690245393663645, "keep": 0.11573054600739852, "leak": 1.1926554404199123, "replacement": 0.8670587316155434}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 7.1861090660095215, "loss_scale": 4.0, "retries": 0, "seconds": 325.7, "max_memory_gb": 28.37, "pair": 0.5409712288528681, "sign": 0.5749579668045044, "keep": 0.13670727610588074, "leak": 1.225504853995517, "replacement": 1.20318004488945}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 6.036001205444336, "loss_scale": 4.0, "retries": 0, "seconds": 327.1, "max_memory_gb": 28.37, "pair": 0.3220694065093994, "sign": 0.3036532513797283, "keep": 0.26699700113385916, "leak": 0.8093028217554092, "replacement": 0.8097202628850937}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 4.138459205627441, "loss_scale": 4.0, "retries": 0, "seconds": 328.5, "max_memory_gb": 28.37, "pair": 0.36429034918546677, "sign": 0.18331595230847597, "keep": 0.03782674024114385, "leak": 0.9666877798736095, "replacement": 0.7165336161851883}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 5.285159111022949, "loss_scale": 4.0, "retries": 0, "seconds": 329.9, "max_memory_gb": 28.37, "pair": 0.2948478199541569, "sign": 0.36540013551712036, "keep": 0.04504365846514702, "leak": 0.9107643142342567, "replacement": 0.7277608215808868}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 5.46368408203125, "loss_scale": 4.0, "retries": 0, "seconds": 331.4, "max_memory_gb": 28.37, "pair": 0.5200063548982143, "sign": 0.6587456585839391, "keep": 0.4839286720380187, "leak": 0.7319680098444223, "replacement": 1.274335503578186}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 6.073727130889893, "loss_scale": 4.0, "retries": 0, "seconds": 332.8, "max_memory_gb": 28.37, "pair": 0.5866223536431789, "sign": 0.6855435110628605, "keep": 0.6307890668394975, "leak": 0.5380110112018883, "replacement": 1.379291445016861}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 4.922842025756836, "loss_scale": 4.0, "retries": 0, "seconds": 334.2, "max_memory_gb": 28.37, "pair": 0.6165857315063477, "sign": 0.7924806252121925, "keep": 0.6221732188714668, "leak": 0.46328649669885635, "replacement": 1.4397342503070831}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 7.126873970031738, "loss_scale": 4.0, "retries": 0, "seconds": 335.6, "max_memory_gb": 28.37, "pair": 0.3217037543654442, "sign": 0.4227573471143842, "keep": 0.06894293706864119, "leak": 0.7275045234709978, "replacement": 0.7494300454854965}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 5.59698486328125, "loss_scale": 4.0, "retries": 0, "seconds": 337.0, "max_memory_gb": 28.37, "pair": 0.30472956970334053, "sign": 0.16841888520866632, "keep": 0.21799901500344276, "leak": 1.4452353537781164, "replacement": 0.8592473566532135}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.392347812652588, "loss_scale": 4.0, "retries": 0, "seconds": 338.4, "max_memory_gb": 28.37, "pair": 0.41559878922998905, "sign": 0.44951371755450964, "keep": 0.29292169271502644, "leak": 1.4644641447812319, "replacement": 1.1529325544834137}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 5.892461776733398, "loss_scale": 4.0, "retries": 0, "seconds": 339.9, "max_memory_gb": 28.37, "pair": 0.5915030762553215, "sign": 0.6661906354129314, "keep": 0.7121813483536243, "leak": 0.5970121435821056, "replacement": 1.4299420267343521}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 5.296928882598877, "loss_scale": 4.0, "retries": 0, "seconds": 341.3, "max_memory_gb": 28.37, "pair": 0.4797920435667038, "sign": 0.4476396180689335, "keep": 0.16971809114329517, "leak": 0.2561220689676702, "replacement": 0.8525014221668243}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 6.216341972351074, "loss_scale": 4.0, "retries": 0, "seconds": 342.7, "max_memory_gb": 28.37, "pair": 0.5328200608491898, "sign": 0.4425325281918049, "keep": 0.37079996382817626, "leak": 0.34075320744886994, "replacement": 1.0246745944023132}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 8.199986457824707, "loss_scale": 4.0, "retries": 0, "seconds": 344.2, "max_memory_gb": 28.37, "pair": 0.3040697015821934, "sign": 0.17127122450619936, "keep": 0.08365784073248506, "leak": 1.2431832179427147, "replacement": 0.7423300296068192}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 6.0023627281188965, "loss_scale": 4.0, "retries": 0, "seconds": 345.7, "max_memory_gb": 28.37, "pair": 0.5002739373594522, "sign": 0.5513903833925724, "keep": 0.3350032309535891, "leak": 3.0154076013714075, "replacement": 1.6973226070404053}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 4.021387100219727, "loss_scale": 4.0, "retries": 0, "seconds": 347.1, "max_memory_gb": 28.37, "pair": 0.34139740094542503, "sign": 0.31065870448946953, "keep": 0.09849869587924331, "leak": 0.641532626003027, "replacement": 0.706359252333641}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 7.116018772125244, "loss_scale": 4.0, "retries": 0, "seconds": 348.6, "max_memory_gb": 28.37, "pair": 0.4822409823536873, "sign": 0.588603675365448, "keep": 0.17357548791915178, "leak": 0.28651331178843975, "replacement": 0.934958890080452}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 4.666807174682617, "loss_scale": 4.0, "retries": 0, "seconds": 350.0, "max_memory_gb": 28.37, "pair": 0.523166261613369, "sign": 0.5642849281430244, "keep": 0.21593541803304106, "leak": 0.4907285664230585, "replacement": 1.0359586104750633}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 10.648781776428223, "loss_scale": 4.0, "retries": 0, "seconds": 351.5, "max_memory_gb": 28.37, "pair": 0.23910238593816757, "sign": 0.22487120609730482, "keep": 0.07338312873616815, "leak": 2.9214323349297047, "replacement": 1.1185876429080963}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 12.072288513183594, "loss_scale": 4.0, "retries": 0, "seconds": 353.0, "max_memory_gb": 28.37, "pair": 0.6852559633553028, "sign": 0.891923657618463, "keep": 1.0183011488988996, "leak": 0.7897300105541945, "replacement": 1.8378009051084518}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 5.118011951446533, "loss_scale": 4.0, "retries": 0, "seconds": 354.4, "max_memory_gb": 28.37, "pair": 0.8098533377051353, "sign": 0.9759684465825558, "keep": 0.7427727389149368, "leak": 0.40087753790430725, "replacement": 1.769443392753601}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 8.314445495605469, "loss_scale": 4.0, "retries": 0, "seconds": 355.9, "max_memory_gb": 28.37, "pair": 0.15211952291429043, "sign": 0.15174829168245196, "keep": 0.04792264656862244, "leak": 1.7467328608036041, "replacement": 0.6886382102966309}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 6.8692450523376465, "loss_scale": 4.0, "retries": 0, "seconds": 357.4, "max_memory_gb": 28.37, "pair": 0.27302291989326477, "sign": 0.28967806324362755, "keep": 0.11436539917485788, "leak": 0.8180382400751114, "replacement": 0.6795541942119598}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 3.8959834575653076, "loss_scale": 4.0, "retries": 0, "seconds": 358.8, "max_memory_gb": 28.37, "pair": 0.22585074417293072, "sign": 0.13008584454655647, "keep": 0.12283281423151493, "leak": 0.58791592909256, "replacement": 0.49928905069828033}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 8.443492889404297, "loss_scale": 4.0, "retries": 0, "seconds": 360.3, "max_memory_gb": 28.37, "pair": 0.43395455181598663, "sign": 0.6741192936897278, "keep": 0.10149966180324554, "leak": 0.6135628196643665, "replacement": 0.9751547276973724}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 9.990182876586914, "loss_scale": 4.0, "retries": 0, "seconds": 361.8, "max_memory_gb": 28.37, "pair": 0.25003407150506973, "sign": 0.23453772254288197, "keep": 0.05861474073026329, "leak": 1.9127352088689804, "replacement": 0.8747940957546234}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 5.807175159454346, "loss_scale": 4.0, "retries": 0, "seconds": 363.3, "max_memory_gb": 28.37, "pair": 0.36935727298259735, "sign": 0.4146005380898714, "keep": 0.193703032535268, "leak": 2.4011567023117095, "replacement": 1.2737982273101807}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 4.732166767120361, "loss_scale": 4.0, "retries": 0, "seconds": 364.8, "max_memory_gb": 28.37, "pair": 0.21505618281662464, "sign": 0.13687241449952126, "keep": 0.03182921890402213, "leak": 0.9628989398479462, "replacement": 0.5401317328214645}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 10.25930118560791, "loss_scale": 4.0, "retries": 0, "seconds": 366.2, "max_memory_gb": 28.37, "pair": 0.2793750911951065, "sign": 0.31766193732619286, "keep": 0.22918576686060987, "leak": 2.1005323380231857, "replacement": 1.0779320150613785}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 6.246370792388916, "loss_scale": 4.0, "retries": 0, "seconds": 367.6, "max_memory_gb": 28.37, "pair": 0.2706964276731014, "sign": 0.33544470835477114, "keep": 0.2954288423061371, "leak": 1.692791630513966, "replacement": 1.0093311071395874}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 7.053839206695557, "loss_scale": 4.0, "retries": 0, "seconds": 369.1, "max_memory_gb": 28.37, "pair": 0.5133144035935402, "sign": 0.6754869222640991, "keep": 0.16848304262384772, "leak": 0.771323386579752, "replacement": 1.1281301975250244}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 6.120074272155762, "loss_scale": 4.0, "retries": 0, "seconds": 370.6, "max_memory_gb": 28.37, "pair": 0.4716994855552912, "sign": 0.5539268078282475, "keep": 0.08062080480158329, "leak": 0.9799497523345053, "replacement": 1.0339607447385788}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 6.294834136962891, "loss_scale": 4.0, "retries": 0, "seconds": 372.0, "max_memory_gb": 28.37, "pair": 0.4242624007165432, "sign": 0.47222703229635954, "keep": 0.10685550756170414, "leak": 0.7714304168907802, "replacement": 0.9066613167524338}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 7.4538960456848145, "loss_scale": 4.0, "retries": 0, "seconds": 373.5, "max_memory_gb": 28.37, "pair": 0.16824666410684586, "sign": 0.1341029405593872, "keep": 0.13539164676330984, "leak": 1.6163583099842072, "replacement": 0.7070835381746292}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 3.8117125034332275, "loss_scale": 4.0, "retries": 0, "seconds": 375.0, "max_memory_gb": 28.37, "pair": 0.24768565595149994, "sign": 0.20762541517615318, "keep": 0.006956203607842326, "leak": 1.0051289070397615, "replacement": 0.6062586605548859}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 5.812534809112549, "loss_scale": 4.0, "retries": 0, "seconds": 376.4, "max_memory_gb": 28.37, "pair": 0.28150092903524637, "sign": 0.33771632984280586, "keep": 0.20786436973139644, "leak": 1.2423400059342384, "replacement": 0.8648762851953506}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 5.271564483642578, "loss_scale": 4.0, "retries": 0, "seconds": 377.9, "max_memory_gb": 28.37, "pair": 0.3666035979986191, "sign": 0.3280688086524606, "keep": 0.23903467785567045, "leak": 1.2251013850327581, "replacement": 0.956430695950985}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 7.44493293762207, "loss_scale": 4.0, "retries": 0, "seconds": 379.4, "max_memory_gb": 28.37, "pair": 0.5012763366103172, "sign": 0.540094519034028, "keep": 0.5231416739989072, "leak": 0.3622865527868271, "replacement": 1.1234660670161247}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 5.405226230621338, "loss_scale": 4.0, "retries": 0, "seconds": 380.8, "max_memory_gb": 28.37, "pair": 0.3782059829682112, "sign": 0.3639532132074237, "keep": 0.3606599271297455, "leak": 1.558120807632804, "replacement": 1.1300427466630936}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 7.8380889892578125, "loss_scale": 4.0, "retries": 0, "seconds": 382.3, "max_memory_gb": 28.37, "pair": 0.21502091363072395, "sign": 0.12009360454976559, "keep": 0.4867963647702709, "leak": 2.9049605950713158, "replacement": 1.2447060197591782}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 5.098642826080322, "loss_scale": 4.0, "retries": 0, "seconds": 383.8, "max_memory_gb": 28.37, "pair": 0.39486295357346535, "sign": 0.36501441057771444, "keep": 0.12400927441194654, "leak": 1.5617474821046926, "replacement": 1.029811680316925}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 7.6037373542785645, "loss_scale": 4.0, "retries": 0, "seconds": 385.2, "max_memory_gb": 28.37, "pair": 0.5928183495998383, "sign": 0.5780994743108749, "keep": 1.0683261109516025, "leak": 0.3089633407071233, "replacement": 1.4932719469070435}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 3.2641985416412354, "loss_scale": 4.0, "retries": 0, "seconds": 386.7, "max_memory_gb": 28.37, "pair": 0.4609631523489952, "sign": 0.47542229387909174, "keep": 0.349722818704322, "leak": 0.5787228010594845, "replacement": 1.018216460943222}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 3.5399041175842285, "loss_scale": 4.0, "retries": 0, "seconds": 388.2, "max_memory_gb": 28.37, "pair": 0.6111657842993736, "sign": 0.7363540679216385, "keep": 0.40040274360217154, "leak": 0.40864909562515095, "replacement": 1.2817064225673676}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 7.566380023956299, "loss_scale": 4.0, "retries": 0, "seconds": 389.6, "max_memory_gb": 28.37, "pair": 0.409272238612175, "sign": 0.4393145926296711, "keep": 1.0849203947000206, "leak": 1.0736201908439398, "replacement": 1.4397947937250137}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 6.100890159606934, "loss_scale": 4.0, "retries": 0, "seconds": 391.1, "max_memory_gb": 28.37, "pair": 0.3574051409959793, "sign": 0.4103644387796521, "keep": 0.42734905052930117, "leak": 0.9448608905076981, "replacement": 1.012477084994316}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 6.231600761413574, "loss_scale": 4.0, "retries": 0, "seconds": 392.6, "max_memory_gb": 28.37, "pair": 0.2754503898322582, "sign": 0.26474694535136223, "keep": 0.08929564384743571, "leak": 2.3202709518373013, "replacement": 1.0325394347310066}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 10.852154731750488, "loss_scale": 4.0, "retries": 0, "seconds": 394.0, "max_memory_gb": 28.37, "pair": 0.27777815237641335, "sign": 0.07234255876392126, "keep": 0.07531699258834124, "leak": 0.33563762134872377, "replacement": 0.4355173408985138}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 8.177433967590332, "loss_scale": 4.0, "retries": 0, "seconds": 395.5, "max_memory_gb": 28.37, "pair": 0.3484485000371933, "sign": 0.1679881550371647, "keep": 0.05560422706184909, "leak": 0.16082542203366756, "replacement": 0.5004510581493378}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 7.634413719177246, "loss_scale": 4.0, "retries": 0, "seconds": 396.9, "max_memory_gb": 28.37, "pair": 0.557063490152359, "sign": 0.7931545525789261, "keep": 0.14915989711880684, "leak": 0.3744560023478698, "replacement": 1.1218346953392029}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 4.309587001800537, "loss_scale": 4.0, "retries": 0, "seconds": 398.4, "max_memory_gb": 28.37, "pair": 0.4180501028895378, "sign": 0.47457945067435503, "keep": 1.4089243607595563, "leak": 0.44099824223667383, "replacement": 1.4700515195727348}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 3.96734881401062, "loss_scale": 4.0, "retries": 0, "seconds": 399.8, "max_memory_gb": 28.37, "pair": 0.21057649329304695, "sign": 0.10230915527790785, "keep": 0.05217176442965865, "leak": 0.41869824193418026, "replacement": 0.39249151945114136}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 3.6571905612945557, "loss_scale": 4.0, "retries": 0, "seconds": 401.3, "max_memory_gb": 28.37, "pair": 0.3163229525089264, "sign": 0.27730968594551086, "keep": 0.026380298659205437, "leak": 0.4675651118159294, "replacement": 0.5850592330098152}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 5.728474140167236, "loss_scale": 4.0, "retries": 0, "seconds": 402.7, "max_memory_gb": 28.37, "pair": 0.4255019314587116, "sign": 0.5161459501832724, "keep": 0.37588803691323847, "leak": 0.47445673937909305, "replacement": 0.9901330918073654}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 7.305995464324951, "loss_scale": 4.0, "retries": 0, "seconds": 404.1, "max_memory_gb": 28.37, "pair": 0.32263581082224846, "sign": 0.23573634773492813, "keep": 0.11709603609051555, "leak": 1.719079795992002, "replacement": 0.9288219437003136}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 3.834123373031616, "loss_scale": 4.0, "retries": 0, "seconds": 405.5, "max_memory_gb": 28.37, "pair": 0.22347762249410152, "sign": 0.17079305369406939, "keep": 0.05645661090966314, "leak": 0.7197877203579992, "replacement": 0.5170493796467781}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 5.45281982421875, "loss_scale": 4.0, "retries": 0, "seconds": 407.0, "max_memory_gb": 28.37, "pair": 0.24717599526047707, "sign": 0.2722293664701283, "keep": 0.3401945761288516, "leak": 1.0034009888768196, "replacement": 0.8042382076382637}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 7.389954566955566, "loss_scale": 4.0, "retries": 0, "seconds": 408.4, "max_memory_gb": 28.37, "pair": 0.2734398702159524, "sign": 0.31239357171580195, "keep": 0.15544665697962046, "leak": 0.7909244857728481, "replacement": 0.7050910666584969}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 8.526514053344727, "loss_scale": 4.0, "retries": 0, "seconds": 409.9, "max_memory_gb": 28.37, "pair": 0.1473893988877535, "sign": 0.07149208802729845, "keep": 0.10590321570634842, "leak": 3.231438398361206, "replacement": 1.0439466685056686}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 6.029798984527588, "loss_scale": 4.0, "retries": 0, "seconds": 411.4, "max_memory_gb": 28.37, "pair": 0.17238435707986355, "sign": 0.18211788311600685, "keep": 0.11780750739853829, "leak": 1.240386739373207, "replacement": 0.6324437111616135}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 14.442825317382812, "loss_scale": 4.0, "retries": 0, "seconds": 412.9, "max_memory_gb": 28.37, "pair": 0.3427831679582596, "sign": 0.48707984015345573, "keep": 0.14483188069425523, "leak": 2.072767920792103, "replacement": 1.176931008696556}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 11.164200782775879, "loss_scale": 4.0, "retries": 0, "seconds": 414.4, "max_memory_gb": 28.37, "pair": 0.2469866368919611, "sign": 0.4199677035212517, "keep": 0.04768775880802423, "leak": 2.323000743985176, "replacement": 1.0615645200014114}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 8.809417724609375, "loss_scale": 4.0, "retries": 0, "seconds": 416.0, "max_memory_gb": 28.37, "pair": 0.3380378596484661, "sign": 0.3915906734764576, "keep": 0.4411014709621668, "leak": 0.2148330992177998, "replacement": 0.8080921918153763}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 14.476689338684082, "loss_scale": 4.0, "retries": 0, "seconds": 417.4, "max_memory_gb": 28.37, "pair": 0.3744212780147791, "sign": 0.33399986336007714, "keep": 0.05924154072999954, "leak": 0.6784734497778118, "replacement": 0.7406603395938873}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 7.121931076049805, "loss_scale": 4.0, "retries": 0, "seconds": 418.9, "max_memory_gb": 28.37, "pair": 0.3115442544221878, "sign": 0.2137979557737708, "keep": 0.14389170426875353, "leak": 1.5527186281979084, "replacement": 0.8785687834024429}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 5.604940891265869, "loss_scale": 4.0, "retries": 0, "seconds": 420.4, "max_memory_gb": 28.37, "pair": 0.35577412508428097, "sign": 0.4387021018192172, "keep": 0.3422041059238836, "leak": 0.7958904132246971, "replacement": 0.9451998099684715}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 8.255584716796875, "loss_scale": 4.0, "retries": 0, "seconds": 421.9, "max_memory_gb": 28.37, "pair": 0.1237829178571701, "sign": 0.05670149903744459, "keep": 0.03507166763301939, "leak": 2.358546420931816, "replacement": 0.7593061178922653}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 5.015164852142334, "loss_scale": 4.0, "retries": 0, "seconds": 423.3, "max_memory_gb": 28.37, "pair": 0.2337728776037693, "sign": 0.1831588326022029, "keep": 0.055699304677546024, "leak": 0.6977177653461695, "replacement": 0.5276313945651054}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 6.333771705627441, "loss_scale": 4.0, "retries": 0, "seconds": 424.8, "max_memory_gb": 28.37, "pair": 0.27252318523824215, "sign": 0.26706329360604286, "keep": 0.12016441230662167, "leak": 0.6011965721845627, "replacement": 0.616436168551445}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 8.336138725280762, "loss_scale": 4.0, "retries": 0, "seconds": 426.2, "max_memory_gb": 28.37, "pair": 0.316357359290123, "sign": 0.1566027868539095, "keep": 0.2542200442403555, "leak": 1.2633144482970238, "replacement": 0.8375973701477051}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 8.843623161315918, "loss_scale": 4.0, "retries": 0, "seconds": 427.6, "max_memory_gb": 28.37, "pair": 0.3217923454940319, "sign": 0.4139086343348026, "keep": 0.36412781244143844, "leak": 2.5076049864292145, "replacement": 1.3377117812633514}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 7.96827507019043, "loss_scale": 4.0, "retries": 0, "seconds": 429.1, "max_memory_gb": 28.37, "pair": 0.5020170025527477, "sign": 0.6957926321774721, "keep": 0.5651709353551269, "leak": 1.1942473268136382, "replacement": 1.4310605749487877}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 7.848089694976807, "loss_scale": 4.0, "retries": 0, "seconds": 430.5, "max_memory_gb": 28.37, "pair": 0.4989441819489002, "sign": 0.8330141343176365, "keep": 0.31151326338294894, "leak": 0.951568866148591, "replacement": 1.309100091457367}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 5.2140889167785645, "loss_scale": 4.0, "retries": 0, "seconds": 432.0, "max_memory_gb": 28.37, "pair": 0.5288670659065247, "sign": 0.6783533990383148, "keep": 0.7487271374557167, "leak": 0.3472473605070263, "replacement": 1.3292191848158836}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 8.088099479675293, "loss_scale": 4.0, "retries": 0, "seconds": 433.4, "max_memory_gb": 28.37, "pair": 0.39018198661506176, "sign": 0.13777676224708557, "keep": 0.07986572105437517, "leak": 1.0185584295541048, "replacement": 0.7536428272724152}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 13.92957592010498, "loss_scale": 4.0, "retries": 0, "seconds": 434.9, "max_memory_gb": 28.37, "pair": 0.6961708962917328, "sign": 0.8526216764003038, "keep": 0.514302670955658, "leak": 1.7678819925058633, "replacement": 1.8216036260128021}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 6.047259330749512, "loss_scale": 4.0, "retries": 0, "seconds": 436.3, "max_memory_gb": 28.37, "pair": 0.3360816612839699, "sign": 0.38719226885586977, "keep": 0.13445405196398497, "leak": 0.6578804743476212, "replacement": 0.7613749355077744}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 7.991765975952148, "loss_scale": 4.0, "retries": 0, "seconds": 437.9, "max_memory_gb": 28.37, "pair": 0.18792200088500977, "sign": 0.053747380152344704, "keep": 0.09818678232841194, "leak": 0.7361555048264563, "replacement": 0.44792795926332474}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 9.742389678955078, "loss_scale": 4.0, "retries": 0, "seconds": 439.3, "max_memory_gb": 28.37, "pair": 0.3043743818998337, "sign": 0.268004069570452, "keep": 0.061480697826482356, "leak": 2.2721865624189377, "replacement": 1.0371633619070053}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 12.82282543182373, "loss_scale": 4.0, "retries": 0, "seconds": 440.8, "max_memory_gb": 28.37, "pair": 0.4715992733836174, "sign": 0.5402830895036459, "keep": 0.07549039833247662, "leak": 1.280829736031592, "replacement": 1.0996934734284878}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 8.29389476776123, "loss_scale": 4.0, "retries": 0, "seconds": 442.3, "max_memory_gb": 28.37, "pair": 0.28741494938731194, "sign": 0.14344855956733227, "keep": 0.16268033150117844, "leak": 1.3243547854945064, "replacement": 0.7715680748224258}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 9.662795066833496, "loss_scale": 4.0, "retries": 0, "seconds": 443.7, "max_memory_gb": 28.37, "pair": 0.3126180060207844, "sign": 0.06330016814172268, "keep": 0.06764826737344265, "leak": 0.7565535356407054, "replacement": 0.5672306045889854}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 10.373615264892578, "loss_scale": 4.0, "retries": 0, "seconds": 445.2, "max_memory_gb": 28.37, "pair": 0.4537472575902939, "sign": 0.4238853733986616, "keep": 0.36202074657194316, "leak": 0.8136428948491812, "replacement": 1.0501110032200813}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 8.317516326904297, "loss_scale": 4.0, "retries": 0, "seconds": 446.7, "max_memory_gb": 28.37, "pair": 0.2942013405263424, "sign": 0.47812915593385696, "keep": 0.09457155625568703, "leak": 1.3236816450953484, "replacement": 0.9114720970392227}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 17.797273635864258, "loss_scale": 4.0, "retries": 0, "seconds": 448.0, "max_memory_gb": 28.37, "pair": 0.6375418957322836, "sign": 0.7138351523317397, "keep": 0.44765327870845795, "leak": 1.3291297312825918, "replacement": 1.5505685210227966}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 5.1176886558532715, "loss_scale": 4.0, "retries": 0, "seconds": 449.4, "max_memory_gb": 28.37, "pair": 0.2376703843474388, "sign": 0.10329788085073233, "keep": 0.050830120337195694, "leak": 1.4387374073266983, "replacement": 0.6744187325239182}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 5.573484420776367, "loss_scale": 4.0, "retries": 0, "seconds": 450.9, "max_memory_gb": 28.37, "pair": 0.40201621875166893, "sign": 0.4526950605213642, "keep": 0.30475278571248055, "leak": 0.5135147385299206, "replacement": 0.9091188237071037}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 9.632946014404297, "loss_scale": 4.0, "retries": 0, "seconds": 452.4, "max_memory_gb": 28.37, "pair": 0.31836239621043205, "sign": 0.2155514983460307, "keep": 0.04824644446489401, "leak": 0.4126743735396303, "replacement": 0.5534299463033676}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 15.688261985778809, "loss_scale": 4.0, "retries": 0, "seconds": 453.8, "max_memory_gb": 28.37, "pair": 0.5382833331823349, "sign": 0.615279080811888, "keep": 0.6928738714195788, "leak": 0.8233697712421417, "replacement": 1.398202270269394}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 8.917548179626465, "loss_scale": 4.0, "retries": 0, "seconds": 455.4, "max_memory_gb": 28.37, "pair": 0.3719535991549492, "sign": 0.382720866240561, "keep": 0.3193600226659328, "leak": 1.020599365234375, "replacement": 0.9781438857316971}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 12.039088249206543, "loss_scale": 4.0, "retries": 0, "seconds": 456.9, "max_memory_gb": 28.37, "pair": 0.26150714233517647, "sign": 0.23431283980607986, "keep": 0.08040515566244721, "leak": 1.598891844972968, "replacement": 0.8185891136527061}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 6.227328300476074, "loss_scale": 4.0, "retries": 0, "seconds": 458.3, "max_memory_gb": 28.37, "pair": 0.721128560602665, "sign": 0.7879800014197826, "keep": 0.39931368082761765, "leak": 0.5938348113395477, "replacement": 1.4632340967655182}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 9.41157054901123, "loss_scale": 4.0, "retries": 0, "seconds": 459.7, "max_memory_gb": 28.37, "pair": 0.3696216978132725, "sign": 0.42579600121825933, "keep": 0.25245169177651405, "leak": 1.5427909567952156, "replacement": 1.0944433212280273}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 5.79132080078125, "loss_scale": 4.0, "retries": 0, "seconds": 461.1, "max_memory_gb": 28.37, "pair": 0.33507320284843445, "sign": 0.24426223896443844, "keep": 0.06859658093890175, "leak": 0.9196551740169525, "replacement": 0.7214163839817047}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 6.441504955291748, "loss_scale": 4.0, "retries": 0, "seconds": 462.5, "max_memory_gb": 28.37, "pair": 0.38762839138507843, "sign": 0.3373361090198159, "keep": 0.15996486134827137, "leak": 0.4842686108313501, "replacement": 0.7573460415005684}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 7.428499698638916, "loss_scale": 4.0, "retries": 0, "seconds": 463.8, "max_memory_gb": 28.37, "pair": 0.39992118440568447, "sign": 0.40554908849298954, "keep": 0.28865616262191907, "leak": 0.4623822271823883, "replacement": 0.8626193776726723}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 6.274840354919434, "loss_scale": 4.0, "retries": 0, "seconds": 465.2, "max_memory_gb": 28.37, "pair": 0.33283429220318794, "sign": 0.2660166546702385, "keep": 0.06406231981236488, "leak": 0.6129860430955887, "replacement": 0.6511202752590179}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 11.553864479064941, "loss_scale": 4.0, "retries": 0, "seconds": 466.6, "max_memory_gb": 28.37, "pair": 0.47456590458750725, "sign": 0.4357467480003834, "keep": 0.05454619601368904, "leak": 0.19861514843069017, "replacement": 0.7693661451339722}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 16.101539611816406, "loss_scale": 4.0, "retries": 0, "seconds": 468.0, "max_memory_gb": 28.37, "pair": 0.11821994185447693, "sign": 0.10160049889236689, "keep": 0.04967475542798638, "leak": 4.375592112541199, "replacement": 1.2877555638551712}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 8.173523902893066, "loss_scale": 4.0, "retries": 0, "seconds": 469.4, "max_memory_gb": 28.37, "pair": 0.1129191042855382, "sign": 0.06482542026787996, "keep": 0.075313615729101, "leak": 1.4606927558779716, "replacement": 0.5481618195772171}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 8.878228187561035, "loss_scale": 4.0, "retries": 0, "seconds": 470.8, "max_memory_gb": 28.37, "pair": 0.6583136543631554, "sign": 0.6512535992078483, "keep": 0.557379615289392, "leak": 0.17009059176325536, "replacement": 1.3051529377698898}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 6.539189338684082, "loss_scale": 4.0, "retries": 0, "seconds": 472.2, "max_memory_gb": 28.37, "pair": 0.3981249388307333, "sign": 0.4965632017701864, "keep": 0.22164067850098945, "leak": 1.1895366162061691, "replacement": 1.0546109899878502}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 13.02089786529541, "loss_scale": 4.0, "retries": 0, "seconds": 473.6, "max_memory_gb": 28.37, "pair": 0.19206297397613525, "sign": 0.11474031372927129, "keep": 0.5054673242848366, "leak": 1.2204204469453543, "replacement": 0.807271920144558}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 9.040365219116211, "loss_scale": 4.0, "retries": 0, "seconds": 475.0, "max_memory_gb": 28.37, "pair": 0.537328690290451, "sign": 0.6271246448159218, "keep": 0.1387984047178179, "leak": 0.21829916606657207, "replacement": 0.9748649895191193}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 11.00690746307373, "loss_scale": 4.0, "retries": 0, "seconds": 476.4, "max_memory_gb": 28.37, "pair": 0.305960395373404, "sign": 0.4032547944225371, "keep": 0.33636172174010426, "leak": 1.2272358760237694, "replacement": 0.9825776070356369}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 6.470784664154053, "loss_scale": 4.0, "retries": 0, "seconds": 477.8, "max_memory_gb": 28.37, "pair": 0.43502763099968433, "sign": 0.48926426470279694, "keep": 0.30349044769536704, "leak": 0.7491916622966528, "replacement": 1.0187028646469116}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 10.219966888427734, "loss_scale": 4.0, "retries": 0, "seconds": 479.3, "max_memory_gb": 28.37, "pair": 0.5628334730863571, "sign": 0.6155900172889233, "keep": 0.486568061838625, "leak": 2.0203493235167116, "replacement": 1.618999905884266}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 7.042323589324951, "loss_scale": 4.0, "retries": 0, "seconds": 480.8, "max_memory_gb": 28.37, "pair": 0.3117324896156788, "sign": 0.3000434171408415, "keep": 0.4704415373271331, "leak": 1.017846044152975, "replacement": 0.9514365196228027}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 5.998631954193115, "loss_scale": 4.0, "retries": 0, "seconds": 482.2, "max_memory_gb": 28.37, "pair": 0.6249199900776148, "sign": 0.7661183997988701, "keep": 0.5343471184605733, "leak": 0.22096755215898156, "replacement": 1.3303946554660797}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 10.483612060546875, "loss_scale": 4.0, "retries": 0, "seconds": 483.7, "max_memory_gb": 28.37, "pair": 0.47400934249162674, "sign": 0.5473509849980474, "keep": 0.524584949016571, "leak": 2.023194979003165, "replacement": 1.5157761350274086}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 7.852602005004883, "loss_scale": 4.0, "retries": 0, "seconds": 485.2, "max_memory_gb": 28.37, "pair": 0.4477906860411167, "sign": 0.5356154497712851, "keep": 0.8041046625003219, "leak": 0.5344406943768263, "replacement": 1.2512608766555786}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 8.992766380310059, "loss_scale": 4.0, "retries": 0, "seconds": 486.7, "max_memory_gb": 28.37, "pair": 0.25883733481168747, "sign": 0.13709469139575958, "keep": 0.06043120659887791, "leak": 0.9878596747294068, "replacement": 0.6045651957392693}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 10.539037704467773, "loss_scale": 4.0, "retries": 0, "seconds": 488.2, "max_memory_gb": 28.37, "pair": 0.49561989307403564, "sign": 0.7028119442984462, "keep": 0.40645646303892136, "leak": 1.133996931836009, "replacement": 1.3337532877922058}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 9.226024627685547, "loss_scale": 4.0, "retries": 0, "seconds": 489.7, "max_memory_gb": 28.37, "pair": 0.4438214134424925, "sign": 0.5602168845944107, "keep": 0.49571747065056115, "leak": 1.1039336621761322, "replacement": 1.2477720081806183}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 9.446505546569824, "loss_scale": 4.0, "retries": 0, "seconds": 491.1, "max_memory_gb": 28.37, "pair": 0.46597660332918167, "sign": 0.7359793204814196, "keep": 0.2955497794318944, "leak": 1.5416909512132406, "replacement": 1.3671638891100883}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 10.432514190673828, "loss_scale": 4.0, "retries": 0, "seconds": 492.6, "max_memory_gb": 28.37, "pair": 0.36228659003973007, "sign": 0.3643824025057256, "keep": 0.7544612450292334, "leak": 0.6006538126384839, "replacement": 1.0718718655407429}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 14.850263595581055, "loss_scale": 4.0, "retries": 0, "seconds": 494.0, "max_memory_gb": 28.37, "pair": 0.270405113697052, "sign": 0.3327216515317559, "keep": 0.41461411165073514, "leak": 1.9619390964508057, "replacement": 1.1345578134059906}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 13.461203575134277, "loss_scale": 4.0, "retries": 0, "seconds": 495.4, "max_memory_gb": 28.37, "pair": 0.629226460121572, "sign": 0.7123521128669381, "keep": 0.6342345792800188, "leak": 1.0739525748649612, "replacement": 1.571007952094078}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 14.465938568115234, "loss_scale": 4.0, "retries": 0, "seconds": 496.8, "max_memory_gb": 28.37, "pair": 0.5224695987999439, "sign": 0.6021936684846878, "keep": 0.026901131495833397, "leak": 0.36574308201670647, "replacement": 0.9284527599811554}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 6.618311405181885, "loss_scale": 4.0, "retries": 0, "seconds": 498.2, "max_memory_gb": 28.37, "pair": 0.18249474838376045, "sign": 0.15649915859103203, "keep": 0.023536926659289747, "leak": 1.099970068782568, "replacement": 0.5475053042173386}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 9.238042831420898, "loss_scale": 4.0, "retries": 0, "seconds": 499.7, "max_memory_gb": 28.37, "pair": 0.3700405973941088, "sign": 0.431674397084862, "keep": 0.3067233432084322, "leak": 0.6566646338906139, "replacement": 0.9034056067466736}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 9.432011604309082, "loss_scale": 4.0, "retries": 0, "seconds": 501.2, "max_memory_gb": 28.37, "pair": 0.3669324778020382, "sign": 0.4714326411485672, "keep": 0.553769636695506, "leak": 1.96427071839571, "replacement": 1.370601311326027}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 6.41206693649292, "loss_scale": 4.0, "retries": 0, "seconds": 502.6, "max_memory_gb": 28.37, "pair": 0.4508361518383026, "sign": 0.38546975515782833, "keep": 0.32648119097575545, "leak": 0.35345456533832476, "replacement": 0.8951752483844757}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 6.75233268737793, "loss_scale": 4.0, "retries": 0, "seconds": 504.1, "max_memory_gb": 28.37, "pair": 0.3170965751633048, "sign": 0.1582242250442505, "keep": 0.030534778954461217, "leak": 1.844988630648004, "replacement": 0.8727232068777084}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 9.816843032836914, "loss_scale": 4.0, "retries": 0, "seconds": 505.7, "max_memory_gb": 28.37, "pair": 0.328601804561913, "sign": 0.36838546535000205, "keep": 0.18077432864811271, "leak": 2.2617818884318694, "replacement": 1.1686271727085114}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 10.220748901367188, "loss_scale": 4.0, "retries": 0, "seconds": 507.1, "max_memory_gb": 28.37, "pair": 0.4987865090370178, "sign": 0.7053969874978065, "keep": 0.2571908449754119, "leak": 0.5067792134359479, "replacement": 1.1067752540111542}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 5.865315914154053, "loss_scale": 4.0, "retries": 0, "seconds": 508.5, "max_memory_gb": 28.37, "pair": 0.2197986813262105, "sign": 0.1051212577149272, "keep": 0.07357175834476948, "leak": 0.4587130583822727, "replacement": 0.4238234609365463}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 15.928409576416016, "loss_scale": 4.0, "retries": 0, "seconds": 510.0, "max_memory_gb": 28.37, "pair": 0.3707271758466959, "sign": 0.575886461418122, "keep": 0.5002427429426461, "leak": 1.2725310698151588, "replacement": 1.2269245311617851}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 8.843475341796875, "loss_scale": 4.0, "retries": 0, "seconds": 511.4, "max_memory_gb": 28.37, "pair": 0.4151994623243809, "sign": 0.450317669659853, "keep": 0.41560051357373595, "leak": 0.6378801898099482, "replacement": 1.0076286643743515}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 3.6286909580230713, "loss_scale": 4.0, "retries": 0, "seconds": 512.9, "max_memory_gb": 28.37, "pair": 0.40081945434212685, "sign": 0.4302601292729378, "keep": 0.29991381004219875, "leak": 0.30983316828496754, "replacement": 0.8433647230267525}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 9.877729415893555, "loss_scale": 4.0, "retries": 0, "seconds": 514.4, "max_memory_gb": 28.37, "pair": 0.41287122294306755, "sign": 0.41166362748481333, "keep": 0.28329800348728895, "leak": 1.8887361651286483, "replacement": 1.2325360774993896}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 11.411191940307617, "loss_scale": 4.0, "retries": 0, "seconds": 515.9, "max_memory_gb": 28.37, "pair": 0.139734148979187, "sign": 0.15921726264059544, "keep": 0.03275496396236122, "leak": 3.403267413377762, "replacement": 1.0865371376276016}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 8.746541023254395, "loss_scale": 4.0, "retries": 0, "seconds": 517.4, "max_memory_gb": 28.37, "pair": 0.18516011908650398, "sign": 0.10126369167119265, "keep": 0.15147379785776138, "leak": 1.6389585472643375, "replacement": 0.7212685123085976}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 10.405474662780762, "loss_scale": 4.0, "retries": 0, "seconds": 518.8, "max_memory_gb": 28.37, "pair": 0.6283829659223557, "sign": 0.7405269872397184, "keep": 0.5174788238946348, "leak": 0.2631001519039273, "replacement": 1.3231608718633652}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 5.974414825439453, "loss_scale": 4.0, "retries": 0, "seconds": 520.3, "max_memory_gb": 28.37, "pair": 0.363947581499815, "sign": 0.5539943315088749, "keep": 0.0760648634750396, "leak": 1.0262184627354145, "replacement": 0.9355317875742912}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 8.65456771850586, "loss_scale": 4.0, "retries": 0, "seconds": 521.8, "max_memory_gb": 28.37, "pair": 0.17541689425706863, "sign": 0.18626624532043934, "keep": 0.08281506263301708, "leak": 1.9147829711437225, "replacement": 0.7886532843112946}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 8.26901626586914, "loss_scale": 4.0, "retries": 0, "seconds": 523.3, "max_memory_gb": 28.37, "pair": 0.1544751850888133, "sign": 0.2876862548291683, "keep": 0.033721778308972716, "leak": 1.769864559173584, "replacement": 0.7576453536748886}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 11.296820640563965, "loss_scale": 4.0, "retries": 0, "seconds": 524.7, "max_memory_gb": 28.37, "pair": 0.43630565889179707, "sign": 0.4529283056035638, "keep": 1.042772532091476, "leak": 0.6551826912909746, "replacement": 1.3479517251253128}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 12.707585334777832, "loss_scale": 4.0, "retries": 0, "seconds": 526.1, "max_memory_gb": 28.37, "pair": 0.5270511284470558, "sign": 0.9347762204706669, "keep": 0.06061624363064766, "leak": 0.455385803245008, "replacement": 1.1385938227176666}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 10.519611358642578, "loss_scale": 4.0, "retries": 0, "seconds": 527.6, "max_memory_gb": 28.37, "pair": 0.18074636720120907, "sign": 0.04876583535224199, "keep": 0.05725214444100857, "leak": 1.7458104584366083, "replacement": 0.6702079847455025}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 4.846485614776611, "loss_scale": 4.0, "retries": 0, "seconds": 529.1, "max_memory_gb": 28.37, "pair": 0.37805773690342903, "sign": 0.37566882371902466, "keep": 0.27980957261752337, "leak": 0.2263627052307129, "replacement": 0.7623876333236694}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 9.519440650939941, "loss_scale": 4.0, "retries": 0, "seconds": 530.6, "max_memory_gb": 28.37, "pair": 0.3744376618415117, "sign": 0.34505322575569153, "keep": 0.09666409716010094, "leak": 1.832685612142086, "replacement": 1.0534677058458328}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 4.949957847595215, "loss_scale": 4.0, "retries": 0, "seconds": 532.0, "max_memory_gb": 28.37, "pair": 0.42629941925406456, "sign": 0.4497091304510832, "keep": 0.3964043263113126, "leak": 0.8631762228906155, "replacement": 1.0651501715183258}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 7.768413066864014, "loss_scale": 4.0, "retries": 0, "seconds": 533.5, "max_memory_gb": 28.37, "pair": 0.24762052297592163, "sign": 0.12302058655768633, "keep": 0.06365393206942827, "leak": 0.7082599642453715, "replacement": 0.518022783100605}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 11.81740951538086, "loss_scale": 4.0, "retries": 0, "seconds": 535.0, "max_memory_gb": 28.37, "pair": 0.5274633951485157, "sign": 0.38745883759111166, "keep": 0.06696761620696634, "leak": 0.8368168600136414, "replacement": 0.9638808518648148}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 9.639424324035645, "loss_scale": 4.0, "retries": 0, "seconds": 536.5, "max_memory_gb": 28.37, "pair": 0.28260164335370064, "sign": 0.31511274725198746, "keep": 0.8282296881079674, "leak": 0.5444422103464603, "replacement": 0.9903833866119385}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 5.47838830947876, "loss_scale": 4.0, "retries": 0, "seconds": 537.9, "max_memory_gb": 28.37, "pair": 0.20924518443644047, "sign": 0.07616599556058645, "keep": 0.23804973997175694, "leak": 0.8549537621438503, "replacement": 0.5800914764404297}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 7.9787211418151855, "loss_scale": 4.0, "retries": 0, "seconds": 539.3, "max_memory_gb": 28.37, "pair": 0.2905949791893363, "sign": 0.26383086759597063, "keep": 0.06845447153318673, "leak": 0.9138304449734278, "replacement": 0.6851952448487282}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 6.90522575378418, "loss_scale": 4.0, "retries": 0, "seconds": 540.8, "max_memory_gb": 28.37, "pair": 0.5902224108576775, "sign": 0.7731197625398636, "keep": 0.08958584221545607, "leak": 0.3209667393086981, "replacement": 1.1018169075250626}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 8.301621437072754, "loss_scale": 4.0, "retries": 0, "seconds": 542.3, "max_memory_gb": 28.37, "pair": 0.33664868399500847, "sign": 0.30940290354192257, "keep": 0.5120032927952707, "leak": 0.994111929088831, "replacement": 0.9958797097206116}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 4.545530796051025, "loss_scale": 4.0, "retries": 0, "seconds": 543.7, "max_memory_gb": 28.37, "pair": 0.45514559745788574, "sign": 0.4939294457435608, "keep": 0.4078300445107743, "leak": 0.3811947423964739, "replacement": 1.0013240277767181}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 4.278798580169678, "loss_scale": 4.0, "retries": 0, "seconds": 545.2, "max_memory_gb": 28.37, "pair": 0.2531236857175827, "sign": 0.22838771156966686, "keep": 0.1147460036445409, "leak": 0.2641811668872833, "replacement": 0.49073585867881775}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 5.661510944366455, "loss_scale": 4.0, "retries": 0, "seconds": 546.6, "max_memory_gb": 28.37, "pair": 0.3282428327947855, "sign": 0.365213125012815, "keep": 0.041049518156796694, "leak": 0.4276161529123783, "replacement": 0.6382782235741615}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 8.171171188354492, "loss_scale": 4.0, "retries": 0, "seconds": 548.1, "max_memory_gb": 28.37, "pair": 0.24263913929462433, "sign": 0.1821243204176426, "keep": 0.10221388004720211, "leak": 0.9564718715846539, "replacement": 0.6239261999726295}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 7.7025556564331055, "loss_scale": 4.0, "retries": 0, "seconds": 549.6, "max_memory_gb": 28.37, "pair": 0.2868943326175213, "sign": 0.3535363972187042, "keep": 0.1998775601387024, "leak": 0.48949225014075637, "replacement": 0.6859743744134903}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 10.534259796142578, "loss_scale": 4.0, "retries": 0, "seconds": 551.0, "max_memory_gb": 28.37, "pair": 0.30737245827913284, "sign": 0.3854898549616337, "keep": 0.09792153211310506, "leak": 3.2174364775419235, "replacement": 1.3534372746944427}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 8.017963409423828, "loss_scale": 4.0, "retries": 0, "seconds": 552.4, "max_memory_gb": 28.37, "pair": 0.6530538350343704, "sign": 0.8463924378156662, "keep": 0.43574974313378334, "leak": 0.22500325832515955, "replacement": 1.3503757417201996}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 8.879739761352539, "loss_scale": 4.0, "retries": 0, "seconds": 553.9, "max_memory_gb": 28.37, "pair": 0.41617846488952637, "sign": 0.5326914549805224, "keep": 1.0643138531595469, "leak": 1.076178926974535, "replacement": 1.4837258830666542}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 6.1708502769470215, "loss_scale": 4.0, "retries": 0, "seconds": 555.4, "max_memory_gb": 28.37, "pair": 0.3067941218614578, "sign": 0.30736502818763256, "keep": 0.14297715201973915, "leak": 0.8378389869467355, "replacement": 0.7414249777793884}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 5.996629238128662, "loss_scale": 4.0, "retries": 0, "seconds": 556.8, "max_memory_gb": 28.37, "pair": 0.13550515845417976, "sign": 0.0702125309035182, "keep": 0.03445114195346832, "leak": 0.4675125745088735, "replacement": 0.3047151379287243}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 4.8742756843566895, "loss_scale": 4.0, "retries": 0, "seconds": 558.3, "max_memory_gb": 28.37, "pair": 0.39831838570535183, "sign": 0.48987946286797523, "keep": 0.34683709498494864, "leak": 0.5197246682364494, "replacement": 0.94660784304142}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 12.985743522644043, "loss_scale": 4.0, "retries": 0, "seconds": 559.8, "max_memory_gb": 28.37, "pair": 0.31542924605309963, "sign": 0.18276991555467248, "keep": 0.034303815162274987, "leak": 0.5533079990418628, "replacement": 0.5622931122779846}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 8.11854076385498, "loss_scale": 4.0, "retries": 0, "seconds": 561.3, "max_memory_gb": 28.37, "pair": 0.28178284876048565, "sign": 0.46268741972744465, "keep": 0.10283100651577115, "leak": 1.1154457919765264, "replacement": 0.8434035405516624}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 7.407924175262451, "loss_scale": 4.0, "retries": 0, "seconds": 562.8, "max_memory_gb": 28.37, "pair": 0.17209863290190697, "sign": 0.1576454285532236, "keep": 0.051650885376147926, "leak": 1.49902671482414, "replacement": 0.6515034511685371}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 10.093544006347656, "loss_scale": 4.0, "retries": 0, "seconds": 564.3, "max_memory_gb": 28.37, "pair": 0.25805103592574596, "sign": 0.11529786512255669, "keep": 0.1862955540418625, "leak": 2.295264283195138, "replacement": 0.9826638326048851}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 13.80710506439209, "loss_scale": 4.0, "retries": 0, "seconds": 565.8, "max_memory_gb": 28.37, "pair": 0.3250166866928339, "sign": 0.42329862527549267, "keep": 0.29621099832002074, "leak": 0.9533393867313862, "replacement": 0.9231063649058342}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 7.1091837882995605, "loss_scale": 4.0, "retries": 0, "seconds": 567.3, "max_memory_gb": 28.37, "pair": 0.5641685873270035, "sign": 0.7484457977116108, "keep": 0.558185700327158, "leak": 0.6562138618901372, "replacement": 1.3815378546714783}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 5.201996326446533, "loss_scale": 4.0, "retries": 0, "seconds": 568.7, "max_memory_gb": 28.37, "pair": 0.13732279278337955, "sign": 0.0801788792014122, "keep": 0.05511538463179022, "leak": 0.2934685545042157, "replacement": 0.2783370614051819}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 6.426707744598389, "loss_scale": 4.0, "retries": 0, "seconds": 570.2, "max_memory_gb": 28.37, "pair": 0.11781154293566942, "sign": 0.16717762500047684, "keep": 0.09824824519455433, "leak": 0.5598128917627037, "replacement": 0.3904776871204376}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 8.344503402709961, "loss_scale": 4.0, "retries": 0, "seconds": 571.7, "max_memory_gb": 28.37, "pair": 0.3443186320364475, "sign": 0.44092587660998106, "keep": 0.1665930226445198, "leak": 0.24625913798809052, "replacement": 0.7096428759396076}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 10.451592445373535, "loss_scale": 4.0, "retries": 0, "seconds": 573.2, "max_memory_gb": 28.37, "pair": 0.10371148958802223, "sign": 0.12564911134541035, "keep": 0.05320699920412153, "leak": 2.269387871026993, "replacement": 0.760486513376236}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 11.319659233093262, "loss_scale": 4.0, "retries": 0, "seconds": 574.6, "max_memory_gb": 28.37, "pair": 0.24978389963507652, "sign": 0.08124088216573, "keep": 0.10681868530809879, "leak": 0.6432941358798416, "replacement": 0.5046372190117836}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 7.212383270263672, "loss_scale": 4.0, "retries": 0, "seconds": 576.0, "max_memory_gb": 28.37, "pair": 0.3695280719548464, "sign": 0.5011265659704804, "keep": 0.2346592340618372, "leak": 0.5068181464448571, "replacement": 0.8641254976391792}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 10.169415473937988, "loss_scale": 4.0, "retries": 0, "seconds": 577.5, "max_memory_gb": 28.37, "pair": 0.3275071382522583, "sign": 0.345546274445951, "keep": 0.20822033862350509, "leak": 0.8269585818052292, "replacement": 0.811130091547966}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 14.29847240447998, "loss_scale": 4.0, "retries": 0, "seconds": 579.0, "max_memory_gb": 28.37, "pair": 0.4752224013209343, "sign": 0.698249920271337, "keep": 0.7654363792389631, "leak": 0.3779040053486824, "replacement": 1.3015415370464325}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 14.178327560424805, "loss_scale": 4.0, "retries": 0, "seconds": 580.5, "max_memory_gb": 28.37, "pair": 0.3715657666325569, "sign": 0.2879192382097244, "keep": 0.1812227927148342, "leak": 1.613494070363231, "replacement": 1.0095102787017822}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 50.99598693847656, "loss_scale": 4.0, "retries": 0, "seconds": 582.1, "max_memory_gb": 28.37, "pair": 0.5813659355044365, "sign": 0.7148777926340699, "keep": 0.7344538868055679, "leak": 1.610146897379309, "replacement": 1.7085685208439827}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 9.550865173339844, "loss_scale": 4.0, "retries": 0, "seconds": 583.5, "max_memory_gb": 28.37, "pair": 0.623141709715128, "sign": 0.7531332513317466, "keep": 0.2286184150725603, "leak": 0.8001192016527057, "replacement": 1.3140473514795303}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 8.749368667602539, "loss_scale": 4.0, "retries": 0, "seconds": 585.0, "max_memory_gb": 28.37, "pair": 0.18158412724733353, "sign": 0.1090209404937923, "keep": 0.032106566708534956, "leak": 1.6080139130353928, "replacement": 0.6541513726115227}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 15.849543571472168, "loss_scale": 4.0, "retries": 0, "seconds": 586.4, "max_memory_gb": 28.37, "pair": 0.3687057960778475, "sign": 0.479481284506619, "keep": 0.33553219959139824, "leak": 0.9079492874443531, "replacement": 1.0031998828053474}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 8.383478164672852, "loss_scale": 4.0, "retries": 0, "seconds": 587.8, "max_memory_gb": 28.37, "pair": 0.12541471421718597, "sign": 0.12382939318194985, "keep": 0.07257435884093866, "leak": 0.8377407118678093, "replacement": 0.4330517500638962}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 10.400249481201172, "loss_scale": 4.0, "retries": 0, "seconds": 589.3, "max_memory_gb": 28.37, "pair": 0.4819031208753586, "sign": 0.6385264471173286, "keep": 0.23745483974926174, "leak": 0.8400152996182442, "replacement": 1.1298975571990013}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 16.788288116455078, "loss_scale": 4.0, "retries": 0, "seconds": 590.8, "max_memory_gb": 28.37, "pair": 0.6903941184282303, "sign": 0.9498427454382181, "keep": 2.4314197972416878, "leak": 0.913254464045167, "replacement": 2.6093389242887497}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 9.858122825622559, "loss_scale": 4.0, "retries": 0, "seconds": 592.2, "max_memory_gb": 28.37, "pair": 0.4084237636998296, "sign": 0.5463005872443318, "keep": 0.49797686259262264, "leak": 0.41176754725165665, "replacement": 1.0335044041275978}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 16.632360458374023, "loss_scale": 4.0, "retries": 0, "seconds": 593.7, "max_memory_gb": 28.37, "pair": 0.36311429273337126, "sign": 0.4085689075291157, "keep": 0.40898137041949667, "leak": 1.4227515682578087, "replacement": 1.1275773234665394}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 11.010592460632324, "loss_scale": 4.0, "retries": 0, "seconds": 595.1, "max_memory_gb": 28.37, "pair": 0.21186582185328007, "sign": 0.15500474907457829, "keep": 0.24827254583942704, "leak": 1.3488363474607468, "replacement": 0.7507135644555092}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 14.72938346862793, "loss_scale": 4.0, "retries": 0, "seconds": 596.6, "max_memory_gb": 28.37, "pair": 0.36753519251942635, "sign": 0.6116405576467514, "keep": 0.9486821369500831, "leak": 1.4450167063623667, "replacement": 1.5089507214725018}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 17.485515594482422, "loss_scale": 4.0, "retries": 0, "seconds": 598.1, "max_memory_gb": 28.37, "pair": 0.5071905422955751, "sign": 0.6727189645171165, "keep": 0.32398472039494663, "leak": 0.294169208384119, "replacement": 1.0790846943855286}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 8.240002632141113, "loss_scale": 4.0, "retries": 0, "seconds": 599.5, "max_memory_gb": 28.37, "pair": 0.22919701784849167, "sign": 0.16314767487347126, "keep": 0.04865352623164654, "leak": 1.4162952899932861, "replacement": 0.6891714334487915}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 15.068557739257812, "loss_scale": 4.0, "retries": 0, "seconds": 601.0, "max_memory_gb": 28.37, "pair": 0.3648567274212837, "sign": 0.3918978408910334, "keep": 0.7971814181655645, "leak": 0.4216074254363775, "replacement": 1.0647982023656368}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 12.106219291687012, "loss_scale": 4.0, "retries": 0, "seconds": 602.4, "max_memory_gb": 28.37, "pair": 0.3890036530792713, "sign": 0.3892408860847354, "keep": 1.3214009692892432, "leak": 0.6376972818979993, "replacement": 1.4037488996982574}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 5.239052772521973, "loss_scale": 4.0, "retries": 0, "seconds": 603.8, "max_memory_gb": 28.37, "pair": 0.21263466402888298, "sign": 0.14275077544152737, "keep": 0.05635675828671083, "leak": 0.7235622648149729, "replacement": 0.4930789992213249}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 13.489142417907715, "loss_scale": 4.0, "retries": 0, "seconds": 605.1, "max_memory_gb": 28.37, "pair": 0.8488235548138618, "sign": 1.1483380421996117, "keep": 1.2154652322642505, "leak": 0.27554700151085854, "replacement": 2.0996119379997253}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 7.238080024719238, "loss_scale": 4.0, "retries": 0, "seconds": 606.5, "max_memory_gb": 28.37, "pair": 0.21777338534593582, "sign": 0.21751638688147068, "keep": 0.05227926513180137, "leak": 0.4375258916988969, "replacement": 0.46205267310142517}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 14.849616050720215, "loss_scale": 4.0, "retries": 0, "seconds": 607.8, "max_memory_gb": 28.37, "pair": 0.5190407149493694, "sign": 0.7282169163227081, "keep": 0.08503575203940272, "leak": 1.30148636829108, "replacement": 1.2510386556386948}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 12.75859260559082, "loss_scale": 4.0, "retries": 0, "seconds": 609.2, "max_memory_gb": 28.37, "pair": 0.3023824468255043, "sign": 0.4102302361279726, "keep": 0.6168857682496309, "leak": 1.60593281686306, "replacement": 1.2174236178398132}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 11.921027183532715, "loss_scale": 4.0, "retries": 0, "seconds": 610.6, "max_memory_gb": 28.37, "pair": 0.3147286865860224, "sign": 0.29405676294118166, "keep": 0.07132330373860896, "leak": 1.764248807565309, "replacement": 0.9384808987379074}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 14.81674575805664, "loss_scale": 4.0, "retries": 0, "seconds": 612.0, "max_memory_gb": 28.37, "pair": 0.6875059306621552, "sign": 0.5820583291351795, "keep": 0.5143332215957344, "leak": 0.13524272595532238, "replacement": 1.2695123851299286}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 12.639368057250977, "loss_scale": 4.0, "retries": 0, "seconds": 613.4, "max_memory_gb": 28.37, "pair": 0.6239723600447178, "sign": 0.789553489536047, "keep": 0.9072572626173496, "leak": 1.2433899007737637, "replacement": 1.7832252085208893}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 6.3902974128723145, "loss_scale": 4.0, "retries": 0, "seconds": 614.8, "max_memory_gb": 28.37, "pair": 0.19838479161262512, "sign": 0.07989024743437767, "keep": 0.09245885442942381, "leak": 0.8462041616439819, "replacement": 0.49611037969589233}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 20.411190032958984, "loss_scale": 4.0, "retries": 0, "seconds": 616.3, "max_memory_gb": 28.37, "pair": 0.3996655158698559, "sign": 0.23339782655239105, "keep": 0.0638133017346263, "leak": 0.7598587474494707, "replacement": 0.7382357567548752}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 11.512556076049805, "loss_scale": 4.0, "retries": 0, "seconds": 617.8, "max_memory_gb": 28.37, "pair": 0.677406519651413, "sign": 0.6698492709547281, "keep": 0.45996355824172497, "leak": 0.2569583561271429, "replacement": 1.3065525367856026}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 11.214838027954102, "loss_scale": 4.0, "retries": 0, "seconds": 619.3, "max_memory_gb": 28.37, "pair": 0.3273177668452263, "sign": 0.30128978844732046, "keep": 0.19228563085198402, "leak": 2.45156155154109, "replacement": 1.1869958341121674}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 4.520030498504639, "loss_scale": 4.0, "retries": 0, "seconds": 620.8, "max_memory_gb": 28.37, "pair": 0.1905958317220211, "sign": 0.10652870032936335, "keep": 0.018516571610234678, "leak": 0.49303373508155346, "replacement": 0.3763769045472145}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 18.645212173461914, "loss_scale": 4.0, "retries": 0, "seconds": 622.2, "max_memory_gb": 28.37, "pair": 0.33085401728749275, "sign": 0.08171931374818087, "keep": 0.04656771582085639, "leak": 0.3840206232853234, "replacement": 0.4910027012228966}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 5.603272914886475, "loss_scale": 4.0, "retries": 0, "seconds": 623.7, "max_memory_gb": 28.37, "pair": 0.29034803435206413, "sign": 0.26344143925234675, "keep": 0.19702649384271353, "leak": 0.7315925005823374, "replacement": 0.7034801244735718}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 12.894454002380371, "loss_scale": 4.0, "retries": 0, "seconds": 625.2, "max_memory_gb": 28.37, "pair": 0.12046471238136292, "sign": 0.05562175204977393, "keep": 0.07162079945555888, "leak": 1.6903486847877502, "replacement": 0.6066731736063957}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 10.730412483215332, "loss_scale": 4.0, "retries": 0, "seconds": 626.6, "max_memory_gb": 28.37, "pair": 0.28003518655896187, "sign": 0.26173690147697926, "keep": 0.12569568073377013, "leak": 0.294686546549201, "replacement": 0.5474231131374836}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 6.040855884552002, "loss_scale": 4.0, "retries": 0, "seconds": 628.0, "max_memory_gb": 28.37, "pair": 0.23674500361084938, "sign": 0.27314438112080097, "keep": 0.07422337587922812, "leak": 1.2794731110334396, "replacement": 0.730297140777111}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 4.7096943855285645, "loss_scale": 4.0, "retries": 0, "seconds": 629.4, "max_memory_gb": 28.37, "pair": 0.14999178051948547, "sign": 0.12427319586277008, "keep": 0.01858957455260679, "leak": 0.21610946394503117, "replacement": 0.27545053511857986}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 13.720458984375, "loss_scale": 4.0, "retries": 0, "seconds": 630.8, "max_memory_gb": 28.37, "pair": 0.4238753952085972, "sign": 0.4120347872376442, "keep": 0.4448103138129227, "leak": 0.7659773230552673, "replacement": 1.0437922850251198}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 9.413061141967773, "loss_scale": 4.0, "retries": 0, "seconds": 632.3, "max_memory_gb": 28.37, "pair": 0.2472736593335867, "sign": 0.2833958212286234, "keep": 0.1414107009768486, "leak": 1.260917380452156, "replacement": 0.7749062702059746}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 24.039953231811523, "loss_scale": 4.0, "retries": 0, "seconds": 633.8, "max_memory_gb": 28.37, "pair": 0.4059290625154972, "sign": 0.5485941246151924, "keep": 0.6469873981550336, "leak": 1.7290453985333443, "replacement": 1.4359811395406723}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 5.148892879486084, "loss_scale": 4.0, "retries": 0, "seconds": 635.2, "max_memory_gb": 28.37, "pair": 0.10740473493933678, "sign": 0.0784720522351563, "keep": 0.07498865475645289, "leak": 0.7790492251515388, "replacement": 0.37889739125967026}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 7.474455833435059, "loss_scale": 4.0, "retries": 0, "seconds": 636.7, "max_memory_gb": 28.37, "pair": 0.4273308590054512, "sign": 0.4508784129284322, "keep": 0.36468489468097687, "leak": 0.3337655554059893, "replacement": 0.9185539335012436}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 7.878944396972656, "loss_scale": 4.0, "retries": 0, "seconds": 638.2, "max_memory_gb": 28.37, "pair": 0.31121494621038437, "sign": 0.385280504822731, "keep": 0.03438411274692044, "leak": 0.5319946240633726, "replacement": 0.6540459021925926}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 19.182401657104492, "loss_scale": 4.0, "retries": 0, "seconds": 639.6, "max_memory_gb": 28.37, "pair": 0.6496417000889778, "sign": 0.7478434890508652, "keep": 0.9437733464583289, "leak": 0.6648941487073898, "replacement": 1.6616736352443695}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 10.06530475616455, "loss_scale": 4.0, "retries": 0, "seconds": 641.0, "max_memory_gb": 28.37, "pair": 0.3696793932467699, "sign": 0.39648433960974216, "keep": 0.11275065838708542, "leak": 2.46080575697124, "replacement": 1.23949833214283}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 6.63407039642334, "loss_scale": 4.0, "retries": 0, "seconds": 642.5, "max_memory_gb": 28.37, "pair": 0.2384296916425228, "sign": 0.30954206734895706, "keep": 0.0657401600619778, "leak": 0.6420638598501682, "replacement": 0.5865867510437965}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 12.186230659484863, "loss_scale": 4.0, "retries": 0, "seconds": 643.9, "max_memory_gb": 28.37, "pair": 0.38053683564066887, "sign": 0.5042543057352304, "keep": 0.35174395330250263, "leak": 0.35635396745055914, "replacement": 0.8976244404911995}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 9.765357971191406, "loss_scale": 4.0, "retries": 0, "seconds": 645.3, "max_memory_gb": 28.37, "pair": 0.6106062084436417, "sign": 0.7909626718610525, "keep": 0.9181450312025845, "leak": 1.525070725940168, "replacement": 1.8464277237653732}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 8.619842529296875, "loss_scale": 4.0, "retries": 0, "seconds": 646.7, "max_memory_gb": 28.37, "pair": 0.16955082677304745, "sign": 0.21178370155394077, "keep": 0.2171654049307108, "leak": 1.1055275918915868, "replacement": 0.6604072600603104}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 10.241110801696777, "loss_scale": 4.0, "retries": 0, "seconds": 648.1, "max_memory_gb": 28.37, "pair": 0.44223774038255215, "sign": 0.5822137035429478, "keep": 0.42930381174664944, "leak": 1.7950002254219726, "replacement": 1.3967465311288834}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 8.003361701965332, "loss_scale": 4.0, "retries": 0, "seconds": 649.5, "max_memory_gb": 28.37, "pair": 0.20400265976786613, "sign": 0.24799080938100815, "keep": 0.18833850789815187, "leak": 0.7541033774614334, "replacement": 0.6106931567192078}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 8.319498062133789, "loss_scale": 4.0, "retries": 0, "seconds": 651.0, "max_memory_gb": 28.37, "pair": 0.14128345623612404, "sign": 0.14973778277635574, "keep": 0.07299594243522733, "leak": 1.379629097878933, "replacement": 0.5975575894117355}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 11.268753051757812, "loss_scale": 4.0, "retries": 0, "seconds": 652.4, "max_memory_gb": 28.37, "pair": 0.2567314775660634, "sign": 0.14560260251164436, "keep": 0.32368107413640246, "leak": 1.2274964228272438, "replacement": 0.7982473969459534}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 7.833223819732666, "loss_scale": 4.0, "retries": 0, "seconds": 653.9, "max_memory_gb": 28.37, "pair": 0.15303139202296734, "sign": 0.11115360166877508, "keep": 0.043379404523875564, "leak": 1.429443210363388, "replacement": 0.5876587107777596}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 7.1822052001953125, "loss_scale": 4.0, "retries": 0, "seconds": 655.3, "max_memory_gb": 28.37, "pair": 0.38456812500953674, "sign": 0.34934677090495825, "keep": 0.381433024071157, "leak": 1.376050403341651, "replacement": 1.093970648944378}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 9.257671356201172, "loss_scale": 4.0, "retries": 0, "seconds": 656.8, "max_memory_gb": 28.37, "pair": 0.32816967368125916, "sign": 0.2666151477023959, "keep": 0.043005701154470444, "leak": 0.9451026826282032, "replacement": 0.7192557603120804}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 14.979154586791992, "loss_scale": 4.0, "retries": 0, "seconds": 658.3, "max_memory_gb": 28.37, "pair": 0.4351990707218647, "sign": 0.29591714683920145, "keep": 0.37314883060753345, "leak": 1.1349549740552902, "replacement": 1.0534708201885223}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 5.710115432739258, "loss_scale": 4.0, "retries": 0, "seconds": 659.8, "max_memory_gb": 28.37, "pair": 0.2371082790195942, "sign": 0.24958348274230957, "keep": 0.22266985103487968, "leak": 1.7477619051933289, "replacement": 0.9101754426956177}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 10.461949348449707, "loss_scale": 4.0, "retries": 0, "seconds": 661.3, "max_memory_gb": 28.37, "pair": 0.44663823023438454, "sign": 0.7436109120026231, "keep": 0.08569779060781002, "leak": 0.38797702733427286, "replacement": 0.9582868367433548}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 5.900652885437012, "loss_scale": 4.0, "retries": 0, "seconds": 662.7, "max_memory_gb": 28.37, "pair": 0.18785216473042965, "sign": 0.2711083609610796, "keep": 0.043051436834502965, "leak": 0.48415743187069893, "replacement": 0.4659714177250862}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 12.219132423400879, "loss_scale": 4.0, "retries": 0, "seconds": 664.2, "max_memory_gb": 28.37, "pair": 0.3788014054298401, "sign": 0.483220137655735, "keep": 0.30099307000637054, "leak": 1.4494727769633755, "replacement": 1.1332762241363525}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 7.550717830657959, "loss_scale": 4.0, "retries": 0, "seconds": 665.7, "max_memory_gb": 28.37, "pair": 0.27589226141572, "sign": 0.2847518064081669, "keep": 0.06676334142684937, "leak": 0.3172990658786148, "replacement": 0.5309746190905571}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 8.979776382446289, "loss_scale": 4.0, "retries": 0, "seconds": 667.1, "max_memory_gb": 28.37, "pair": 0.2120429165661335, "sign": 0.24177252687513828, "keep": 0.09102417004760355, "leak": 0.3732895366847515, "replacement": 0.47176366299390793}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 13.991700172424316, "loss_scale": 4.0, "retries": 0, "seconds": 668.6, "max_memory_gb": 28.37, "pair": 0.33578261360526085, "sign": 0.6150618083775043, "keep": 0.5662563545629382, "leak": 1.0415920168161392, "replacement": 1.186839684844017}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 5.969209671020508, "loss_scale": 4.0, "retries": 0, "seconds": 670.1, "max_memory_gb": 28.37, "pair": 0.21263504028320312, "sign": 0.22346235066652298, "keep": 0.02955137484241277, "leak": 0.6494059152901173, "replacement": 0.5014933720231056}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 13.336238861083984, "loss_scale": 4.0, "retries": 0, "seconds": 671.6, "max_memory_gb": 28.37, "pair": 0.33147056587040424, "sign": 0.23249244131147861, "keep": 0.058660398004576564, "leak": 0.23356831027194858, "replacement": 0.5354390777647495}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 7.460275650024414, "loss_scale": 4.0, "retries": 0, "seconds": 673.0, "max_memory_gb": 28.37, "pair": 0.152268729172647, "sign": 0.10248358640819788, "keep": 0.026731934864073992, "leak": 2.0616170316934586, "replacement": 0.7322807163000107}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 10.47463607788086, "loss_scale": 4.0, "retries": 0, "seconds": 674.5, "max_memory_gb": 28.37, "pair": 0.15867257490754128, "sign": 0.20725965779274702, "keep": 0.1603495089802891, "leak": 1.8560780603438616, "replacement": 0.8064966760575771}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 12.684396743774414, "loss_scale": 4.0, "retries": 0, "seconds": 676.0, "max_memory_gb": 28.37, "pair": 0.3278169743716717, "sign": 0.22571242600679398, "keep": 0.145388375967741, "leak": 1.0284054297953844, "replacement": 0.7704687267541885}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 7.5833420753479, "loss_scale": 4.0, "retries": 0, "seconds": 677.4, "max_memory_gb": 28.37, "pair": 0.21708719991147518, "sign": 0.12554840557277203, "keep": 0.05189772881567478, "leak": 0.8463254794478416, "replacement": 0.5173916518688202}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 6.631997585296631, "loss_scale": 4.0, "retries": 0, "seconds": 678.9, "max_memory_gb": 28.37, "pair": 0.198301300406456, "sign": 0.15660322410985827, "keep": 0.03650638181716204, "leak": 0.08903863094747066, "replacement": 0.31711576506495476}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 15.433690071105957, "loss_scale": 4.0, "retries": 0, "seconds": 680.3, "max_memory_gb": 28.37, "pair": 0.4279845114797354, "sign": 0.15605992451310158, "keep": 0.43235941301099956, "leak": 1.2688200548291206, "replacement": 1.039399191737175}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 6.14856481552124, "loss_scale": 4.0, "retries": 0, "seconds": 681.8, "max_memory_gb": 28.37, "pair": 0.1444168221205473, "sign": 0.06837597116827965, "keep": 0.030581839673686773, "leak": 0.467817023396492, "replacement": 0.31084999069571495}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 10.917768478393555, "loss_scale": 4.0, "retries": 0, "seconds": 683.3, "max_memory_gb": 28.37, "pair": 0.49262600019574165, "sign": 0.5519608613103628, "keep": 0.4090923359617591, "leak": 0.8757605479331687, "replacement": 1.192092776298523}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 11.559871673583984, "loss_scale": 4.0, "retries": 0, "seconds": 684.7, "max_memory_gb": 28.37, "pair": 0.3183916201815009, "sign": 0.30315890815109015, "keep": 0.06300797127187252, "leak": 0.7901136241853237, "replacement": 0.6990034654736519}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 13.23290729522705, "loss_scale": 4.0, "retries": 0, "seconds": 686.1, "max_memory_gb": 28.37, "pair": 0.2454347088932991, "sign": 0.1675223857164383, "keep": 0.1456221322878264, "leak": 0.11975162476301193, "replacement": 0.4319448545575142}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 13.832677841186523, "loss_scale": 4.0, "retries": 0, "seconds": 687.6, "max_memory_gb": 28.37, "pair": 0.31921568885445595, "sign": 0.4256689175963402, "keep": 0.3563290309975855, "leak": 0.31881353874086926, "replacement": 0.7899180836975574}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 13.465764045715332, "loss_scale": 4.0, "retries": 0, "seconds": 689.0, "max_memory_gb": 28.37, "pair": 0.23631275445222855, "sign": 0.3788193017244339, "keep": 0.3849975580815226, "leak": 0.6768281906843185, "replacement": 0.7874282374978065}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 9.144339561462402, "loss_scale": 4.0, "retries": 0, "seconds": 690.5, "max_memory_gb": 28.37, "pair": 0.2552371993660927, "sign": 0.17108905129134655, "keep": 0.059395160991698503, "leak": 2.4657213985919952, "replacement": 0.9869096577167511}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 14.460222244262695, "loss_scale": 4.0, "retries": 0, "seconds": 692.0, "max_memory_gb": 28.37, "pair": 0.2501209508627653, "sign": 0.4915539654903114, "keep": 0.02424933377187699, "leak": 0.7865572860464454, "replacement": 0.7046619206666946}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 12.26041030883789, "loss_scale": 4.0, "retries": 0, "seconds": 693.4, "max_memory_gb": 28.37, "pair": 0.29664371348917484, "sign": 0.236637556925416, "keep": 0.034693918365519494, "leak": 0.5955515969544649, "replacement": 0.5811973437666893}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 15.923044204711914, "loss_scale": 4.0, "retries": 0, "seconds": 695.0, "max_memory_gb": 28.37, "pair": 0.45855190232396126, "sign": 0.5068040378391743, "keep": 0.6210218289634213, "leak": 0.738362907897681, "replacement": 1.207055613398552}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 9.292012214660645, "loss_scale": 4.0, "retries": 0, "seconds": 696.4, "max_memory_gb": 28.37, "pair": 0.45858342200517654, "sign": 0.5577289797365665, "keep": 0.46632018126547337, "leak": 0.8908810811117291, "replacement": 1.1933282911777496}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 7.958188056945801, "loss_scale": 4.0, "retries": 0, "seconds": 697.8, "max_memory_gb": 28.37, "pair": 0.13539971131831408, "sign": 0.12808100041002035, "keep": 0.11286233030841686, "leak": 1.0126438066363335, "replacement": 0.5090323314070702}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 15.982488632202148, "loss_scale": 4.0, "retries": 0, "seconds": 699.3, "max_memory_gb": 28.37, "pair": 0.27817573957145214, "sign": 0.4451961927115917, "keep": 0.29005611955653876, "leak": 0.8640948664397001, "replacement": 0.861825630068779}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 14.870759963989258, "loss_scale": 4.0, "retries": 0, "seconds": 700.7, "max_memory_gb": 28.37, "pair": 0.47022683918476105, "sign": 0.7903390675783157, "keep": 0.5098860934376717, "leak": 0.2658108612522483, "replacement": 1.1867921650409698}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 22.06114387512207, "loss_scale": 4.0, "retries": 0, "seconds": 702.2, "max_memory_gb": 28.37, "pair": 0.6801295652985573, "sign": 0.941843057051301, "keep": 1.2883498016744852, "leak": 0.49738754844293, "replacement": 1.9195728600025177}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 17.316415786743164, "loss_scale": 4.0, "retries": 0, "seconds": 703.6, "max_memory_gb": 28.37, "pair": 0.22547730524092913, "sign": 0.47592560946941376, "keep": 0.05434641405008733, "leak": 1.4010789394378662, "replacement": 0.8408830240368843}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 12.583410263061523, "loss_scale": 4.0, "retries": 0, "seconds": 705.0, "max_memory_gb": 28.37, "pair": 0.462256520986557, "sign": 0.5879348563030362, "keep": 0.6303326105698943, "leak": 0.3268040579278022, "replacement": 1.1530912518501282}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 5.760562896728516, "loss_scale": 4.0, "retries": 0, "seconds": 706.4, "max_memory_gb": 28.37, "pair": 0.18720080889761448, "sign": 0.17171417735517025, "keep": 0.11765203869435936, "leak": 0.6553499652072787, "replacement": 0.49572139233350754}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 11.179559707641602, "loss_scale": 4.0, "retries": 0, "seconds": 707.9, "max_memory_gb": 28.37, "pair": 0.287935109809041, "sign": 0.33260492980480194, "keep": 0.17553858319297433, "leak": 2.140435006469488, "replacement": 1.0771155804395676}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 12.44533920288086, "loss_scale": 4.0, "retries": 0, "seconds": 709.4, "max_memory_gb": 28.37, "pair": 0.22145674843341112, "sign": 0.37677386216819286, "keep": 0.7218004275346175, "leak": 1.5098361670970917, "replacement": 1.1482029929757118}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 9.543554306030273, "loss_scale": 4.0, "retries": 0, "seconds": 710.8, "max_memory_gb": 28.37, "pair": 0.27688170596957207, "sign": 0.3134241015650332, "keep": 0.08606941078323871, "leak": 0.6635853096377105, "replacement": 0.6425247676670551}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 17.142623901367188, "loss_scale": 4.0, "retries": 0, "seconds": 712.3, "max_memory_gb": 28.37, "pair": 0.26859848015010357, "sign": 0.3857684638351202, "keep": 0.16935199641739018, "leak": 1.1261102338739875, "replacement": 0.8276862613856792}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 7.185181140899658, "loss_scale": 4.0, "retries": 0, "seconds": 713.7, "max_memory_gb": 28.37, "pair": 0.21235709637403488, "sign": 0.1677123922854662, "keep": 0.03577719273744151, "leak": 0.6018151603639126, "replacement": 0.46455567330121994}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 16.253860473632812, "loss_scale": 4.0, "retries": 0, "seconds": 715.1, "max_memory_gb": 28.37, "pair": 0.5594073589891195, "sign": 0.8255747109651566, "keep": 0.44089093268848956, "leak": 0.5627742558717728, "replacement": 1.3333337754011154}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 13.546945571899414, "loss_scale": 4.0, "retries": 0, "seconds": 716.6, "max_memory_gb": 28.37, "pair": 0.16698837839066982, "sign": 0.2548279478214681, "keep": 0.09040814358741045, "leak": 0.8352986760437489, "replacement": 0.5484311208128929}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 10.291513442993164, "loss_scale": 4.0, "retries": 0, "seconds": 718.0, "max_memory_gb": 28.37, "pair": 0.17929751798510551, "sign": 0.22882011905312538, "keep": 0.054619637317955494, "leak": 0.9873503223061562, "replacement": 0.56785499304533}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 7.9048943519592285, "loss_scale": 4.0, "retries": 0, "seconds": 719.5, "max_memory_gb": 28.37, "pair": 0.24052266217768192, "sign": 0.06688921712338924, "keep": 0.11911647592205554, "leak": 0.8952218648046255, "replacement": 0.5573309659957886}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 34.18135452270508, "loss_scale": 4.0, "retries": 0, "seconds": 721.0, "max_memory_gb": 28.37, "pair": 0.36909477040171623, "sign": 0.5718461684882641, "keep": 0.8205364282475784, "leak": 1.6365895122289658, "replacement": 1.4744334518909454}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 9.015326499938965, "loss_scale": 4.0, "retries": 0, "seconds": 722.4, "max_memory_gb": 28.37, "pair": 0.42294277623295784, "sign": 0.37769151013344526, "keep": 0.26541154040023685, "leak": 1.3216709718981292, "replacement": 1.0749120265245438}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 13.086395263671875, "loss_scale": 4.0, "retries": 0, "seconds": 723.9, "max_memory_gb": 28.37, "pair": 0.40352432057261467, "sign": 0.37081106193363667, "keep": 0.33878790884045884, "leak": 0.6628562789410353, "replacement": 0.9240378588438034}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 22.144500732421875, "loss_scale": 4.0, "retries": 0, "seconds": 725.3, "max_memory_gb": 28.37, "pair": 0.47660175897181034, "sign": 0.5761090572923422, "keep": 0.7739287143340334, "leak": 0.7747287005186081, "replacement": 1.3453027978539467}
diff --git a/runs/single_llava/replacement/L02_s13/adapter/README.md b/runs/single_llava/replacement/L02_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L02_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L02_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L02_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..ba18852dc0e98f82bf4115874691b0d709b8bc88
--- /dev/null
+++ b/runs/single_llava/replacement/L02_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.2.mlp.gate_proj",
+    "model.language_model.layers.2.self_attn.k_proj",
+    "model.language_model.layers.2.self_attn.o_proj",
+    "model.language_model.layers.2.self_attn.v_proj",
+    "model.language_model.layers.2.self_attn.q_proj",
+    "model.language_model.layers.2.mlp.up_proj",
+    "model.language_model.layers.2.mlp.down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L02_s13/config.json b/runs/single_llava/replacement/L02_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e563ba84799a1cdeb37d748b4a655aa97acbe96f
--- /dev/null
+++ b/runs/single_llava/replacement/L02_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 2,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L02_s13/steps.jsonl b/runs/single_llava/replacement/L02_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..dbd53078bb645983fd44a99761d847eb5de5330c
--- /dev/null
+++ b/runs/single_llava/replacement/L02_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.2817051410675049, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 27.64, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.4738163650035858, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 27.66, "pair": 1.097527801990509, "sign": 0.5363437794148922, "keep": 5.372197345364839, "leak": 9.845238878369855e-06, "replacement": 4.051800727844238}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.713406503200531, "loss_scale": 4.0, "retries": 0, "seconds": 4.6, "max_memory_gb": 27.78, "pair": 1.0961261689662933, "sign": 0.5418421141803265, "keep": 10.803682327270508, "leak": 4.5084161740760464e-05, "replacement": 6.76889955997467}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.5934984087944031, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 27.79, "pair": 1.0924069285392761, "sign": 0.4717551190406084, "keep": 7.671070475131273, "leak": 2.539782953192571e-05, "replacement": 5.163826048374176}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.3141006827354431, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 27.79, "pair": 1.0980133712291718, "sign": 0.7765458524227142, "keep": 0.15900829061865807, "leak": 1.6421627313523146e-06, "replacement": 1.565790832042694}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.29982316493988037, "loss_scale": 4.0, "retries": 0, "seconds": 8.9, "max_memory_gb": 27.79, "pair": 1.0957257747650146, "sign": 0.7367488890886307, "keep": 2.9357824586331844, "leak": 5.509062293640454e-05, "replacement": 2.932005226612091}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.691548228263855, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 27.79, "pair": 1.085175335407257, "sign": 0.5467642629519105, "keep": 6.880534393712878, "leak": 0.00026791538766701706, "replacement": 4.798891723155975}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.6026771664619446, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 27.79, "pair": 1.08881676197052, "sign": 0.5875262171030045, "keep": 7.388506159186363, "leak": 0.0002416540742729012, "replacement": 5.076893329620361}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.5860347747802734, "loss_scale": 4.0, "retries": 0, "seconds": 13.2, "max_memory_gb": 27.79, "pair": 1.085835188627243, "sign": 0.6188745154067874, "keep": 5.715079002082348, "leak": 0.00038848950574532637, "replacement": 4.252909034490585}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.44116926193237305, "loss_scale": 4.0, "retries": 0, "seconds": 14.7, "max_memory_gb": 27.79, "pair": 1.08646821975708, "sign": 0.7801981754601002, "keep": 3.2745020240545273, "leak": 0.00019059260029496983, "replacement": 3.113865941762924}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.694628119468689, "loss_scale": 4.0, "retries": 0, "seconds": 16.2, "max_memory_gb": 27.79, "pair": 1.0708692073822021, "sign": 0.453110633417964, "keep": 5.2710761446505785, "leak": 0.0007706085486489656, "replacement": 3.933155208826065}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.5920441150665283, "loss_scale": 4.0, "retries": 0, "seconds": 17.6, "max_memory_gb": 27.79, "pair": 1.0794925391674042, "sign": 0.6098762229084969, "keep": 3.0586095778271556, "leak": 0.0007365669408159192, "replacement": 2.913919597864151}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.5355865359306335, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 27.79, "pair": 1.0799387991428375, "sign": 0.50962795317173, "keep": 5.357793655246496, "leak": 0.0008255865744539737, "replacement": 4.013855934143066}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.8635652661323547, "loss_scale": 4.0, "retries": 0, "seconds": 20.5, "max_memory_gb": 27.79, "pair": 1.0579308569431305, "sign": 0.39653499238193035, "keep": 7.804916150867939, "leak": 0.0012576627185376665, "replacement": 5.158970803022385}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.5774020552635193, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 27.79, "pair": 1.0738334655761719, "sign": 0.6898348145186901, "keep": 5.190639819251373, "leak": 0.0009304663281000103, "replacement": 4.01430344581604}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.543117105960846, "loss_scale": 4.0, "retries": 0, "seconds": 23.3, "max_memory_gb": 27.79, "pair": 1.0800548195838928, "sign": 0.7812766656279564, "keep": 2.765119630843401, "leak": 0.00043817378991661826, "replacement": 2.853362560272217}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.4505954086780548, "loss_scale": 4.0, "retries": 0, "seconds": 24.7, "max_memory_gb": 27.79, "pair": 1.0790804326534271, "sign": 0.7122386395931244, "keep": 3.1923536341637373, "leak": 0.002064391770545626, "replacement": 3.0318925976753235}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.7485449314117432, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 27.87, "pair": 1.066823571920395, "sign": 0.5362308099865913, "keep": 7.207710199058056, "leak": 0.0009074541412701365, "replacement": 4.939021021127701}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.2003628015518188, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 27.87, "pair": 1.0274905264377594, "sign": 0.530842125415802, "keep": 10.731139659881592, "leak": 0.0051172979874536395, "replacement": 6.65976083278656}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.4456416666507721, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 27.87, "pair": 1.0896302163600922, "sign": 0.4305683374404907, "keep": 0.023883539717644453, "leak": 0.00012690736686238324, "replacement": 1.3168878555297852}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.701176106929779, "loss_scale": 4.0, "retries": 0, "seconds": 30.5, "max_memory_gb": 27.87, "pair": 1.0592089146375656, "sign": 0.3748345822095871, "keep": 5.572023785673082, "leak": 0.0011363871290086536, "replacement": 4.032922178506851}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.7722740173339844, "loss_scale": 4.0, "retries": 0, "seconds": 31.9, "max_memory_gb": 27.87, "pair": 1.049563080072403, "sign": 0.5720532834529877, "keep": 5.9628886468708515, "leak": 0.0066590812921276665, "replacement": 4.3186987936496735}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.9206231236457825, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 27.87, "pair": 1.0225822180509567, "sign": 0.24823899753391743, "keep": 5.342677770648152, "leak": 0.007096724788880238, "replacement": 3.8198148608207703}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.4648168087005615, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 27.87, "pair": 0.9992198795080185, "sign": 0.25653447583317757, "keep": 10.093616247177124, "leak": 0.005709298973670229, "replacement": 6.175722599029541}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.1040608882904053, "loss_scale": 4.0, "retries": 0, "seconds": 36.2, "max_memory_gb": 27.87, "pair": 1.00673246383667, "sign": 0.38951220363378525, "keep": 8.50822164863348, "leak": 0.008919862351831398, "replacement": 5.45782932639122}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.3639613389968872, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 27.87, "pair": 0.9825603365898132, "sign": 0.6585562964901328, "keep": 10.583276987075806, "leak": 0.014366950141265988, "replacement": 6.607068657875061}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.1044178009033203, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 27.87, "pair": 1.0126438289880753, "sign": 0.5889389477670193, "keep": 8.477174527943134, "leak": 0.008816969666497698, "replacement": 5.547904849052429}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.8031459450721741, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 27.87, "pair": 1.0514820218086243, "sign": 0.7764229625463486, "keep": 5.677723973989487, "leak": 0.006051160044421522, "replacement": 4.280068278312683}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 0.6688621044158936, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 27.87, "pair": 1.0541159212589264, "sign": 0.8368307575583458, "keep": 5.26480358466506, "leak": 0.004022828336957218, "replacement": 4.105938762426376}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 0.9950656294822693, "loss_scale": 4.0, "retries": 0, "seconds": 43.5, "max_memory_gb": 27.87, "pair": 0.9872885346412659, "sign": 0.34663252905011177, "keep": 5.5740710981190205, "leak": 0.01832933407786186, "replacement": 3.95222270488739}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 0.8738688826560974, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 27.87, "pair": 1.0072946697473526, "sign": 0.8013875838369131, "keep": 5.7591873705387115, "leak": 0.024146111525624292, "replacement": 4.293618738651276}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.552808403968811, "loss_scale": 4.0, "retries": 0, "seconds": 46.5, "max_memory_gb": 27.87, "pair": 0.9712538123130798, "sign": 0.3870690315961838, "keep": 10.861537218093872, "leak": 0.020697631174698472, "replacement": 6.6007314920425415}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 0.9252247214317322, "loss_scale": 4.0, "retries": 0, "seconds": 47.9, "max_memory_gb": 27.87, "pair": 1.0147127509117126, "sign": 0.62191341817379, "keep": 4.958471050485969, "leak": 0.030267088926367203, "replacement": 3.8124715983867645}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 0.881714403629303, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 27.87, "pair": 1.0147426277399063, "sign": 0.5228221714496613, "keep": 5.360461314208806, "leak": 0.00805834704078734, "replacement": 3.9583989679813385}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.1765532493591309, "loss_scale": 4.0, "retries": 0, "seconds": 50.8, "max_memory_gb": 27.87, "pair": 0.9646505117416382, "sign": 0.4296038933098316, "keep": 5.562486752867699, "leak": 0.024705652191187255, "replacement": 3.9668723046779633}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.0553098917007446, "loss_scale": 4.0, "retries": 0, "seconds": 52.3, "max_memory_gb": 27.87, "pair": 0.9573881924152374, "sign": 0.4307584837079048, "keep": 5.241551697254181, "leak": 0.023097464891179698, "replacement": 3.799317628145218}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.1666111946105957, "loss_scale": 4.0, "retries": 0, "seconds": 53.7, "max_memory_gb": 27.87, "pair": 0.9666561335325241, "sign": 0.5830428618937731, "keep": 5.031005397439003, "leak": 0.03597020350025559, "replacement": 3.7826728522777557}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 1.7937922477722168, "loss_scale": 4.0, "retries": 0, "seconds": 55.3, "max_memory_gb": 27.87, "pair": 0.9226193279027939, "sign": 0.7762796431779861, "keep": 9.619816064834595, "leak": 0.029460467863827944, "replacement": 6.128032326698303}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.880584180355072, "loss_scale": 4.0, "retries": 0, "seconds": 56.7, "max_memory_gb": 27.87, "pair": 1.0242395848035812, "sign": 0.5619047358632088, "keep": 2.942239135503769, "leak": 0.028486058406997472, "replacement": 2.783433049917221}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.25542414188385, "loss_scale": 4.0, "retries": 0, "seconds": 58.1, "max_memory_gb": 27.87, "pair": 0.9664637893438339, "sign": 0.49157142639160156, "keep": 5.608846627175808, "leak": 0.021220909664407372, "replacement": 4.021978080272675}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.5027607679367065, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 27.87, "pair": 1.062016025185585, "sign": 1.021115705370903, "keep": 3.0198273956775665, "leak": 0.008895198292520945, "replacement": 3.0847113132476807}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.1915478706359863, "loss_scale": 4.0, "retries": 0, "seconds": 61.0, "max_memory_gb": 27.87, "pair": 0.9972256124019623, "sign": 0.3990702498704195, "keep": 4.79452108964324, "leak": 0.007337034883676097, "replacement": 3.5958555042743683}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 1.3559471368789673, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 27.87, "pair": 0.9777195155620575, "sign": 0.5096406303346157, "keep": 4.995556244626641, "leak": 0.010260346854920499, "replacement": 3.732883095741272}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 0.9398077130317688, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 27.87, "pair": 0.9941593706607819, "sign": 0.5727969333529472, "keep": 2.8399616088718176, "leak": 0.015150561812333763, "replacement": 2.7043263018131256}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 0.6429174542427063, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 27.87, "pair": 1.0550654530525208, "sign": 0.8918667882680893, "keep": 2.885459393262863, "leak": 0.005488894249992882, "replacement": 2.9451008439064026}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.2298613786697388, "loss_scale": 4.0, "retries": 0, "seconds": 66.8, "max_memory_gb": 27.87, "pair": 1.0000884234905243, "sign": 0.7234296053647995, "keep": 4.716262340545654, "leak": 0.030319855615482005, "replacement": 3.7275144159793854}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 1.310543417930603, "loss_scale": 4.0, "retries": 0, "seconds": 68.3, "max_memory_gb": 27.87, "pair": 0.9050084054470062, "sign": 0.6586335599422455, "keep": 4.240762133616954, "leak": 0.04558258812176064, "replacement": 3.3661020398139954}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 2.23781418800354, "loss_scale": 4.0, "retries": 0, "seconds": 69.8, "max_memory_gb": 27.87, "pair": 0.8608633279800415, "sign": 0.2006770996376872, "keep": 7.811100021004677, "leak": 0.06585047880071215, "replacement": 4.883214473724365}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 2.075930595397949, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 27.87, "pair": 0.8451159745454788, "sign": 0.5837158337235451, "keep": 7.397205214947462, "leak": 0.06984100461704656, "replacement": 4.8530367612838745}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 1.9474495649337769, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 27.87, "pair": 0.9376567900180817, "sign": 0.4371692091226578, "keep": 7.460464388132095, "leak": 0.03004725056598545, "replacement": 4.893985480070114}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 0.9588145613670349, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 27.87, "pair": 0.9919336438179016, "sign": 0.5769950933754444, "keep": 2.784848570358008, "leak": 0.03478321839338605, "replacement": 2.6815512478351593}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 1.541501760482788, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 27.87, "pair": 0.898185133934021, "sign": 0.6872389931231737, "keep": 4.888366959989071, "leak": 0.06473492373334011, "replacement": 3.702171891927719}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 1.7949137687683105, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 27.87, "pair": 0.9524824321269989, "sign": 0.743928924202919, "keep": 7.71500489115715, "leak": 0.04206311326472645, "replacement": 5.19246506690979}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 2.5610086917877197, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 27.87, "pair": 0.8472298830747604, "sign": 0.6242691352963448, "keep": 8.123515948653221, "leak": 0.07222445838851854, "replacement": 5.239178359508514}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 2.1381921768188477, "loss_scale": 4.0, "retries": 0, "seconds": 79.8, "max_memory_gb": 27.87, "pair": 0.8903504312038422, "sign": 0.4369201511144638, "keep": 4.981209269259125, "leak": 0.0495209875080036, "replacement": 3.6117953956127167}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 3.849403142929077, "loss_scale": 4.0, "retries": 0, "seconds": 81.2, "max_memory_gb": 27.87, "pair": 0.6940256655216217, "sign": 0.16934731602668762, "keep": 9.668060779571533, "leak": 0.11903379671275616, "replacement": 5.6424880027771}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.0048151016235352, "loss_scale": 4.0, "retries": 0, "seconds": 82.7, "max_memory_gb": 27.87, "pair": 0.9976796358823776, "sign": 0.7882027626037598, "keep": 2.5365168899297714, "leak": 0.011614224180902966, "replacement": 2.662943035364151}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 2.768709659576416, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 27.87, "pair": 0.9182140976190567, "sign": 0.27432925067842007, "keep": 4.89855498727411, "leak": 0.09424454902182333, "replacement": 3.528217315673828}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.2158105373382568, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 27.87, "pair": 1.0098975151777267, "sign": 0.6653846949338913, "keep": 2.8875851817429066, "leak": 0.004041189815325197, "replacement": 2.787392795085907}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 2.2881429195404053, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 27.87, "pair": 0.8793729692697525, "sign": 0.6219838615506887, "keep": 4.933015971444547, "leak": 0.03587932465597987, "replacement": 3.6658427715301514}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 4.177335739135742, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 27.87, "pair": 0.8694346249103546, "sign": 0.6455254033207893, "keep": 6.91050061583519, "leak": 0.09696044845622964, "replacement": 4.6716877818107605}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.5776876211166382, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 27.87, "pair": 0.979543998837471, "sign": 0.7091646604239941, "keep": 2.744979538023472, "leak": 0.013986970551364664, "replacement": 2.710112899541855}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 4.574782371520996, "loss_scale": 4.0, "retries": 0, "seconds": 91.3, "max_memory_gb": 27.87, "pair": 0.901161327958107, "sign": 0.5634399354457855, "keep": 7.272267548367381, "leak": 0.040871518838685006, "replacement": 4.829232901334763}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 5.373219966888428, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 27.92, "pair": 0.7818914651870728, "sign": 0.46606624824926257, "keep": 7.077554356306791, "leak": 0.10046732128830627, "replacement": 4.578818619251251}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 5.316930294036865, "loss_scale": 4.0, "retries": 0, "seconds": 94.3, "max_memory_gb": 27.92, "pair": 0.9247929751873016, "sign": 0.5652987919747829, "keep": 7.094744265079498, "leak": 0.05828233773354441, "replacement": 4.7693851590156555}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.5536004304885864, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 27.92, "pair": 1.095348984003067, "sign": 0.6589651331305504, "keep": 1.8588012433610857, "leak": 0.008061797845584806, "replacement": 2.3562475740909576}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 8.228431701660156, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 27.92, "pair": 0.7811040282249451, "sign": 0.3817225620150566, "keep": 6.074716925621033, "leak": 0.2588745355751598, "replacement": 4.07404226064682}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 2.5259978771209717, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 27.92, "pair": 1.0032947063446045, "sign": 0.5076137520372868, "keep": 2.3795631255488843, "leak": 0.024693098428542726, "replacement": 2.4530563354492188}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 5.534269332885742, "loss_scale": 4.0, "retries": 0, "seconds": 100.1, "max_memory_gb": 27.92, "pair": 0.882193848490715, "sign": 0.6983588561415672, "keep": 4.130822024773806, "leak": 0.02855140440078685, "replacement": 3.3039221465587616}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 5.057831764221191, "loss_scale": 4.0, "retries": 0, "seconds": 101.6, "max_memory_gb": 27.92, "pair": 1.0564977526664734, "sign": 0.6465602815151215, "keep": 3.859792845323682, "leak": 0.11639182631915901, "replacement": 3.3387722074985504}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 6.9065351486206055, "loss_scale": 4.0, "retries": 0, "seconds": 103.0, "max_memory_gb": 27.92, "pair": 0.881095364689827, "sign": 0.1818632259964943, "keep": 5.903035033494234, "leak": 0.129525897475105, "replacement": 3.9559259116649628}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 8.467080116271973, "loss_scale": 4.0, "retries": 0, "seconds": 104.5, "max_memory_gb": 27.92, "pair": 0.7914260029792786, "sign": 0.28561828657984734, "keep": 6.9822224378585815, "leak": 0.10957149881869555, "replacement": 4.452739238739014}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.8260660767555237, "loss_scale": 4.0, "retries": 0, "seconds": 105.9, "max_memory_gb": 27.92, "pair": 1.0341999381780624, "sign": 0.8046500459313393, "keep": 0.43838122859597206, "leak": 0.010726227705163183, "replacement": 1.6583971679210663}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 4.829427719116211, "loss_scale": 4.0, "retries": 0, "seconds": 107.3, "max_memory_gb": 27.92, "pair": 1.013728305697441, "sign": 0.5049932710826397, "keep": 3.493757002055645, "leak": 0.03568386898058762, "replacement": 3.0220245122909546}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 2.423180103302002, "loss_scale": 4.0, "retries": 0, "seconds": 108.8, "max_memory_gb": 27.92, "pair": 0.9524993896484375, "sign": 0.896068599075079, "keep": 2.938564456999302, "leak": 0.06040322737680981, "replacement": 2.8849167227745056}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 5.179271221160889, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 27.92, "pair": 0.9086727052927017, "sign": 0.45000268146395683, "keep": 3.03153208270669, "leak": 0.05698558106087148, "replacement": 2.663686454296112}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 6.680859088897705, "loss_scale": 4.0, "retries": 0, "seconds": 111.6, "max_memory_gb": 27.92, "pair": 1.0294935405254364, "sign": 0.5052573047578335, "keep": 5.113830983638763, "leak": 0.17660897719906643, "replacement": 3.8831897974014282}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 3.1929142475128174, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 27.92, "pair": 1.081667184829712, "sign": 0.7647160813212395, "keep": 2.683749794960022, "leak": 0.19353516265982762, "replacement": 2.8542838990688324}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 2.155067205429077, "loss_scale": 4.0, "retries": 0, "seconds": 114.5, "max_memory_gb": 27.92, "pair": 0.9861599951982498, "sign": 0.588654175400734, "keep": 1.44864420639351, "leak": 0.007681820909965609, "replacement": 2.0067296028137207}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 3.896568536758423, "loss_scale": 4.0, "retries": 0, "seconds": 115.9, "max_memory_gb": 27.92, "pair": 1.1834287345409393, "sign": 0.4237107113003731, "keep": 2.4655057596974075, "leak": 0.06298096411046572, "replacement": 2.643782287836075}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 7.365220546722412, "loss_scale": 4.0, "retries": 0, "seconds": 117.3, "max_memory_gb": 27.92, "pair": 0.960928276181221, "sign": 0.5122857987880707, "keep": 3.9684286639094353, "leak": 0.30858558183535933, "replacement": 3.278431922197342}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 6.437034606933594, "loss_scale": 4.0, "retries": 0, "seconds": 118.8, "max_memory_gb": 27.92, "pair": 1.131926417350769, "sign": 0.42868334613740444, "keep": 3.726194389164448, "leak": 0.016179860236661625, "replacement": 3.2134103775024414}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 6.195207595825195, "loss_scale": 4.0, "retries": 0, "seconds": 120.3, "max_memory_gb": 27.92, "pair": 1.0422490239143372, "sign": 0.5311847180128098, "keep": 3.481399592012167, "leak": 0.3068983417470008, "replacement": 3.125265747308731}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 3.6707494258880615, "loss_scale": 4.0, "retries": 0, "seconds": 121.8, "max_memory_gb": 27.92, "pair": 1.0937244594097137, "sign": 0.8359365612268448, "keep": 1.5838745336513966, "leak": 0.11309360167069826, "replacement": 2.3319033682346344}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.9022011756896973, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 27.92, "pair": 1.0716237723827362, "sign": 0.5822483375668526, "keep": 0.14190179342404008, "leak": 0.013081865232379641, "replacement": 1.43696928024292}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 6.8447794914245605, "loss_scale": 4.0, "retries": 0, "seconds": 124.6, "max_memory_gb": 27.92, "pair": 1.12120620906353, "sign": 0.6883782297372818, "keep": 3.6324906051158905, "leak": 0.037482374229057314, "replacement": 3.291011244058609}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 7.71906042098999, "loss_scale": 4.0, "retries": 0, "seconds": 126.0, "max_memory_gb": 27.92, "pair": 1.0132759660482407, "sign": 0.4957813583314419, "keep": 3.09979485347867, "leak": 0.30371557967737317, "replacement": 2.8869930505752563}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 5.314063549041748, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 27.92, "pair": 1.1604770720005035, "sign": 0.6497489400207996, "keep": 2.1126679070293903, "leak": 0.03611437836843834, "replacement": 2.550714135169983}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 2.195168972015381, "loss_scale": 4.0, "retries": 0, "seconds": 128.8, "max_memory_gb": 27.92, "pair": 1.1312993466854095, "sign": 0.7108869403600693, "keep": 0.22213704464957118, "leak": 0.021232865681497515, "replacement": 1.6031195521354675}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 3.6007227897644043, "loss_scale": 4.0, "retries": 0, "seconds": 130.2, "max_memory_gb": 27.92, "pair": 0.9547291249036789, "sign": 0.6090961694717407, "keep": 1.9288879153318703, "leak": 0.24418438132852316, "replacement": 2.2847672402858734}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.1978750228881836, "loss_scale": 4.0, "retries": 0, "seconds": 131.8, "max_memory_gb": 27.92, "pair": 1.035583734512329, "sign": 1.0116755813360214, "keep": 0.650710579007864, "leak": 0.008366592802803163, "replacement": 1.8688684403896332}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.1668190956115723, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 27.92, "pair": 1.113675832748413, "sign": 0.6000074446201324, "keep": 0.9872471252456307, "leak": 0.5939547044945357, "replacement": 2.0557918548583984}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 5.7982258796691895, "loss_scale": 4.0, "retries": 0, "seconds": 134.7, "max_memory_gb": 27.92, "pair": 1.1600822061300278, "sign": 0.34470370411872864, "keep": 1.8808086616918445, "leak": 0.08268649410456419, "replacement": 2.2935099601745605}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 6.972667694091797, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 27.92, "pair": 0.920767292380333, "sign": 0.4164048880338669, "keep": 2.9817773699760437, "leak": 0.22589005960617214, "replacement": 2.676330953836441}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 6.014079570770264, "loss_scale": 4.0, "retries": 0, "seconds": 137.6, "max_memory_gb": 27.92, "pair": 0.8569213300943375, "sign": 0.4857693389058113, "keep": 2.0127386935055256, "leak": 0.10247556422837079, "replacement": 2.1317942142486572}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 6.751667022705078, "loss_scale": 4.0, "retries": 0, "seconds": 139.0, "max_memory_gb": 27.92, "pair": 1.157761886715889, "sign": 0.5030520632863045, "keep": 2.522524893283844, "leak": 1.270546951360302, "replacement": 2.988187074661255}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.7043375968933105, "loss_scale": 4.0, "retries": 0, "seconds": 140.5, "max_memory_gb": 27.92, "pair": 1.0685083866119385, "sign": 0.5537822358310223, "keep": 1.0269620046019554, "leak": 0.09290012379642576, "replacement": 1.882105529308319}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 4.8062567710876465, "loss_scale": 4.0, "retries": 0, "seconds": 141.9, "max_memory_gb": 27.92, "pair": 0.985029011964798, "sign": 0.5654890164732933, "keep": 1.404345914721489, "leak": 0.18854632656439207, "replacement": 2.0170831382274628}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 2.5402166843414307, "loss_scale": 4.0, "retries": 0, "seconds": 143.4, "max_memory_gb": 27.92, "pair": 1.1265856325626373, "sign": 0.9075963720679283, "keep": 0.41934819938614964, "leak": 0.030911674606613815, "replacement": 1.7977857887744904}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 6.870230197906494, "loss_scale": 4.0, "retries": 0, "seconds": 144.9, "max_memory_gb": 27.92, "pair": 1.0005512982606888, "sign": 0.6736618205904961, "keep": 2.1258621215820312, "leak": 0.7866710438393056, "replacement": 2.5969809889793396}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 7.293521404266357, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 27.92, "pair": 1.1100357174873352, "sign": 0.2154662199318409, "keep": 1.794827163219452, "leak": 0.10631618555635214, "replacement": 2.14176145195961}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 6.131606101989746, "loss_scale": 4.0, "retries": 0, "seconds": 147.8, "max_memory_gb": 27.92, "pair": 0.9069455862045288, "sign": 0.37448861077427864, "keep": 2.0900131165981293, "leak": 0.41190626146271825, "replacement": 2.2421730756759644}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 5.436519145965576, "loss_scale": 4.0, "retries": 0, "seconds": 149.3, "max_memory_gb": 27.92, "pair": 0.9042467623949051, "sign": 0.40541767701506615, "keep": 1.22108005406335, "leak": 0.5608740779571235, "replacement": 1.857714205980301}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 5.051082611083984, "loss_scale": 4.0, "retries": 0, "seconds": 150.6, "max_memory_gb": 27.92, "pair": 1.053589791059494, "sign": 0.8528179675340652, "keep": 0.8512593945488334, "leak": 0.013082880526781082, "replacement": 1.908899188041687}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 4.479809284210205, "loss_scale": 4.0, "retries": 0, "seconds": 152.1, "max_memory_gb": 27.92, "pair": 0.9482328444719315, "sign": 0.4457002691924572, "keep": 1.177616324275732, "leak": 1.3735558241605759, "replacement": 2.10328009724617}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 4.574153900146484, "loss_scale": 4.0, "retries": 0, "seconds": 153.5, "max_memory_gb": 27.92, "pair": 0.8715285062789917, "sign": 0.5192628726363182, "keep": 1.3045514822006226, "leak": 1.3623146649333648, "replacement": 2.124014377593994}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.656514883041382, "loss_scale": 4.0, "retries": 0, "seconds": 155.0, "max_memory_gb": 27.92, "pair": 1.040057212114334, "sign": 0.7869739830493927, "keep": 0.5850851610302925, "leak": 0.01788283872883767, "replacement": 1.7305574417114258}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 4.142504692077637, "loss_scale": 4.0, "retries": 0, "seconds": 156.4, "max_memory_gb": 27.92, "pair": 1.0672423988580704, "sign": 0.7378170788288116, "keep": 0.7235728995874524, "leak": 0.9452746036404278, "replacement": 2.034256100654602}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 4.101782321929932, "loss_scale": 4.0, "retries": 0, "seconds": 157.9, "max_memory_gb": 27.92, "pair": 0.831515945494175, "sign": 0.4988696463406086, "keep": 1.1909357011318207, "leak": 1.2580912067496683, "replacement": 1.9909414052963257}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 2.484180450439453, "loss_scale": 4.0, "retries": 0, "seconds": 159.4, "max_memory_gb": 27.92, "pair": 0.9298763945698738, "sign": 0.571414602920413, "keep": 0.8872232139110565, "leak": 0.46242629271000624, "replacement": 1.7748018503189087}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 3.350961685180664, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 27.92, "pair": 0.8120973855257034, "sign": 0.4599667228758335, "keep": 0.5993910990655422, "leak": 0.05797666046419181, "replacement": 1.3562705218791962}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 2.688408613204956, "loss_scale": 4.0, "retries": 0, "seconds": 162.1, "max_memory_gb": 27.92, "pair": 0.9020007252693176, "sign": 0.49374543502926826, "keep": 0.5922650191932917, "leak": 0.18934357957914472, "replacement": 1.4923418760299683}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 2.7400217056274414, "loss_scale": 4.0, "retries": 0, "seconds": 163.5, "max_memory_gb": 27.92, "pair": 0.8878176361322403, "sign": 0.5599335357546806, "keep": 0.5279499553143978, "leak": 0.10639831784646958, "replacement": 1.4583590030670166}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 4.719062805175781, "loss_scale": 4.0, "retries": 0, "seconds": 165.0, "max_memory_gb": 27.92, "pair": 0.7956054210662842, "sign": 0.5305228997021914, "keep": 1.3617252707481384, "leak": 0.6519017331884243, "replacement": 1.90470489859581}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 3.0144314765930176, "loss_scale": 4.0, "retries": 0, "seconds": 166.4, "max_memory_gb": 27.92, "pair": 0.9793079048395157, "sign": 0.618606872856617, "keep": 0.700616255402565, "leak": 0.49706856213742867, "replacement": 1.763186663389206}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 2.4115207195281982, "loss_scale": 4.0, "retries": 0, "seconds": 167.9, "max_memory_gb": 27.92, "pair": 0.9370575845241547, "sign": 0.689009502530098, "keep": 0.2457240717485547, "leak": 0.01154053647769615, "replacement": 1.4073095172643661}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 4.117080211639404, "loss_scale": 4.0, "retries": 0, "seconds": 169.3, "max_memory_gb": 27.92, "pair": 0.8552332073450089, "sign": 0.4616558849811554, "keep": 1.171617990359664, "leak": 0.4372344095099834, "replacement": 1.7811787724494934}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 2.6963586807250977, "loss_scale": 4.0, "retries": 0, "seconds": 170.7, "max_memory_gb": 27.92, "pair": 1.0281845331192017, "sign": 0.5522761978209019, "keep": 0.4397464250214398, "leak": 0.03954682479979965, "replacement": 1.5340825468301773}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.4870681762695312, "loss_scale": 4.0, "retries": 0, "seconds": 172.2, "max_memory_gb": 27.92, "pair": 1.066824048757553, "sign": 0.8258656412363052, "keep": 0.47821962740272284, "leak": 0.015052129456307739, "replacement": 1.722629725933075}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 3.314645767211914, "loss_scale": 4.0, "retries": 0, "seconds": 173.7, "max_memory_gb": 27.92, "pair": 0.8054813742637634, "sign": 0.5465064095333219, "keep": 1.0657717771828175, "leak": 0.5351944300346076, "replacement": 1.7454191148281097}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.2601804733276367, "loss_scale": 4.0, "retries": 0, "seconds": 175.2, "max_memory_gb": 27.92, "pair": 0.9700733125209808, "sign": 0.7509144842624664, "keep": 0.8434886857867241, "leak": 1.3111220778664574, "replacement": 2.0950554609298706}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 5.6250104904174805, "loss_scale": 4.0, "retries": 0, "seconds": 176.7, "max_memory_gb": 27.92, "pair": 0.8670108169317245, "sign": 0.6328742839396, "keep": 0.39639852568507195, "leak": 0.5600619042925246, "replacement": 1.5216627418994904}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.4569929838180542, "loss_scale": 4.0, "retries": 0, "seconds": 178.1, "max_memory_gb": 27.92, "pair": 0.8052246533334255, "sign": 0.7787858694791794, "keep": 0.9136897921562195, "leak": 0.6438450059358729, "replacement": 1.8124237060546875}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 2.3102941513061523, "loss_scale": 4.0, "retries": 0, "seconds": 179.5, "max_memory_gb": 27.92, "pair": 0.9794245511293411, "sign": 1.0045789778232574, "keep": 0.6089527504518628, "leak": 0.02079791634969297, "replacement": 1.791389912366867}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.8308730125427246, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 27.92, "pair": 0.9331511855125427, "sign": 0.8141053393483162, "keep": 0.433837590739131, "leak": 0.009438297769520432, "replacement": 1.5594821870326996}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 6.392273902893066, "loss_scale": 4.0, "retries": 0, "seconds": 182.3, "max_memory_gb": 27.92, "pair": 0.7433666586875916, "sign": 0.6162210120819509, "keep": 1.0597608536481857, "leak": 0.2236364441923797, "replacement": 1.6372667253017426}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 3.1857750415802, "loss_scale": 4.0, "retries": 0, "seconds": 183.7, "max_memory_gb": 27.92, "pair": 0.8351365327835083, "sign": 0.3809318244457245, "keep": 0.26190393045544624, "leak": 0.06718520552385598, "replacement": 1.1733507215976715}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 2.872934579849243, "loss_scale": 4.0, "retries": 0, "seconds": 185.1, "max_memory_gb": 27.92, "pair": 0.7045166343450546, "sign": 0.3330305591225624, "keep": 0.47875534649938345, "leak": 0.5391647217329592, "replacement": 1.2452007681131363}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 7.9504289627075195, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 27.92, "pair": 0.5700589120388031, "sign": 0.23511163517832756, "keep": 0.6828533485531807, "leak": 1.3922748556360602, "replacement": 1.3771101534366608}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 5.986687183380127, "loss_scale": 4.0, "retries": 0, "seconds": 188.0, "max_memory_gb": 27.92, "pair": 0.8091654479503632, "sign": 0.835724301636219, "keep": 0.9260016046464443, "leak": 0.38897296669892967, "replacement": 1.7872715890407562}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 4.872713088989258, "loss_scale": 4.0, "retries": 0, "seconds": 189.4, "max_memory_gb": 27.92, "pair": 0.8136517703533173, "sign": 0.7030271738767624, "keep": 0.773453988134861, "leak": 0.09729224229522515, "replacement": 1.5762154161930084}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 3.598450183868408, "loss_scale": 4.0, "retries": 0, "seconds": 190.9, "max_memory_gb": 27.92, "pair": 0.5324597284197807, "sign": 0.14074093848466873, "keep": 1.128944996278733, "leak": 1.2301849341019988, "replacement": 1.4748488664627075}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 2.967154026031494, "loss_scale": 4.0, "retries": 0, "seconds": 192.4, "max_memory_gb": 27.92, "pair": 0.5948712080717087, "sign": 0.42383187636733055, "keep": 0.7773461192846298, "leak": 1.6416928535618354, "replacement": 1.605883464217186}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 3.707376003265381, "loss_scale": 4.0, "retries": 0, "seconds": 193.9, "max_memory_gb": 27.92, "pair": 0.9470904916524887, "sign": 0.7126045152544975, "keep": 0.3847911134362221, "leak": 0.036691137007437646, "replacement": 1.5049611330032349}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.2659311294555664, "loss_scale": 4.0, "retries": 0, "seconds": 195.3, "max_memory_gb": 27.92, "pair": 0.5769525840878487, "sign": 0.4391287434846163, "keep": 0.5221685022115707, "leak": 2.4797761087102117, "replacement": 1.677545189857483}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 3.8735251426696777, "loss_scale": 4.0, "retries": 0, "seconds": 196.7, "max_memory_gb": 27.92, "pair": 0.7164201140403748, "sign": 0.477360500022769, "keep": 0.5738645866513252, "leak": 0.27285630168626085, "replacement": 1.3102467358112335}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 2.6517436504364014, "loss_scale": 4.0, "retries": 0, "seconds": 198.3, "max_memory_gb": 27.92, "pair": 0.6861568838357925, "sign": 0.4560894388705492, "keep": 0.3785748742520809, "leak": 1.8864648373564705, "replacement": 1.5751053094863892}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 3.9625496864318848, "loss_scale": 4.0, "retries": 0, "seconds": 199.7, "max_memory_gb": 27.92, "pair": 0.6484499648213387, "sign": 0.308500443585217, "keep": 0.3295781798660755, "leak": 0.0799245669040829, "replacement": 0.987470418214798}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 5.058800220489502, "loss_scale": 4.0, "retries": 0, "seconds": 201.2, "max_memory_gb": 27.92, "pair": 0.6569050624966621, "sign": 0.5718522183597088, "keep": 0.7579122930765152, "leak": 0.8229261150117964, "replacement": 1.5275188237428665}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 2.3834943771362305, "loss_scale": 4.0, "retries": 0, "seconds": 202.7, "max_memory_gb": 27.92, "pair": 0.6714385077357292, "sign": 0.5158358328044415, "keep": 0.36828087642788887, "leak": 0.36957317776978016, "replacement": 1.20589017868042}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 1.8481433391571045, "loss_scale": 4.0, "retries": 0, "seconds": 204.1, "max_memory_gb": 27.92, "pair": 0.7058520689606667, "sign": 0.5231708865612745, "keep": 0.4595742058008909, "leak": 0.9531231219880283, "replacement": 1.435505360364914}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.586540460586548, "loss_scale": 4.0, "retries": 0, "seconds": 205.5, "max_memory_gb": 27.92, "pair": 0.5742584988474846, "sign": 0.40057874377816916, "keep": 0.3451932189054787, "leak": 0.2253076104734646, "replacement": 1.0034714043140411}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 2.231788396835327, "loss_scale": 4.0, "retries": 0, "seconds": 206.9, "max_memory_gb": 27.92, "pair": 0.9477445781230927, "sign": 0.9829470366239548, "keep": 0.25395901466254145, "leak": 0.08142305689580098, "replacement": 1.5865534543991089}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 2.817294120788574, "loss_scale": 4.0, "retries": 0, "seconds": 208.3, "max_memory_gb": 27.92, "pair": 0.5995264947414398, "sign": 0.39622986130416393, "keep": 0.46326310373842716, "leak": 1.3150943638756871, "replacement": 1.3580465614795685}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 4.726444721221924, "loss_scale": 4.0, "retries": 0, "seconds": 209.7, "max_memory_gb": 27.92, "pair": 0.6956913322210312, "sign": 0.322244584094733, "keep": 0.15986865106970072, "leak": 0.5794575996696949, "replacement": 1.0816123634576797}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.0263025760650635, "loss_scale": 4.0, "retries": 0, "seconds": 211.2, "max_memory_gb": 27.92, "pair": 0.6318031698465347, "sign": 0.5548917092382908, "keep": 0.37279454432427883, "leak": 0.8441921607591212, "replacement": 1.3066943883895874}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.0237748622894287, "loss_scale": 4.0, "retries": 0, "seconds": 212.6, "max_memory_gb": 27.92, "pair": 0.6063450090587139, "sign": 0.5125967338681221, "keep": 0.4989710212685168, "leak": 0.9589736866764724, "replacement": 1.351872280240059}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 3.5971620082855225, "loss_scale": 4.0, "retries": 0, "seconds": 214.0, "max_memory_gb": 27.92, "pair": 0.6605605781078339, "sign": 0.3198021985590458, "keep": 0.107519042911008, "leak": 0.1510483775564353, "replacement": 0.9119832962751389}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 1.806618571281433, "loss_scale": 4.0, "retries": 0, "seconds": 215.4, "max_memory_gb": 27.92, "pair": 0.6726933903992176, "sign": 0.4967665635049343, "keep": 0.270909221842885, "leak": 0.07685337658040226, "replacement": 1.0757446736097336}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 4.4616899490356445, "loss_scale": 4.0, "retries": 0, "seconds": 216.8, "max_memory_gb": 27.92, "pair": 0.41593121364712715, "sign": 0.3411238696426153, "keep": 0.6702345088124275, "leak": 2.5129079916514456, "replacement": 1.5498373806476593}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 5.010674953460693, "loss_scale": 4.0, "retries": 0, "seconds": 218.2, "max_memory_gb": 27.92, "pair": 0.5705323368310928, "sign": 0.4397939518094063, "keep": 0.5636117164976895, "leak": 0.3157450743019581, "replacement": 1.1511714458465576}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 4.246367454528809, "loss_scale": 4.0, "retries": 0, "seconds": 219.7, "max_memory_gb": 27.92, "pair": 0.3548904359340668, "sign": 0.23698561545461416, "keep": 0.34942373633384705, "leak": 2.585587367415428, "replacement": 1.294491931796074}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 3.23175311088562, "loss_scale": 4.0, "retries": 0, "seconds": 221.2, "max_memory_gb": 27.92, "pair": 0.4445769041776657, "sign": 0.2704806588590145, "keep": 0.21580338943749666, "leak": 1.66253150254488, "replacement": 1.1033517867326736}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.2297263145446777, "loss_scale": 4.0, "retries": 0, "seconds": 222.6, "max_memory_gb": 27.92, "pair": 0.3607415445148945, "sign": 0.21657798811793327, "keep": 0.8469343017786741, "leak": 0.9029425531625748, "replacement": 1.1182333081960678}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.2777085304260254, "loss_scale": 4.0, "retries": 0, "seconds": 224.1, "max_memory_gb": 27.92, "pair": 0.5795822665095329, "sign": 0.5212722569704056, "keep": 1.2135894750244915, "leak": 1.9970794832333922, "replacement": 1.9462830424308777}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 2.2073819637298584, "loss_scale": 4.0, "retries": 0, "seconds": 225.6, "max_memory_gb": 27.92, "pair": 0.6564125195145607, "sign": 0.5458661615848541, "keep": 0.3277836716733873, "leak": 0.3780953595414758, "replacement": 1.1877612620592117}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 2.0463855266571045, "loss_scale": 4.0, "retries": 0, "seconds": 227.0, "max_memory_gb": 27.92, "pair": 0.8407216221094131, "sign": 1.1786244362592697, "keep": 0.31505388766527176, "leak": 0.1657157390145585, "replacement": 1.6289897412061691}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.944172978401184, "loss_scale": 4.0, "retries": 0, "seconds": 228.5, "max_memory_gb": 27.92, "pair": 0.6843973696231842, "sign": 0.6810384541749954, "keep": 0.5290023293346167, "leak": 0.3278931600507349, "replacement": 1.3713910579681396}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 2.5677943229675293, "loss_scale": 4.0, "retries": 0, "seconds": 230.0, "max_memory_gb": 27.92, "pair": 0.3671464715152979, "sign": 0.19326928444206715, "keep": 0.13501478917896748, "leak": 2.353594172745943, "replacement": 1.1196870505809784}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 3.9325270652770996, "loss_scale": 4.0, "retries": 0, "seconds": 231.4, "max_memory_gb": 27.92, "pair": 0.5621059685945511, "sign": 0.3649474112316966, "keep": 0.29033179208636284, "leak": 1.768824048456736, "replacement": 1.3319516032934189}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 1.7466861009597778, "loss_scale": 4.0, "retries": 0, "seconds": 232.8, "max_memory_gb": 27.92, "pair": 0.697392787784338, "sign": 0.6094485819339752, "keep": 0.4223004821687937, "leak": 0.4337830433360068, "replacement": 1.3217131048440933}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 1.8817802667617798, "loss_scale": 4.0, "retries": 0, "seconds": 234.1, "max_memory_gb": 27.92, "pair": 0.6851447653025389, "sign": 0.6307441564276814, "keep": 0.15770718501880765, "leak": 0.3146571878751274, "replacement": 1.1580347642302513}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 3.2805731296539307, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 27.92, "pair": 0.40471720695495605, "sign": 0.30345575883984566, "keep": 0.14861528761684895, "leak": 0.8457008376717567, "replacement": 0.8421779572963715}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 2.3740291595458984, "loss_scale": 4.0, "retries": 0, "seconds": 237.0, "max_memory_gb": 27.92, "pair": 0.712919294834137, "sign": 0.925743225030601, "keep": 0.4674336239695549, "leak": 0.9249332027829951, "replacement": 1.6407410800457}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 2.7463862895965576, "loss_scale": 4.0, "retries": 0, "seconds": 238.5, "max_memory_gb": 27.92, "pair": 0.5822571441531181, "sign": 0.596087820827961, "keep": 0.2726209657266736, "leak": 1.5904546715319157, "replacement": 1.414225161075592}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 2.308887004852295, "loss_scale": 4.0, "retries": 0, "seconds": 240.0, "max_memory_gb": 27.92, "pair": 0.5206529796123505, "sign": 0.5043289437890053, "keep": 0.2701084800064564, "leak": 1.403846422326751, "replacement": 1.2588332742452621}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 4.619688510894775, "loss_scale": 4.0, "retries": 0, "seconds": 241.5, "max_memory_gb": 27.92, "pair": 0.37981530372053385, "sign": 0.3899712981656194, "keep": 0.6220295019447803, "leak": 1.5652132836985402, "replacement": 1.2771189957857132}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.925746440887451, "loss_scale": 4.0, "retries": 0, "seconds": 242.9, "max_memory_gb": 27.92, "pair": 0.5414483398199081, "sign": 0.4844621140509844, "keep": 0.19873314164578915, "leak": 1.3183282911777496, "replacement": 1.2126280516386032}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 6.670689582824707, "loss_scale": 4.0, "retries": 0, "seconds": 244.4, "max_memory_gb": 27.92, "pair": 0.4074275866150856, "sign": 0.41290103644132614, "keep": 0.775314636528492, "leak": 2.7483919244259596, "replacement": 1.6886334121227264}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 5.130143642425537, "loss_scale": 4.0, "retries": 0, "seconds": 245.8, "max_memory_gb": 27.92, "pair": 0.34247449040412903, "sign": 0.11409724410623312, "keep": 0.14566484279930592, "leak": 1.1190714053809643, "replacement": 0.7521233931183815}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 2.8341524600982666, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 27.92, "pair": 0.575661800801754, "sign": 0.4179831091314554, "keep": 0.16450740420259535, "leak": 1.896458761766553, "replacement": 1.3410217463970184}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 4.878784656524658, "loss_scale": 4.0, "retries": 0, "seconds": 248.8, "max_memory_gb": 27.92, "pair": 0.3089998345822096, "sign": 0.08941757120192051, "keep": 0.26434325240552425, "leak": 4.277626875322312, "replacement": 1.5552869588136673}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 3.9145238399505615, "loss_scale": 4.0, "retries": 0, "seconds": 250.2, "max_memory_gb": 27.92, "pair": 0.34024539217352867, "sign": 0.39659283589571714, "keep": 0.3784014396369457, "leak": 2.268213503062725, "replacement": 1.2947958707809448}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 4.339803695678711, "loss_scale": 4.0, "retries": 0, "seconds": 251.7, "max_memory_gb": 27.92, "pair": 0.5039034709334373, "sign": 0.3022967167198658, "keep": 0.29198299162089825, "leak": 0.8899349123239517, "replacement": 1.0235270410776138}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.712006092071533, "loss_scale": 4.0, "retries": 0, "seconds": 253.2, "max_memory_gb": 27.92, "pair": 0.62059535831213, "sign": 0.5683944672346115, "keep": 0.3551945686340332, "leak": 1.61612682742998, "replacement": 1.486421525478363}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 3.111572027206421, "loss_scale": 4.0, "retries": 0, "seconds": 254.6, "max_memory_gb": 27.92, "pair": 0.3663810268044472, "sign": 0.16360763553529978, "keep": 0.18366330116987228, "leak": 2.080526935867965, "replacement": 1.0601481944322586}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 5.009747505187988, "loss_scale": 4.0, "retries": 0, "seconds": 256.0, "max_memory_gb": 27.92, "pair": 0.5481803119182587, "sign": 0.3442468252032995, "keep": 0.3119359673000872, "leak": 0.6948166699148715, "replacement": 1.0499758571386337}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.107419490814209, "loss_scale": 4.0, "retries": 0, "seconds": 257.5, "max_memory_gb": 27.92, "pair": 0.649363324046135, "sign": 0.5722992867231369, "keep": 0.32317775930278003, "leak": 1.3634035220893566, "replacement": 1.4379527419805527}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 2.453667640686035, "loss_scale": 4.0, "retries": 0, "seconds": 258.8, "max_memory_gb": 27.92, "pair": 0.7041068207472563, "sign": 0.7614123579114676, "keep": 0.6076923805521801, "leak": 0.8218586160364794, "replacement": 1.5941238403320312}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 5.445296287536621, "loss_scale": 4.0, "retries": 0, "seconds": 260.2, "max_memory_gb": 27.92, "pair": 0.639087438583374, "sign": 0.5277860853821039, "keep": 0.7006356846541166, "leak": 0.17838185839354992, "replacement": 1.2978938221931458}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 2.8859922885894775, "loss_scale": 4.0, "retries": 0, "seconds": 261.6, "max_memory_gb": 27.92, "pair": 0.29354529455304146, "sign": 0.1338361739180982, "keep": 0.22637304663658142, "leak": 2.8784780306741595, "replacement": 1.1932693496346474}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.7788429260253906, "loss_scale": 4.0, "retries": 0, "seconds": 263.0, "max_memory_gb": 27.92, "pair": 0.6779220849275589, "sign": 0.6025860197842121, "keep": 0.3586729026865214, "leak": 0.6198962172493339, "replacement": 1.3135255575180054}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 4.6953558921813965, "loss_scale": 4.0, "retries": 0, "seconds": 264.4, "max_memory_gb": 27.92, "pair": 0.3127294331789017, "sign": 0.15617684368044138, "keep": 0.041404867835808545, "leak": 1.4246994853019714, "replacement": 0.7676951736211777}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 5.20165491104126, "loss_scale": 4.0, "retries": 0, "seconds": 265.9, "max_memory_gb": 27.92, "pair": 0.5119736604392529, "sign": 0.3594448519870639, "keep": 0.21228946250630543, "leak": 0.6955825984477997, "replacement": 0.9717364609241486}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 6.960307598114014, "loss_scale": 4.0, "retries": 0, "seconds": 267.4, "max_memory_gb": 27.92, "pair": 0.37522072345018387, "sign": 0.26163829024881124, "keep": 0.07755676470696926, "leak": 3.3667461993172765, "replacement": 1.3865047991275787}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 2.855854034423828, "loss_scale": 4.0, "retries": 0, "seconds": 268.9, "max_memory_gb": 27.92, "pair": 0.3860817477107048, "sign": 0.23498146422207355, "keep": 0.1837142383446917, "leak": 0.8519723564386368, "replacement": 0.8084226697683334}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 3.3236021995544434, "loss_scale": 4.0, "retries": 0, "seconds": 270.5, "max_memory_gb": 27.92, "pair": 0.33566558361053467, "sign": 0.07410695031285286, "keep": 0.16746208723634481, "leak": 1.4562178842024878, "replacement": 0.8205045908689499}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 2.6844308376312256, "loss_scale": 4.0, "retries": 0, "seconds": 271.9, "max_memory_gb": 27.92, "pair": 0.6556978821754456, "sign": 0.5849581360816956, "keep": 0.25601747271139175, "leak": 0.2594078550755512, "replacement": 1.1410376578569412}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.755528211593628, "loss_scale": 4.0, "retries": 0, "seconds": 273.3, "max_memory_gb": 27.92, "pair": 0.40568653866648674, "sign": 0.23428360745310783, "keep": 0.10331507865339518, "leak": 1.433144232723862, "replacement": 0.9327719286084175}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 2.008619785308838, "loss_scale": 4.0, "retries": 0, "seconds": 274.8, "max_memory_gb": 27.92, "pair": 0.2767917886376381, "sign": 0.19046671129763126, "keep": 0.09858326055109501, "leak": 1.2749934047460556, "replacement": 0.7400651276111603}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 2.9217875003814697, "loss_scale": 4.0, "retries": 0, "seconds": 276.1, "max_memory_gb": 27.92, "pair": 0.660210732370615, "sign": 0.6368458596989512, "keep": 0.631723141297698, "leak": 0.46020028553903103, "replacement": 1.4095453321933746}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 1.5840452909469604, "loss_scale": 4.0, "retries": 0, "seconds": 277.6, "max_memory_gb": 27.92, "pair": 0.626971960067749, "sign": 0.49904789589345455, "keep": 0.36130540166050196, "leak": 0.602144683674851, "replacement": 1.2076847404241562}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 4.756386756896973, "loss_scale": 4.0, "retries": 0, "seconds": 279.1, "max_memory_gb": 27.92, "pair": 0.4155435711145401, "sign": 0.1962762102484703, "keep": 0.15627337899059057, "leak": 0.5448883697390556, "replacement": 0.7280404716730118}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 4.0828633308410645, "loss_scale": 4.0, "retries": 0, "seconds": 280.5, "max_memory_gb": 27.92, "pair": 0.7273704446852207, "sign": 0.7952909022569656, "keep": 0.7704004824627191, "leak": 0.7645344361662865, "replacement": 1.7013497948646545}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 5.969086647033691, "loss_scale": 4.0, "retries": 0, "seconds": 282.0, "max_memory_gb": 27.92, "pair": 0.5140619799494743, "sign": 0.36813079938292503, "keep": 0.26649938710033894, "leak": 2.301175310043618, "replacement": 1.4066709280014038}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 3.1904730796813965, "loss_scale": 4.0, "retries": 0, "seconds": 283.5, "max_memory_gb": 27.92, "pair": 0.2907288745045662, "sign": 0.11831157840788364, "keep": 0.08942510187625885, "leak": 0.672021672129631, "replacement": 0.5626026317477226}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 3.2445950508117676, "loss_scale": 4.0, "retries": 0, "seconds": 284.9, "max_memory_gb": 27.92, "pair": 0.6541071906685829, "sign": 0.5888031255453825, "keep": 0.2885847371071577, "leak": 0.3955282960087061, "replacement": 1.1916831731796265}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 2.2815632820129395, "loss_scale": 4.0, "retries": 0, "seconds": 286.4, "max_memory_gb": 27.92, "pair": 0.6268753707408905, "sign": 0.6159841753542423, "keep": 0.30464607290923595, "leak": 0.9104121942073107, "replacement": 1.3147935718297958}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 8.74718952178955, "loss_scale": 4.0, "retries": 0, "seconds": 287.9, "max_memory_gb": 27.92, "pair": 0.286952868103981, "sign": 0.13862737454473972, "keep": 0.23848482687026262, "leak": 2.223671570420265, "replacement": 1.0314268618822098}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 2.5465335845947266, "loss_scale": 4.0, "retries": 0, "seconds": 289.3, "max_memory_gb": 27.92, "pair": 0.6048914790153503, "sign": 0.6529796216636896, "keep": 0.45784357748925686, "leak": 0.4174180394038558, "replacement": 1.2646575719118118}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 2.281707763671875, "loss_scale": 4.0, "retries": 0, "seconds": 290.8, "max_memory_gb": 27.92, "pair": 0.5682807266712189, "sign": 0.6412341184914112, "keep": 0.31732981372624636, "leak": 0.7222152724862099, "replacement": 1.228116475045681}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 2.853466033935547, "loss_scale": 4.0, "retries": 0, "seconds": 292.2, "max_memory_gb": 27.92, "pair": 0.45522890239953995, "sign": 0.3340066932141781, "keep": 0.010424126754514873, "leak": 0.4308387488126755, "replacement": 0.7351540178060532}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 3.389482259750366, "loss_scale": 4.0, "retries": 0, "seconds": 293.7, "max_memory_gb": 27.92, "pair": 0.5600854679942131, "sign": 0.4931773152202368, "keep": 0.31812625750899315, "leak": 1.225635437418532, "replacement": 1.2721461653709412}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 5.982293605804443, "loss_scale": 4.0, "retries": 0, "seconds": 295.1, "max_memory_gb": 27.92, "pair": 0.5269595310091972, "sign": 0.5347744561731815, "keep": 0.27763209491968155, "leak": 1.5101003907620907, "replacement": 1.3106879144906998}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 5.113636493682861, "loss_scale": 4.0, "retries": 0, "seconds": 296.6, "max_memory_gb": 27.92, "pair": 0.3524680957198143, "sign": 0.343585517257452, "keep": 0.1786594024160877, "leak": 1.695582378655672, "replacement": 1.0374861359596252}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 4.415849685668945, "loss_scale": 4.0, "retries": 0, "seconds": 298.0, "max_memory_gb": 27.92, "pair": 0.4587165489792824, "sign": 0.43377984687685966, "keep": 0.1770969171775505, "leak": 2.0493143086787313, "replacement": 1.2764834612607956}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 7.713351726531982, "loss_scale": 4.0, "retries": 0, "seconds": 299.5, "max_memory_gb": 27.92, "pair": 0.9608386904001236, "sign": 1.0436130240559578, "keep": 1.4350320303346962, "leak": 0.2242816801299341, "replacement": 2.256231665611267}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 7.337829113006592, "loss_scale": 4.0, "retries": 0, "seconds": 301.0, "max_memory_gb": 27.92, "pair": 0.2929546497762203, "sign": 0.08752182684838772, "keep": 0.07207493996247649, "leak": 2.16367329005152, "replacement": 0.9136713445186615}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 2.524792194366455, "loss_scale": 4.0, "retries": 0, "seconds": 302.5, "max_memory_gb": 27.92, "pair": 0.5683253556489944, "sign": 0.6070172814652324, "keep": 0.4310515820980072, "leak": 1.1848673975291604, "replacement": 1.383576676249504}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 2.744739294052124, "loss_scale": 4.0, "retries": 0, "seconds": 303.9, "max_memory_gb": 27.92, "pair": 0.43775099143385887, "sign": 0.4821274718269706, "keep": 0.4602986876270734, "leak": 1.8581139957532287, "replacement": 1.373492568731308}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 5.878167152404785, "loss_scale": 4.0, "retries": 0, "seconds": 305.4, "max_memory_gb": 27.92, "pair": 0.5718168830499053, "sign": 0.5554377604275942, "keep": 0.3855358958244324, "leak": 0.6702502795960754, "replacement": 1.2098662108182907}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 7.366751194000244, "loss_scale": 4.0, "retries": 0, "seconds": 306.9, "max_memory_gb": 27.92, "pair": 0.5421989113092422, "sign": 0.6259896494448185, "keep": 0.5921191796660423, "leak": 0.4529250480700284, "replacement": 1.2644846141338348}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 9.254046440124512, "loss_scale": 4.0, "retries": 0, "seconds": 308.4, "max_memory_gb": 27.92, "pair": 0.512195223942399, "sign": 0.46555707789957523, "keep": 0.7329212967306376, "leak": 1.8031325284391642, "replacement": 1.5622175186872482}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 3.4831900596618652, "loss_scale": 4.0, "retries": 0, "seconds": 309.8, "max_memory_gb": 27.92, "pair": 0.6515041068196297, "sign": 0.6493658199906349, "keep": 0.6495836116373539, "leak": 0.22264932846883312, "replacement": 1.3566411286592484}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 4.627473831176758, "loss_scale": 4.0, "retries": 0, "seconds": 311.3, "max_memory_gb": 27.92, "pair": 0.8269630372524261, "sign": 0.9670927450060844, "keep": 0.9777512049186043, "leak": 0.38952215245808475, "replacement": 1.896765559911728}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 2.7080211639404297, "loss_scale": 4.0, "retries": 0, "seconds": 312.8, "max_memory_gb": 27.92, "pair": 0.4101625457406044, "sign": 0.3466045409440994, "keep": 0.43710610643029213, "leak": 0.5371793592348695, "replacement": 0.9363127127289772}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 4.753688812255859, "loss_scale": 4.0, "retries": 0, "seconds": 314.3, "max_memory_gb": 27.92, "pair": 0.32067851163446903, "sign": 0.26206492725759745, "keep": 0.10812511155381799, "leak": 1.7484313622117043, "replacement": 0.9428813755512238}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 3.125364303588867, "loss_scale": 4.0, "retries": 0, "seconds": 315.7, "max_memory_gb": 27.92, "pair": 0.8398526459932327, "sign": 1.0282713025808334, "keep": 0.31725689116865396, "leak": 0.1807161969548261, "replacement": 1.5577957779169083}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 5.730081081390381, "loss_scale": 4.0, "retries": 0, "seconds": 317.2, "max_memory_gb": 27.92, "pair": 0.5014721378684044, "sign": 0.5492990836501122, "keep": 0.1652558408677578, "leak": 0.452135633706348, "replacement": 0.9717834740877151}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 3.3868043422698975, "loss_scale": 4.0, "retries": 0, "seconds": 318.6, "max_memory_gb": 27.92, "pair": 0.41546791791915894, "sign": 0.5197935029864311, "keep": 0.2713790331617929, "leak": 0.7466709017753601, "replacement": 0.9977219253778458}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 3.5179126262664795, "loss_scale": 4.0, "retries": 0, "seconds": 320.0, "max_memory_gb": 27.92, "pair": 0.29531776905059814, "sign": 0.29393526539206505, "keep": 0.14315207581967115, "leak": 1.3542021363973618, "replacement": 0.8524119704961777}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 4.812281131744385, "loss_scale": 4.0, "retries": 0, "seconds": 321.5, "max_memory_gb": 27.92, "pair": 0.3497737869620323, "sign": 0.20589096564799547, "keep": 0.5722535457462072, "leak": 1.6785714885336347, "replacement": 1.1584888994693756}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 3.271240711212158, "loss_scale": 4.0, "retries": 0, "seconds": 323.0, "max_memory_gb": 27.92, "pair": 0.6204440966248512, "sign": 0.6662655547261238, "keep": 0.36783831065986305, "leak": 0.25595819612863124, "replacement": 1.2014855593442917}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 6.429602146148682, "loss_scale": 4.0, "retries": 0, "seconds": 324.4, "max_memory_gb": 27.92, "pair": 0.39962346479296684, "sign": 0.4277372118085623, "keep": 0.17043552588438615, "leak": 1.3102161059869104, "replacement": 1.0262638479471207}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 7.977824687957764, "loss_scale": 4.0, "retries": 0, "seconds": 325.9, "max_memory_gb": 27.92, "pair": 0.47006773948669434, "sign": 0.5149569008499384, "keep": 0.15421291068196297, "leak": 1.5652628298848867, "replacement": 1.1959683001041412}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 3.0048389434814453, "loss_scale": 4.0, "retries": 0, "seconds": 327.4, "max_memory_gb": 27.92, "pair": 0.34738533571362495, "sign": 0.32954778894782066, "keep": 0.26832661777734756, "leak": 0.6007671914994717, "replacement": 0.7965143471956253}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.9893434047698975, "loss_scale": 4.0, "retries": 0, "seconds": 328.8, "max_memory_gb": 27.92, "pair": 0.362200528383255, "sign": 0.18597088381648064, "keep": 0.0283657907275483, "leak": 1.1033777743577957, "replacement": 0.7452132850885391}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 3.515810489654541, "loss_scale": 4.0, "retries": 0, "seconds": 330.2, "max_memory_gb": 27.92, "pair": 0.35392507910728455, "sign": 0.40595027059316635, "keep": 0.07503975555300713, "leak": 0.7496010512113571, "replacement": 0.7818203717470169}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 4.503273963928223, "loss_scale": 4.0, "retries": 0, "seconds": 331.7, "max_memory_gb": 27.92, "pair": 0.5896158069372177, "sign": 0.6852821009233594, "keep": 0.49040508456528187, "leak": 0.7590375058352947, "replacement": 1.3672187626361847}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 2.3751001358032227, "loss_scale": 4.0, "retries": 0, "seconds": 333.1, "max_memory_gb": 27.92, "pair": 0.6258249990642071, "sign": 0.6820901371538639, "keep": 0.6310899831005372, "leak": 0.3346135433530435, "replacement": 1.3660684525966644}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 2.8607702255249023, "loss_scale": 4.0, "retries": 0, "seconds": 334.5, "max_memory_gb": 27.92, "pair": 0.6706112697720528, "sign": 0.7920293882489204, "keep": 0.485823069117032, "leak": 0.36249773763120174, "replacement": 1.4001619219779968}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 5.514828681945801, "loss_scale": 4.0, "retries": 0, "seconds": 335.9, "max_memory_gb": 27.92, "pair": 0.26764850690960884, "sign": 0.43458876945078373, "keep": 0.08836966566741467, "leak": 0.9864282142370939, "replacement": 0.7757347673177719}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 7.278069496154785, "loss_scale": 4.0, "retries": 0, "seconds": 337.4, "max_memory_gb": 27.92, "pair": 0.337135910987854, "sign": 0.1958793718367815, "keep": 0.36554601043462753, "leak": 1.325532583752647, "replacement": 0.949231743812561}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 2.9182686805725098, "loss_scale": 4.0, "retries": 0, "seconds": 338.8, "max_memory_gb": 27.92, "pair": 0.4640035517513752, "sign": 0.4777497835457325, "keep": 0.3843591521726921, "leak": 1.164726721122861, "replacement": 1.1862397491931915}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 3.971522569656372, "loss_scale": 4.0, "retries": 0, "seconds": 340.3, "max_memory_gb": 27.92, "pair": 0.620087057352066, "sign": 0.6650778949260712, "keep": 0.5846264585852623, "leak": 0.477451115148142, "replacement": 1.3643020391464233}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 3.633151054382324, "loss_scale": 4.0, "retries": 0, "seconds": 341.7, "max_memory_gb": 27.92, "pair": 0.5111369714140892, "sign": 0.46873993799090385, "keep": 0.16519829351454973, "leak": 0.3545381754229311, "replacement": 0.9167406260967255}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 6.047452926635742, "loss_scale": 4.0, "retries": 0, "seconds": 343.1, "max_memory_gb": 27.92, "pair": 0.5775037109851837, "sign": 0.4662494957447052, "keep": 0.46246114425593987, "leak": 0.3681219061400043, "replacement": 1.1338894963264465}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 6.923295497894287, "loss_scale": 4.0, "retries": 0, "seconds": 344.6, "max_memory_gb": 27.92, "pair": 0.34709547087550163, "sign": 0.17715162597596645, "keep": 0.0989031670615077, "leak": 1.4104182422161102, "replacement": 0.8377274349331856}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 4.350046634674072, "loss_scale": 4.0, "retries": 0, "seconds": 346.0, "max_memory_gb": 27.92, "pair": 0.5409803688526154, "sign": 0.5870394017547369, "keep": 0.44480600766837597, "leak": 3.062606705352664, "replacement": 1.8225547671318054}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 3.247220039367676, "loss_scale": 4.0, "retries": 0, "seconds": 347.4, "max_memory_gb": 27.92, "pair": 0.37560537829995155, "sign": 0.32878952473402023, "keep": 0.1272934746230021, "leak": 0.5872879065573215, "replacement": 0.750468835234642}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 3.7185239791870117, "loss_scale": 4.0, "retries": 0, "seconds": 348.9, "max_memory_gb": 27.92, "pair": 0.5503467246890068, "sign": 0.6490740552544594, "keep": 0.22534491959959269, "leak": 0.40199939208546476, "replacement": 1.088056057691574}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 3.3419840335845947, "loss_scale": 4.0, "retries": 0, "seconds": 350.3, "max_memory_gb": 27.92, "pair": 0.5994836017489433, "sign": 0.6128574814647436, "keep": 0.27029301202856004, "leak": 0.3036559628963005, "replacement": 1.1169728115200996}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 5.702000141143799, "loss_scale": 4.0, "retries": 0, "seconds": 351.8, "max_memory_gb": 27.92, "pair": 0.31188082695007324, "sign": 0.24572970531880856, "keep": 0.10668373992666602, "leak": 2.5187010252848268, "replacement": 1.1177627891302109}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 7.017834186553955, "loss_scale": 4.0, "retries": 0, "seconds": 353.3, "max_memory_gb": 27.92, "pair": 0.6858908087015152, "sign": 0.8189205983653665, "keep": 0.7505453936755657, "leak": 0.5833931476809084, "replacement": 1.6164721325039864}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 2.909318208694458, "loss_scale": 4.0, "retries": 0, "seconds": 354.7, "max_memory_gb": 27.92, "pair": 0.8834257051348686, "sign": 0.9945696517825127, "keep": 0.6661073230206966, "leak": 0.19910412863828242, "replacement": 1.7635402381420135}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 3.642695188522339, "loss_scale": 4.0, "retries": 0, "seconds": 356.3, "max_memory_gb": 27.92, "pair": 0.21165157482028008, "sign": 0.1943240975961089, "keep": 0.07356974662980065, "leak": 1.3853268772363663, "replacement": 0.6919302195310593}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 3.590590476989746, "loss_scale": 4.0, "retries": 0, "seconds": 357.7, "max_memory_gb": 27.92, "pair": 0.3162106089293957, "sign": 0.32592789083719254, "keep": 0.15861816296819597, "leak": 0.9097249694168568, "replacement": 0.7859148681163788}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 3.672226905822754, "loss_scale": 4.0, "retries": 0, "seconds": 359.2, "max_memory_gb": 27.92, "pair": 0.24176554009318352, "sign": 0.14734310936182737, "keep": 0.15516920946538448, "leak": 0.505747489631176, "replacement": 0.5194585621356964}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 6.61675500869751, "loss_scale": 4.0, "retries": 0, "seconds": 360.6, "max_memory_gb": 27.92, "pair": 0.49868301302194595, "sign": 0.7221693247556686, "keep": 0.10598499979823828, "leak": 0.5867811117786914, "replacement": 1.0594554245471954}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 7.981259346008301, "loss_scale": 4.0, "retries": 0, "seconds": 362.1, "max_memory_gb": 27.92, "pair": 0.2970941551029682, "sign": 0.286094194278121, "keep": 0.07506579742766917, "leak": 1.7420501410961151, "replacement": 0.9131866991519928}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 4.2453179359436035, "loss_scale": 4.0, "retries": 0, "seconds": 363.6, "max_memory_gb": 27.92, "pair": 0.4286113977432251, "sign": 0.45359410159289837, "keep": 0.20441675779875368, "leak": 2.401950139152177, "replacement": 1.3581043481826782}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 3.806727409362793, "loss_scale": 4.0, "retries": 0, "seconds": 365.0, "max_memory_gb": 27.92, "pair": 0.24806120060384274, "sign": 0.158515065908432, "keep": 0.030713296495378017, "leak": 1.0477527379989624, "replacement": 0.6046135723590851}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 8.209921836853027, "loss_scale": 4.0, "retries": 0, "seconds": 366.5, "max_memory_gb": 27.92, "pair": 0.2860459750518203, "sign": 0.3097034227102995, "keep": 0.1762928935349919, "leak": 2.414408898446709, "replacement": 1.1326463371515274}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 6.977404594421387, "loss_scale": 4.0, "retries": 0, "seconds": 367.9, "max_memory_gb": 27.92, "pair": 0.33873188495635986, "sign": 0.37588984332978725, "keep": 0.32588599156588316, "leak": 1.8636513873934746, "replacement": 1.1555326282978058}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 3.635345220565796, "loss_scale": 4.0, "retries": 0, "seconds": 369.4, "max_memory_gb": 27.92, "pair": 0.5596410818397999, "sign": 0.7183416783809662, "keep": 0.2668922608718276, "leak": 0.743188071064651, "replacement": 1.2380551099777222}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 5.193064212799072, "loss_scale": 4.0, "retries": 0, "seconds": 370.8, "max_memory_gb": 27.92, "pair": 0.46786913834512234, "sign": 0.5622003450989723, "keep": 0.12387251108884811, "leak": 1.0911227762699127, "replacement": 1.0836862474679947}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 6.364336967468262, "loss_scale": 4.0, "retries": 0, "seconds": 372.3, "max_memory_gb": 27.92, "pair": 0.5074222236871719, "sign": 0.5202386183664203, "keep": 0.1400287527940236, "leak": 0.7415744343306869, "replacement": 1.022949531674385}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 6.0536675453186035, "loss_scale": 4.0, "retries": 0, "seconds": 373.7, "max_memory_gb": 27.92, "pair": 0.1714789979159832, "sign": 0.13655004557222128, "keep": 0.15357747836969793, "leak": 2.193505585193634, "replacement": 0.8649191409349442}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.2422196865081787, "loss_scale": 4.0, "retries": 0, "seconds": 375.2, "max_memory_gb": 27.92, "pair": 0.2797507084906101, "sign": 0.23500699549913406, "keep": 0.010883700218982995, "leak": 0.9713732749223709, "replacement": 0.6455393582582474}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 4.418679714202881, "loss_scale": 4.0, "retries": 0, "seconds": 376.6, "max_memory_gb": 27.92, "pair": 0.3099918980151415, "sign": 0.3613796140998602, "keep": 0.262295319698751, "leak": 1.203022986650467, "replacement": 0.9225850999355316}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 4.367606163024902, "loss_scale": 4.0, "retries": 0, "seconds": 378.1, "max_memory_gb": 27.92, "pair": 0.37356734462082386, "sign": 0.3337773960083723, "keep": 0.24181413557380438, "leak": 1.2658444863627665, "replacement": 0.977824255824089}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 5.31065559387207, "loss_scale": 4.0, "retries": 0, "seconds": 379.5, "max_memory_gb": 27.92, "pair": 0.5023116320371628, "sign": 0.5630766628310084, "keep": 0.5627356645418331, "leak": 0.481598187237978, "replacement": 1.1856173127889633}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 3.6603755950927734, "loss_scale": 4.0, "retries": 0, "seconds": 381.0, "max_memory_gb": 27.92, "pair": 0.36969433538615704, "sign": 0.3653632178902626, "keep": 0.4233710318803787, "leak": 1.781067369505763, "replacement": 1.2093282788991928}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 7.446108341217041, "loss_scale": 4.0, "retries": 0, "seconds": 382.4, "max_memory_gb": 27.92, "pair": 0.24163874983787537, "sign": 0.13888406939804554, "keep": 0.6200151313096285, "leak": 3.320844627916813, "replacement": 1.4512995034456253}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 3.8636996746063232, "loss_scale": 4.0, "retries": 0, "seconds": 383.9, "max_memory_gb": 27.92, "pair": 0.38523781299591064, "sign": 0.3697144892066717, "keep": 0.13912494492251426, "leak": 1.5565313559491187, "replacement": 1.0287903547286987}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 5.597983360290527, "loss_scale": 4.0, "retries": 0, "seconds": 385.3, "max_memory_gb": 27.92, "pair": 0.5929527170956135, "sign": 0.5730852354317904, "keep": 1.0631744302809238, "leak": 0.4985406845808029, "replacement": 1.5357177406549454}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.148634433746338, "loss_scale": 4.0, "retries": 0, "seconds": 386.8, "max_memory_gb": 27.92, "pair": 0.43160878121852875, "sign": 0.4550820575095713, "keep": 0.2703762366436422, "leak": 0.5794506394304335, "replacement": 0.9392005577683449}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 1.6615338325500488, "loss_scale": 4.0, "retries": 0, "seconds": 388.3, "max_memory_gb": 27.92, "pair": 0.6169662959873676, "sign": 0.7415461577475071, "keep": 0.3916515790624544, "leak": 0.5121411171785439, "replacement": 1.3116004467010498}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 5.456135272979736, "loss_scale": 4.0, "retries": 0, "seconds": 389.7, "max_memory_gb": 27.92, "pair": 0.4409221187233925, "sign": 0.48231243528425694, "keep": 1.22587331966497, "leak": 1.0941085503436625, "replacement": 1.5685421228408813}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 5.654835224151611, "loss_scale": 4.0, "retries": 0, "seconds": 391.2, "max_memory_gb": 27.92, "pair": 0.35617539286613464, "sign": 0.41944450326263905, "keep": 0.5351921105757356, "leak": 0.9513213969767094, "replacement": 1.0713240578770638}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 3.918661117553711, "loss_scale": 4.0, "retries": 0, "seconds": 392.7, "max_memory_gb": 27.92, "pair": 0.2865177020430565, "sign": 0.2807795973494649, "keep": 0.1441197758540511, "leak": 2.3660475593060255, "replacement": 1.090479277074337}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 5.05671501159668, "loss_scale": 4.0, "retries": 0, "seconds": 394.1, "max_memory_gb": 27.92, "pair": 0.2256545051932335, "sign": 0.06101410696282983, "keep": 0.10393469035625458, "leak": 0.5348351619904861, "replacement": 0.44183769822120667}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 5.62105655670166, "loss_scale": 4.0, "retries": 0, "seconds": 395.5, "max_memory_gb": 27.92, "pair": 0.36192090436816216, "sign": 0.1797170303761959, "keep": 0.07536042539868504, "leak": 0.26315470039844513, "replacement": 0.5552482753992081}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 5.2229108810424805, "loss_scale": 4.0, "retries": 0, "seconds": 397.0, "max_memory_gb": 27.92, "pair": 0.5545413456857204, "sign": 0.77545265853405, "keep": 0.17498981300741434, "leak": 0.42517000320367515, "replacement": 1.1360550969839096}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 4.731426239013672, "loss_scale": 4.0, "retries": 0, "seconds": 398.4, "max_memory_gb": 27.92, "pair": 0.465657826513052, "sign": 0.49064292199909687, "keep": 1.2967834174633026, "leak": 0.3102278457954526, "replacement": 1.4369279965758324}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 4.372552871704102, "loss_scale": 4.0, "retries": 0, "seconds": 399.9, "max_memory_gb": 27.92, "pair": 0.25046522356569767, "sign": 0.11960938246920705, "keep": 0.06375356251373887, "leak": 0.5996562810614705, "replacement": 0.4920607879757881}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 2.9070489406585693, "loss_scale": 4.0, "retries": 0, "seconds": 401.4, "max_memory_gb": 27.92, "pair": 0.3091325871646404, "sign": 0.2696725483983755, "keep": 0.034214644343592227, "leak": 0.5211513414978981, "replacement": 0.5913640409708023}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 3.967374086380005, "loss_scale": 4.0, "retries": 0, "seconds": 402.8, "max_memory_gb": 27.92, "pair": 0.49278195202350616, "sign": 0.5315267369151115, "keep": 0.36533506034174934, "leak": 0.5009678667411208, "replacement": 1.0664547830820084}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 6.859506130218506, "loss_scale": 4.0, "retries": 0, "seconds": 404.2, "max_memory_gb": 27.92, "pair": 0.3749486990272999, "sign": 0.2657693177461624, "keep": 0.11448092572391033, "leak": 2.058123731985688, "replacement": 1.0796047449111938}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 3.984981060028076, "loss_scale": 4.0, "retries": 0, "seconds": 405.6, "max_memory_gb": 27.92, "pair": 0.2534469086676836, "sign": 0.18173319473862648, "keep": 0.07332340104039758, "leak": 0.9061646594200283, "replacement": 0.6075163781642914}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.7361087799072266, "loss_scale": 4.0, "retries": 0, "seconds": 407.1, "max_memory_gb": 27.92, "pair": 0.31571851298213005, "sign": 0.30070529179647565, "keep": 0.4142353500938043, "leak": 0.8266289858147502, "replacement": 0.8798460811376572}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 4.462550640106201, "loss_scale": 4.0, "retries": 0, "seconds": 408.5, "max_memory_gb": 27.92, "pair": 0.2996153701096773, "sign": 0.32811327651143074, "keep": 0.1989317153347656, "leak": 0.5326013956218958, "replacement": 0.6962882354855537}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 6.4858784675598145, "loss_scale": 4.0, "retries": 0, "seconds": 410.0, "max_memory_gb": 27.92, "pair": 0.19596902839839458, "sign": 0.08195788133889437, "keep": 0.11235139984637499, "leak": 3.166390597820282, "replacement": 1.0847212970256805}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.807919979095459, "loss_scale": 4.0, "retries": 0, "seconds": 411.5, "max_memory_gb": 27.92, "pair": 0.2161248866468668, "sign": 0.21716385707259178, "keep": 0.10353159997612238, "leak": 1.026468388736248, "replacement": 0.6330897286534309}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 11.003463745117188, "loss_scale": 4.0, "retries": 0, "seconds": 413.0, "max_memory_gb": 27.92, "pair": 0.38079816848039627, "sign": 0.5565888173878193, "keep": 0.16848277766257524, "leak": 2.340408683226997, "replacement": 1.3284361213445663}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 8.560842514038086, "loss_scale": 4.0, "retries": 0, "seconds": 414.5, "max_memory_gb": 27.92, "pair": 0.3535539247095585, "sign": 0.5493403356522322, "keep": 0.033059311215765774, "leak": 2.564232751727104, "replacement": 1.2858119159936905}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 9.951269149780273, "loss_scale": 4.0, "retries": 0, "seconds": 416.0, "max_memory_gb": 27.92, "pair": 0.4256843291223049, "sign": 0.3986785192973912, "keep": 0.38831340335309505, "leak": 0.2705763557423779, "replacement": 0.8868243508040905}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 5.589053153991699, "loss_scale": 4.0, "retries": 0, "seconds": 417.4, "max_memory_gb": 27.92, "pair": 0.31739564798772335, "sign": 0.31550304824486375, "keep": 0.037926810793578625, "leak": 0.5318266206450062, "replacement": 0.6270672306418419}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 4.417564868927002, "loss_scale": 4.0, "retries": 0, "seconds": 418.9, "max_memory_gb": 27.92, "pair": 0.3243899457156658, "sign": 0.22570512862876058, "keep": 0.12178344093263149, "leak": 1.5412477515637875, "replacement": 0.8834461718797684}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 4.5938029289245605, "loss_scale": 4.0, "retries": 0, "seconds": 420.4, "max_memory_gb": 27.92, "pair": 0.42202754877507687, "sign": 0.488420567009598, "keep": 0.35241021891124547, "leak": 0.6909029744565487, "replacement": 1.0151686370372772}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 4.955799579620361, "loss_scale": 4.0, "retries": 0, "seconds": 421.8, "max_memory_gb": 27.92, "pair": 0.15901338309049606, "sign": 0.06322353053838015, "keep": 0.020759725710377097, "leak": 2.3854611963033676, "replacement": 0.7973702922463417}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 4.304324150085449, "loss_scale": 4.0, "retries": 0, "seconds": 423.2, "max_memory_gb": 27.92, "pair": 0.25204798579216003, "sign": 0.19345000432804227, "keep": 0.0721602882258594, "leak": 0.7637618407607079, "replacement": 0.5757935717701912}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 4.477555274963379, "loss_scale": 4.0, "retries": 0, "seconds": 424.7, "max_memory_gb": 27.92, "pair": 0.29906129837036133, "sign": 0.2752265501767397, "keep": 0.11539787275251001, "leak": 0.7803936991840601, "replacement": 0.6894719526171684}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 5.7508864402771, "loss_scale": 4.0, "retries": 0, "seconds": 426.2, "max_memory_gb": 27.92, "pair": 0.26626125164330006, "sign": 0.14805225655436516, "keep": 0.23281138390302658, "leak": 1.4198976652696729, "replacement": 0.8116675168275833}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 8.162921905517578, "loss_scale": 4.0, "retries": 0, "seconds": 427.5, "max_memory_gb": 27.92, "pair": 0.3067585863173008, "sign": 0.451453123241663, "keep": 0.426397904753685, "leak": 2.846386156976223, "replacement": 1.4572805911302567}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 7.415709972381592, "loss_scale": 4.0, "retries": 0, "seconds": 429.0, "max_memory_gb": 27.92, "pair": 0.5196620356291533, "sign": 0.6925051659345627, "keep": 0.522518876940012, "leak": 1.43472831917461, "replacement": 1.485856145620346}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 5.22481107711792, "loss_scale": 4.0, "retries": 0, "seconds": 430.4, "max_memory_gb": 27.92, "pair": 0.552200298756361, "sign": 0.8098303228616714, "keep": 0.2991908221738413, "leak": 0.7520261481404305, "replacement": 1.2947173938155174}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 6.115438938140869, "loss_scale": 4.0, "retries": 0, "seconds": 431.8, "max_memory_gb": 27.92, "pair": 0.5772735811769962, "sign": 0.7423317544162273, "keep": 0.85777307220269, "leak": 0.19771490967832506, "replacement": 1.426754742860794}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 8.159708976745605, "loss_scale": 4.0, "retries": 0, "seconds": 433.3, "max_memory_gb": 27.92, "pair": 0.42010145634412766, "sign": 0.1390826255083084, "keep": 0.08026228379458189, "leak": 0.8845068770460784, "replacement": 0.7509006261825562}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 9.2498140335083, "loss_scale": 4.0, "retries": 0, "seconds": 434.8, "max_memory_gb": 27.92, "pair": 0.7167944088578224, "sign": 0.8528176536783576, "keep": 0.508771812543273, "leak": 1.9270319092320278, "replacement": 1.8793472051620483}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 4.298454761505127, "loss_scale": 4.0, "retries": 0, "seconds": 436.2, "max_memory_gb": 27.92, "pair": 0.3855723366141319, "sign": 0.41075230948626995, "keep": 0.21525371260941029, "leak": 0.505370127139031, "replacement": 0.8249178826808929}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 6.5634050369262695, "loss_scale": 4.0, "retries": 0, "seconds": 437.6, "max_memory_gb": 27.92, "pair": 0.2221682071685791, "sign": 0.06029067747294903, "keep": 0.100975695066154, "leak": 0.5478046331554651, "replacement": 0.4397525489330292}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 10.825055122375488, "loss_scale": 4.0, "retries": 0, "seconds": 439.1, "max_memory_gb": 27.92, "pair": 0.3468644954264164, "sign": 0.26391926081851125, "keep": 0.06697525084018707, "leak": 2.016953431069851, "replacement": 1.0165500938892365}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 6.562521934509277, "loss_scale": 4.0, "retries": 0, "seconds": 440.5, "max_memory_gb": 27.92, "pair": 0.43293861858546734, "sign": 0.5108816213905811, "keep": 0.07710453122854233, "leak": 1.4778687451034784, "replacement": 1.0963988676667213}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 4.981043338775635, "loss_scale": 4.0, "retries": 0, "seconds": 442.0, "max_memory_gb": 27.92, "pair": 0.3144271597266197, "sign": 0.15204500034451485, "keep": 0.10216366592794657, "leak": 1.2310517467558384, "replacement": 0.7492944300174713}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 7.5136399269104, "loss_scale": 4.0, "retries": 0, "seconds": 443.5, "max_memory_gb": 27.92, "pair": 0.23253073543310165, "sign": 0.05220566876232624, "keep": 0.08558501722291112, "leak": 0.7059043286246833, "replacement": 0.4779021590948105}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 7.170452117919922, "loss_scale": 4.0, "retries": 0, "seconds": 444.9, "max_memory_gb": 27.92, "pair": 0.5007235296070576, "sign": 0.40777575271204114, "keep": 0.2570898325648159, "leak": 0.5904646984126884, "replacement": 0.9807724878191948}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 6.650789260864258, "loss_scale": 4.0, "retries": 0, "seconds": 446.4, "max_memory_gb": 27.92, "pair": 0.21215764433145523, "sign": 0.42250197008252144, "keep": 0.06628285977058113, "leak": 1.5266552865505219, "replacement": 0.8382139056921005}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 10.23591423034668, "loss_scale": 4.0, "retries": 0, "seconds": 447.8, "max_memory_gb": 27.92, "pair": 0.6822132132947445, "sign": 0.7311156746000051, "keep": 0.47611564584076405, "leak": 0.9433019183343276, "replacement": 1.5216543674468994}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 4.297537326812744, "loss_scale": 4.0, "retries": 0, "seconds": 449.2, "max_memory_gb": 27.92, "pair": 0.28590779192745686, "sign": 0.1276199882850051, "keep": 0.04531989758834243, "leak": 1.503484033048153, "replacement": 0.7482487484812737}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 2.773747205734253, "loss_scale": 4.0, "retries": 0, "seconds": 450.6, "max_memory_gb": 27.92, "pair": 0.37767794728279114, "sign": 0.4275977313518524, "keep": 0.2141979379230179, "leak": 0.653856985270977, "replacement": 0.8620400279760361}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 5.455564022064209, "loss_scale": 4.0, "retries": 0, "seconds": 452.1, "max_memory_gb": 27.92, "pair": 0.28642624616622925, "sign": 0.20204742345958948, "keep": 0.03952393843792379, "leak": 0.3415862701367587, "replacement": 0.4926084950566292}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 9.08503246307373, "loss_scale": 4.0, "retries": 0, "seconds": 453.6, "max_memory_gb": 27.92, "pair": 0.44249717704951763, "sign": 0.5413175043649971, "keep": 0.5105994008481503, "leak": 0.8673408590257168, "replacement": 1.185290902853012}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 7.91263484954834, "loss_scale": 4.0, "retries": 0, "seconds": 455.1, "max_memory_gb": 27.92, "pair": 0.34597716853022575, "sign": 0.39935474237427115, "keep": 0.35458893119357526, "leak": 1.220993123948574, "replacement": 1.0281973108649254}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 10.791219711303711, "loss_scale": 4.0, "retries": 0, "seconds": 456.6, "max_memory_gb": 27.92, "pair": 0.26218599639832973, "sign": 0.24394777417182922, "keep": 0.08600378688424826, "leak": 1.6823978801257908, "replacement": 0.847761258482933}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 4.852171421051025, "loss_scale": 4.0, "retries": 0, "seconds": 458.0, "max_memory_gb": 27.92, "pair": 0.5889441333711147, "sign": 0.7316589285619557, "keep": 0.3434368968009949, "leak": 1.0321949448916712, "replacement": 1.3845407664775848}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 5.378911972045898, "loss_scale": 4.0, "retries": 0, "seconds": 459.4, "max_memory_gb": 27.92, "pair": 0.40400348231196404, "sign": 0.4474601075053215, "keep": 0.27837675623595715, "leak": 1.3813333101570606, "replacement": 1.1122552454471588}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 5.583462715148926, "loss_scale": 4.0, "retries": 0, "seconds": 460.8, "max_memory_gb": 27.92, "pair": 0.2674327604472637, "sign": 0.2551639750599861, "keep": 0.05484382715076208, "leak": 1.2615731060504913, "replacement": 0.7378299385309219}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 4.3681745529174805, "loss_scale": 4.0, "retries": 0, "seconds": 462.3, "max_memory_gb": 27.92, "pair": 0.3682303074747324, "sign": 0.31866728514432907, "keep": 0.1690778974443674, "leak": 0.36391680501401424, "replacement": 0.7030820921063423}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 4.389791488647461, "loss_scale": 4.0, "retries": 0, "seconds": 463.7, "max_memory_gb": 27.92, "pair": 0.37468146905303, "sign": 0.39207486622035503, "keep": 0.2797807788592763, "leak": 0.6460556145757437, "replacement": 0.872123196721077}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 7.986855983734131, "loss_scale": 4.0, "retries": 0, "seconds": 465.1, "max_memory_gb": 27.92, "pair": 0.3469879478216171, "sign": 0.24998347274959087, "keep": 0.06432520516682416, "leak": 0.7631414432544261, "replacement": 0.6949276477098465}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 7.111616611480713, "loss_scale": 4.0, "retries": 0, "seconds": 466.5, "max_memory_gb": 27.92, "pair": 0.4242113381624222, "sign": 0.4350016564130783, "keep": 0.06068515498191118, "leak": 0.2621472647297196, "replacement": 0.7375915497541428}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 12.530536651611328, "loss_scale": 4.0, "retries": 0, "seconds": 468.0, "max_memory_gb": 27.92, "pair": 0.1177490334957838, "sign": 0.08921088092029095, "keep": 0.06209235265851021, "leak": 4.048623323440552, "replacement": 1.2055564522743225}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 5.642364501953125, "loss_scale": 4.0, "retries": 0, "seconds": 469.4, "max_memory_gb": 27.92, "pair": 0.1468994151800871, "sign": 0.073547069914639, "keep": 0.08085899602156132, "leak": 0.8966236435808241, "replacement": 0.4482583627104759}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 5.9148783683776855, "loss_scale": 4.0, "retries": 0, "seconds": 470.9, "max_memory_gb": 27.92, "pair": 0.6722152456641197, "sign": 0.6236565290018916, "keep": 0.48623091308400035, "leak": 0.11313684136257507, "replacement": 1.2554431408643723}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 6.19317626953125, "loss_scale": 4.0, "retries": 0, "seconds": 472.4, "max_memory_gb": 27.92, "pair": 0.4593472518026829, "sign": 0.5394910201430321, "keep": 0.19343038788065314, "leak": 0.8610695376992226, "replacement": 1.0410753339529037}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 7.897171497344971, "loss_scale": 4.0, "retries": 0, "seconds": 473.9, "max_memory_gb": 27.92, "pair": 0.21155964955687523, "sign": 0.10920335398986936, "keep": 0.44166798144578934, "leak": 0.6027203747071326, "replacement": 0.6376754269003868}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 6.6015191078186035, "loss_scale": 4.0, "retries": 0, "seconds": 475.3, "max_memory_gb": 27.92, "pair": 0.5714621767401695, "sign": 0.6586796343326569, "keep": 0.15875403711106628, "leak": 0.34601655951701105, "replacement": 1.0666831135749817}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 6.700589179992676, "loss_scale": 4.0, "retries": 0, "seconds": 476.7, "max_memory_gb": 27.92, "pair": 0.3317260444164276, "sign": 0.4172239899635315, "keep": 0.35384593077469617, "leak": 1.0326262786984444, "replacement": 0.9754175618290901}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 4.878139019012451, "loss_scale": 4.0, "retries": 0, "seconds": 478.2, "max_memory_gb": 27.92, "pair": 0.46506450325250626, "sign": 0.5059302332811058, "keep": 0.30740583955775946, "leak": 0.5871609458699822, "replacement": 1.0185227543115616}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 7.233245849609375, "loss_scale": 4.0, "retries": 0, "seconds": 479.6, "max_memory_gb": 27.92, "pair": 0.646176666021347, "sign": 0.6350106187164783, "keep": 0.5041388488607481, "leak": 1.4125262713059783, "replacement": 1.5688829720020294}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 6.710005760192871, "loss_scale": 4.0, "retries": 0, "seconds": 481.1, "max_memory_gb": 27.92, "pair": 0.44171667098999023, "sign": 0.36314836516976357, "keep": 0.40249295049579814, "leak": 0.799179682508111, "replacement": 1.0243322551250458}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 4.831238269805908, "loss_scale": 4.0, "retries": 0, "seconds": 482.5, "max_memory_gb": 27.92, "pair": 0.6950079798698425, "sign": 0.8012242242693901, "keep": 0.5449115005321801, "leak": 0.2102966511969413, "replacement": 1.4206499680876732}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 7.631712436676025, "loss_scale": 4.0, "retries": 0, "seconds": 484.0, "max_memory_gb": 27.92, "pair": 0.48771727085113525, "sign": 0.5643512234091759, "keep": 0.4989672312512994, "leak": 1.740905466281447, "replacement": 1.4546028226613998}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 7.549850940704346, "loss_scale": 4.0, "retries": 0, "seconds": 485.5, "max_memory_gb": 27.92, "pair": 0.5252225957810879, "sign": 0.5672761984169483, "keep": 0.7943782133224886, "leak": 0.3019456444308162, "replacement": 1.2815361618995667}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 8.12503433227539, "loss_scale": 4.0, "retries": 0, "seconds": 486.9, "max_memory_gb": 27.92, "pair": 0.2080482877790928, "sign": 0.12568961828947067, "keep": 0.08536712382920086, "leak": 0.891553798574023, "replacement": 0.5364651009440422}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 13.76567268371582, "loss_scale": 4.0, "retries": 0, "seconds": 488.4, "max_memory_gb": 27.92, "pair": 0.6403561905026436, "sign": 0.6919096373021603, "keep": 0.3878699392080307, "leak": 1.543191863456741, "replacement": 1.566043883562088}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 6.476395606994629, "loss_scale": 4.0, "retries": 0, "seconds": 489.9, "max_memory_gb": 27.92, "pair": 0.4742053411900997, "sign": 0.5540369618684053, "keep": 0.37255136179737747, "leak": 0.739448731765151, "replacement": 1.1223617196083069}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 7.491131782531738, "loss_scale": 4.0, "retries": 0, "seconds": 491.4, "max_memory_gb": 27.92, "pair": 0.5002341214567423, "sign": 0.7382785668596625, "keep": 0.2786979943048209, "leak": 1.3946728389710188, "replacement": 1.357390582561493}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 9.3992280960083, "loss_scale": 4.0, "retries": 0, "seconds": 492.8, "max_memory_gb": 27.92, "pair": 0.3660312369465828, "sign": 0.3726705084554851, "keep": 0.7235498480149545, "leak": 0.509841931052506, "replacement": 1.0416018776595592}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 11.949189186096191, "loss_scale": 4.0, "retries": 0, "seconds": 494.3, "max_memory_gb": 27.92, "pair": 0.2696648817509413, "sign": 0.31604779278859496, "keep": 0.34057969658169895, "leak": 1.6813597325235605, "replacement": 1.0183185636997223}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 6.562464714050293, "loss_scale": 4.0, "retries": 0, "seconds": 495.7, "max_memory_gb": 27.92, "pair": 0.5422315765172243, "sign": 0.6465809941291809, "keep": 0.5059410789981484, "leak": 0.7848046571016312, "replacement": 1.314693734049797}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 12.384537696838379, "loss_scale": 4.0, "retries": 0, "seconds": 497.2, "max_memory_gb": 27.92, "pair": 0.4710199050605297, "sign": 0.6175339613109827, "keep": 0.023344623448792845, "leak": 0.42271455749869347, "replacement": 0.8971378058195114}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 5.610244274139404, "loss_scale": 4.0, "retries": 0, "seconds": 498.6, "max_memory_gb": 27.92, "pair": 0.2056137565523386, "sign": 0.17439585085958242, "keep": 0.019725037273019552, "leak": 0.9547879677265882, "replacement": 0.5413711965084076}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 6.90566873550415, "loss_scale": 4.0, "retries": 0, "seconds": 500.0, "max_memory_gb": 27.92, "pair": 0.3911747671663761, "sign": 0.4298776420764625, "keep": 0.2914141418877989, "leak": 0.6096163258771412, "replacement": 0.90422473102808}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 8.571516990661621, "loss_scale": 4.0, "retries": 0, "seconds": 501.5, "max_memory_gb": 27.92, "pair": 0.3911399804055691, "sign": 0.49354162253439426, "keep": 0.5155920044635423, "leak": 1.8698719926178455, "replacement": 1.36317478120327}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 2.872232437133789, "loss_scale": 4.0, "retries": 0, "seconds": 503.0, "max_memory_gb": 27.92, "pair": 0.3899872526526451, "sign": 0.3881260873749852, "keep": 0.2850805353373289, "leak": 0.5419925209134817, "replacement": 0.862088680267334}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 5.733999252319336, "loss_scale": 4.0, "retries": 0, "seconds": 504.5, "max_memory_gb": 27.92, "pair": 0.20949052460491657, "sign": 0.1397867244668305, "keep": 0.016469750669784844, "leak": 2.0811460725963116, "replacement": 0.8079052865505219}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 9.183232307434082, "loss_scale": 4.0, "retries": 0, "seconds": 506.0, "max_memory_gb": 27.92, "pair": 0.3585693808272481, "sign": 0.38312298990786076, "keep": 0.18038279912434518, "leak": 2.2078363803448156, "replacement": 1.1922813951969147}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 6.729488372802734, "loss_scale": 4.0, "retries": 0, "seconds": 507.4, "max_memory_gb": 27.92, "pair": 0.5072018019855022, "sign": 0.6871111541986465, "keep": 0.2226652018725872, "leak": 0.4579005481209606, "replacement": 1.0765650868415833}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 3.8847293853759766, "loss_scale": 4.0, "retries": 0, "seconds": 508.9, "max_memory_gb": 27.92, "pair": 0.17692431900650263, "sign": 0.08707014936953783, "keep": 0.055812896229326725, "leak": 0.6851364243775606, "replacement": 0.41964995115995407}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 11.640119552612305, "loss_scale": 4.0, "retries": 0, "seconds": 510.3, "max_memory_gb": 27.92, "pair": 0.42704994045197964, "sign": 0.6093169916421175, "keep": 0.5821099241729826, "leak": 1.0278516039252281, "replacement": 1.2797263115644455}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 5.601603984832764, "loss_scale": 4.0, "retries": 0, "seconds": 511.8, "max_memory_gb": 27.92, "pair": 0.3932919390499592, "sign": 0.4310174882411957, "keep": 0.3362169931642711, "leak": 0.7749751536175609, "replacement": 0.9706529751420021}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 3.9733188152313232, "loss_scale": 4.0, "retries": 0, "seconds": 513.2, "max_memory_gb": 27.92, "pair": 0.41193830221891403, "sign": 0.43701012805104256, "keep": 0.3135074316523969, "leak": 0.4299464338691905, "replacement": 0.8946836739778519}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 7.799130439758301, "loss_scale": 4.0, "retries": 0, "seconds": 514.7, "max_memory_gb": 27.92, "pair": 0.41726403683423996, "sign": 0.413245246745646, "keep": 0.25497791822999716, "leak": 1.8922131969593465, "replacement": 1.224428877234459}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 12.484423637390137, "loss_scale": 4.0, "retries": 0, "seconds": 516.2, "max_memory_gb": 27.92, "pair": 0.14965348318219185, "sign": 0.1537745576351881, "keep": 0.024129600264132023, "leak": 3.1313566267490387, "replacement": 1.0214447230100632}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 12.697197914123535, "loss_scale": 4.0, "retries": 0, "seconds": 517.7, "max_memory_gb": 27.92, "pair": 0.3149970918893814, "sign": 0.1291006759274751, "keep": 0.2237435095012188, "leak": 1.698266327381134, "replacement": 0.9159857630729675}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 7.503240585327148, "loss_scale": 4.0, "retries": 0, "seconds": 519.1, "max_memory_gb": 27.92, "pair": 0.6224640011787415, "sign": 0.7561200745403767, "keep": 0.6593625973910093, "leak": 0.4649908933788538, "replacement": 1.446453034877777}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 6.9745941162109375, "loss_scale": 4.0, "retries": 0, "seconds": 520.7, "max_memory_gb": 27.92, "pair": 0.3447608947753906, "sign": 0.6079976297914982, "keep": 0.08799270982854068, "leak": 1.2734676599502563, "replacement": 1.0111229717731476}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 6.038405418395996, "loss_scale": 4.0, "retries": 0, "seconds": 522.1, "max_memory_gb": 27.92, "pair": 0.22273258864879608, "sign": 0.2032439149916172, "keep": 0.054013137705624104, "leak": 1.563189536333084, "replacement": 0.7421584874391556}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 10.05644702911377, "loss_scale": 4.0, "retries": 0, "seconds": 523.6, "max_memory_gb": 27.92, "pair": 0.2936036242172122, "sign": 0.525618907995522, "keep": 0.058828720124438405, "leak": 1.4596724659204483, "replacement": 0.950745552778244}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 11.060029029846191, "loss_scale": 4.0, "retries": 0, "seconds": 525.0, "max_memory_gb": 27.92, "pair": 0.4982779137790203, "sign": 0.45176681224256754, "keep": 1.0562248770147562, "leak": 0.17350541707128286, "replacement": 1.2956501320004463}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 12.753893852233887, "loss_scale": 4.0, "retries": 0, "seconds": 526.5, "max_memory_gb": 27.92, "pair": 0.625624306499958, "sign": 0.9461289942264557, "keep": 0.09245309792459011, "leak": 0.27143268939107656, "replacement": 1.2127735167741776}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 6.429512023925781, "loss_scale": 4.0, "retries": 0, "seconds": 528.0, "max_memory_gb": 27.92, "pair": 0.2567793242633343, "sign": 0.05850249994546175, "keep": 0.08727661054581404, "leak": 0.933513056486845, "replacement": 0.5630471631884575}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 6.9805731773376465, "loss_scale": 4.0, "retries": 0, "seconds": 529.4, "max_memory_gb": 27.92, "pair": 0.46904897317290306, "sign": 0.4069896638393402, "keep": 0.2959535283735022, "leak": 0.09987791904337939, "replacement": 0.845490075647831}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 11.759057998657227, "loss_scale": 4.0, "retries": 0, "seconds": 530.9, "max_memory_gb": 27.92, "pair": 0.4532999061048031, "sign": 0.3822006434202194, "keep": 0.19181735813617706, "leak": 1.8585802987217903, "replacement": 1.2049539983272552}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 5.967008113861084, "loss_scale": 4.0, "retries": 0, "seconds": 532.4, "max_memory_gb": 27.92, "pair": 0.5033339560031891, "sign": 0.45395051687955856, "keep": 0.48830713611096144, "leak": 0.6900952607393265, "replacement": 1.1469865888357162}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 5.257540702819824, "loss_scale": 4.0, "retries": 0, "seconds": 533.8, "max_memory_gb": 27.92, "pair": 0.2511250600218773, "sign": 0.12991438433527946, "keep": 0.056010512984357774, "leak": 0.6905338577926159, "replacement": 0.5167209655046463}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 7.461726665496826, "loss_scale": 4.0, "retries": 0, "seconds": 535.2, "max_memory_gb": 27.92, "pair": 0.4895939715206623, "sign": 0.3864290895871818, "keep": 0.11084588803350925, "leak": 0.864054964215029, "replacement": 0.9542452245950699}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 9.296184539794922, "loss_scale": 4.0, "retries": 0, "seconds": 536.8, "max_memory_gb": 27.92, "pair": 0.2598078064620495, "sign": 0.2989598549902439, "keep": 0.8801740184426308, "leak": 0.6180705726146698, "replacement": 1.0038923919200897}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.420310020446777, "loss_scale": 4.0, "retries": 0, "seconds": 538.2, "max_memory_gb": 27.92, "pair": 0.14786400273442268, "sign": 0.06260296516120434, "keep": 0.29035134613513947, "leak": 0.9822903797030449, "replacement": 0.5699137598276138}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 8.411662101745605, "loss_scale": 4.0, "retries": 0, "seconds": 539.6, "max_memory_gb": 27.92, "pair": 0.3092775046825409, "sign": 0.2817769106477499, "keep": 0.06482018774840981, "leak": 1.038551519450266, "replacement": 0.7422139570116997}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 5.3614702224731445, "loss_scale": 4.0, "retries": 0, "seconds": 541.0, "max_memory_gb": 27.92, "pair": 0.5984938070178032, "sign": 0.7784999832510948, "keep": 0.10699835629202425, "leak": 0.39370378060266376, "replacement": 1.1396689116954803}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 7.100438594818115, "loss_scale": 4.0, "retries": 0, "seconds": 542.5, "max_memory_gb": 27.92, "pair": 0.33546140417456627, "sign": 0.30630611814558506, "keep": 0.4379930084105581, "leak": 1.066241068765521, "replacement": 0.9741712510585785}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 6.033018112182617, "loss_scale": 4.0, "retries": 0, "seconds": 544.0, "max_memory_gb": 27.92, "pair": 0.5075983703136444, "sign": 0.5382073279470205, "keep": 0.4670477043837309, "leak": 0.314875356387347, "replacement": 1.0889448001980782}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 4.772689342498779, "loss_scale": 4.0, "retries": 0, "seconds": 545.4, "max_memory_gb": 27.92, "pair": 0.25506122410297394, "sign": 0.2139274850487709, "keep": 0.16015665949089453, "leak": 0.1750224553834414, "replacement": 0.48585890978574753}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 5.640717506408691, "loss_scale": 4.0, "retries": 0, "seconds": 546.8, "max_memory_gb": 27.92, "pair": 0.30940692126750946, "sign": 0.3594202632084489, "keep": 0.06776931905187666, "leak": 0.5362791493535042, "replacement": 0.6570715010166168}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 6.619237899780273, "loss_scale": 4.0, "retries": 0, "seconds": 548.3, "max_memory_gb": 27.92, "pair": 0.23347122594714165, "sign": 0.17870597634464502, "keep": 0.11368632968515158, "leak": 0.8673283345997334, "replacement": 0.5964994877576828}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 8.608811378479004, "loss_scale": 4.0, "retries": 0, "seconds": 549.7, "max_memory_gb": 27.92, "pair": 0.3184906765818596, "sign": 0.3573243161663413, "keep": 0.191452550701797, "leak": 0.43973399279639125, "replacement": 0.7028126046061516}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 11.545001983642578, "loss_scale": 4.0, "retries": 0, "seconds": 551.2, "max_memory_gb": 27.92, "pair": 0.36564282700419426, "sign": 0.4124876353889704, "keep": 0.10556023195385933, "leak": 2.709577202796936, "replacement": 1.302061066031456}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 7.865620136260986, "loss_scale": 4.0, "retries": 0, "seconds": 552.7, "max_memory_gb": 27.92, "pair": 0.7052670642733574, "sign": 0.876009114086628, "keep": 0.41824771650135517, "leak": 0.4194148608657997, "replacement": 1.4572491496801376}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 8.605599403381348, "loss_scale": 4.0, "retries": 0, "seconds": 554.1, "max_memory_gb": 27.92, "pair": 0.4737888276576996, "sign": 0.583702159114182, "keep": 1.0202565034851432, "leak": 0.856513187289238, "replacement": 1.489896446466446}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 5.463141441345215, "loss_scale": 4.0, "retries": 0, "seconds": 555.6, "max_memory_gb": 27.92, "pair": 0.29012218676507473, "sign": 0.30975368712097406, "keep": 0.16796273831278086, "leak": 0.9517454607412219, "replacement": 0.7669167593121529}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 6.434523105621338, "loss_scale": 4.0, "retries": 0, "seconds": 557.0, "max_memory_gb": 27.92, "pair": 0.11235904833301902, "sign": 0.0531343687325716, "keep": 0.04045513179153204, "leak": 0.5543361883610487, "replacement": 0.29773784428834915}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 3.709625482559204, "loss_scale": 4.0, "retries": 0, "seconds": 558.5, "max_memory_gb": 27.92, "pair": 0.40571893751621246, "sign": 0.4831668622791767, "keep": 0.3032959224656224, "leak": 0.5481324959546328, "replacement": 0.9359834790229797}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 13.090877532958984, "loss_scale": 4.0, "retries": 0, "seconds": 560.0, "max_memory_gb": 27.92, "pair": 0.2981752287596464, "sign": 0.1801478574052453, "keep": 0.06498108338564634, "leak": 0.5984715162194334, "replacement": 0.5703575983643532}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 10.962980270385742, "loss_scale": 4.0, "retries": 0, "seconds": 561.4, "max_memory_gb": 27.92, "pair": 0.26607187278568745, "sign": 0.5016974601894617, "keep": 0.10600290261209011, "leak": 1.9119556830846705, "replacement": 1.0479109734296799}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 12.21691608428955, "loss_scale": 4.0, "retries": 0, "seconds": 562.9, "max_memory_gb": 27.92, "pair": 0.25002546794712543, "sign": 0.2512753466144204, "keep": 0.12453700974583626, "leak": 1.242203000932932, "replacement": 0.7484823614358902}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 24.17056655883789, "loss_scale": 4.0, "retries": 0, "seconds": 564.4, "max_memory_gb": 27.92, "pair": 0.32661419082432985, "sign": 0.10758447833359241, "keep": 0.2093580011278391, "leak": 2.314150106161833, "replacement": 1.0636229664087296}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 10.954349517822266, "loss_scale": 4.0, "retries": 0, "seconds": 565.8, "max_memory_gb": 27.92, "pair": 0.32661872543394566, "sign": 0.4210811108350754, "keep": 0.2454417608678341, "leak": 1.3923454601317644, "replacement": 1.0079665631055832}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 3.843838930130005, "loss_scale": 4.0, "retries": 0, "seconds": 567.2, "max_memory_gb": 27.92, "pair": 0.5687436275184155, "sign": 0.7187159471213818, "keep": 0.4394000321626663, "leak": 0.454307359730592, "replacement": 1.2613784968852997}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 5.636519432067871, "loss_scale": 4.0, "retries": 0, "seconds": 568.6, "max_memory_gb": 27.92, "pair": 0.14009645767509937, "sign": 0.08540724962949753, "keep": 0.034267470706254244, "leak": 0.726460540201515, "replacement": 0.3815489560365677}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 9.084641456604004, "loss_scale": 4.0, "retries": 0, "seconds": 570.0, "max_memory_gb": 27.92, "pair": 0.11112346034497023, "sign": 0.16187874414026737, "keep": 0.08770289970561862, "leak": 0.9835589127615094, "replacement": 0.4818040058016777}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 7.550138473510742, "loss_scale": 4.0, "retries": 0, "seconds": 571.5, "max_memory_gb": 27.92, "pair": 0.3485564272850752, "sign": 0.4252130966633558, "keep": 0.14894399186596274, "leak": 0.6102768626878969, "replacement": 0.7882041484117508}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 8.17155933380127, "loss_scale": 4.0, "retries": 0, "seconds": 572.8, "max_memory_gb": 27.92, "pair": 0.16346884984523058, "sign": 0.13971328362822533, "keep": 0.04806876793736592, "leak": 1.7345739752054214, "replacement": 0.6910033747553825}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 7.133337020874023, "loss_scale": 4.0, "retries": 0, "seconds": 574.2, "max_memory_gb": 27.92, "pair": 0.23016124218702316, "sign": 0.08011950924992561, "keep": 0.08597702719271183, "leak": 0.38933090200589504, "replacement": 0.41054222732782364}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 5.471073150634766, "loss_scale": 4.0, "retries": 0, "seconds": 575.7, "max_memory_gb": 27.92, "pair": 0.35900401324033737, "sign": 0.4947139397263527, "keep": 0.3157110894098878, "leak": 0.7068214236060157, "replacement": 0.9409219175577164}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 7.683529853820801, "loss_scale": 4.0, "retries": 0, "seconds": 577.1, "max_memory_gb": 27.92, "pair": 0.33372489735484123, "sign": 0.3497543828561902, "keep": 0.19902956020087004, "leak": 0.8296555066481233, "replacement": 0.8155307583510876}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 11.643818855285645, "loss_scale": 4.0, "retries": 0, "seconds": 578.6, "max_memory_gb": 27.92, "pair": 0.4061986654996872, "sign": 0.6046869745478034, "keep": 0.6375422053970397, "leak": 0.4772246587090194, "replacement": 1.1466194204986095}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 9.95357608795166, "loss_scale": 4.0, "retries": 0, "seconds": 580.1, "max_memory_gb": 27.92, "pair": 0.31111743673682213, "sign": 0.3392690680921078, "keep": 0.19332272466272116, "leak": 2.047300646547228, "replacement": 1.0892384871840477}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 34.7101936340332, "loss_scale": 4.0, "retries": 0, "seconds": 581.7, "max_memory_gb": 27.92, "pair": 0.42516287229955196, "sign": 0.6173991356045008, "keep": 0.5976129075570498, "leak": 1.0577835980802774, "replacement": 1.2971148490905762}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 8.200278282165527, "loss_scale": 4.0, "retries": 0, "seconds": 583.1, "max_memory_gb": 27.92, "pair": 0.678051058202982, "sign": 0.7920862277969718, "keep": 0.23971634171903133, "leak": 0.9690385386347771, "replacement": 1.4362120181322098}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 9.455886840820312, "loss_scale": 4.0, "retries": 0, "seconds": 584.6, "max_memory_gb": 27.92, "pair": 0.159315824508667, "sign": 0.12266522459685802, "keep": 0.019647457695100456, "leak": 1.9475511983036995, "replacement": 0.7173599824309349}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 11.084915161132812, "loss_scale": 4.0, "retries": 0, "seconds": 586.1, "max_memory_gb": 27.92, "pair": 0.4238617680966854, "sign": 0.5074650859460235, "keep": 0.3927982021123171, "leak": 0.9580900093424134, "replacement": 1.1135159134864807}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 7.279848575592041, "loss_scale": 4.0, "retries": 0, "seconds": 587.5, "max_memory_gb": 27.92, "pair": 0.17048239894211292, "sign": 0.1701178690418601, "keep": 0.062479630432790145, "leak": 0.6891074180603027, "replacement": 0.4590580090880394}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 9.355816841125488, "loss_scale": 4.0, "retries": 0, "seconds": 589.0, "max_memory_gb": 27.92, "pair": 0.508415712043643, "sign": 0.6622000960633159, "keep": 0.2869527111761272, "leak": 0.9032315043732524, "replacement": 1.2088000029325485}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 14.778576850891113, "loss_scale": 4.0, "retries": 0, "seconds": 590.5, "max_memory_gb": 27.92, "pair": 0.7037068400532007, "sign": 0.9414828191511333, "keep": 2.191221607849002, "leak": 1.2720844438299537, "replacement": 2.5880801677703857}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 8.555084228515625, "loss_scale": 4.0, "retries": 0, "seconds": 591.9, "max_memory_gb": 27.92, "pair": 0.36023999378085136, "sign": 0.5156825606245548, "keep": 0.46237763355020434, "leak": 0.8040681481361389, "replacement": 1.0502870976924896}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 9.917783737182617, "loss_scale": 4.0, "retries": 0, "seconds": 593.4, "max_memory_gb": 27.92, "pair": 0.38537439703941345, "sign": 0.4210602715611458, "keep": 0.34219529916299507, "leak": 1.445985376369208, "replacement": 1.128498487174511}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 5.695908546447754, "loss_scale": 4.0, "retries": 0, "seconds": 594.9, "max_memory_gb": 27.92, "pair": 0.19517337903380394, "sign": 0.13547835545614362, "keep": 0.17750298570899758, "leak": 1.494296059012413, "replacement": 0.725238062441349}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 12.10234260559082, "loss_scale": 4.0, "retries": 0, "seconds": 596.4, "max_memory_gb": 27.92, "pair": 0.44009047001600266, "sign": 0.5688090864568949, "keep": 0.9614318019594066, "leak": 1.176775461761281, "replacement": 1.499404788017273}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 7.302859783172607, "loss_scale": 4.0, "retries": 0, "seconds": 597.9, "max_memory_gb": 27.92, "pair": 0.4789813533425331, "sign": 0.6010986659675837, "keep": 0.2700460769701749, "leak": 0.17995150992646813, "replacement": 0.9595415815711021}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 5.222929954528809, "loss_scale": 4.0, "retries": 0, "seconds": 599.3, "max_memory_gb": 27.92, "pair": 0.22104433551430702, "sign": 0.1656046975404024, "keep": 0.039833756163716316, "leak": 1.1228330731391907, "replacement": 0.6044718399643898}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 11.52916431427002, "loss_scale": 4.0, "retries": 0, "seconds": 600.8, "max_memory_gb": 27.92, "pair": 0.4293955974280834, "sign": 0.40897105913609266, "keep": 0.8849690305069089, "leak": 0.2772034667432308, "replacement": 1.1456664875149727}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 10.792237281799316, "loss_scale": 4.0, "retries": 0, "seconds": 602.2, "max_memory_gb": 27.92, "pair": 0.35398488119244576, "sign": 0.388911709189415, "keep": 1.0713950442150235, "leak": 0.7689751153811812, "replacement": 1.2763820141553879}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 9.110262870788574, "loss_scale": 4.0, "retries": 0, "seconds": 603.5, "max_memory_gb": 27.92, "pair": 0.2950711604207754, "sign": 0.16087279934436083, "keep": 0.08952991699334234, "leak": 0.3981974897906184, "replacement": 0.5198219045996666}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 16.52223014831543, "loss_scale": 4.0, "retries": 0, "seconds": 605.0, "max_memory_gb": 27.92, "pair": 0.8367163240909576, "sign": 1.136525958776474, "keep": 1.2015184592455626, "leak": 0.3328918293118477, "replacement": 2.0889615416526794}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 9.948040962219238, "loss_scale": 4.0, "retries": 0, "seconds": 606.4, "max_memory_gb": 27.92, "pair": 0.26509943790733814, "sign": 0.23860680870711803, "keep": 0.06392508395947516, "leak": 0.4596445746719837, "replacement": 0.5312765538692474}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 11.289167404174805, "loss_scale": 4.0, "retries": 0, "seconds": 607.7, "max_memory_gb": 27.92, "pair": 0.4278407245874405, "sign": 0.5826753005385399, "keep": 0.08513349760323763, "leak": 1.125023638829589, "replacement": 1.043001040816307}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 17.49469757080078, "loss_scale": 4.0, "retries": 0, "seconds": 609.2, "max_memory_gb": 27.92, "pair": 0.3011069316416979, "sign": 0.40787050779908895, "keep": 0.5952682746574283, "leak": 2.269961006939411, "replacement": 1.3701665848493576}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 14.90798568725586, "loss_scale": 4.0, "retries": 0, "seconds": 610.7, "max_memory_gb": 27.92, "pair": 0.3407090213149786, "sign": 0.3088686503469944, "keep": 0.08930448460159823, "leak": 1.5450073815882206, "replacement": 0.9260474443435669}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 12.96568775177002, "loss_scale": 4.0, "retries": 0, "seconds": 612.1, "max_memory_gb": 27.92, "pair": 0.6300879158079624, "sign": 0.5917348507791758, "keep": 0.5594706269912422, "leak": 0.16888007597299293, "replacement": 1.2479107044637203}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 12.121635437011719, "loss_scale": 4.0, "retries": 0, "seconds": 613.6, "max_memory_gb": 27.92, "pair": 0.6679178010672331, "sign": 0.8092786259949207, "keep": 0.8682095222175121, "leak": 0.8832896728999913, "replacement": 1.7274842709302902}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 8.609647750854492, "loss_scale": 4.0, "retries": 0, "seconds": 615.1, "max_memory_gb": 27.92, "pair": 0.23137588798999786, "sign": 0.10253085102885962, "keep": 0.09209136664867401, "leak": 0.888755064457655, "replacement": 0.5508757680654526}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 14.44556713104248, "loss_scale": 4.0, "retries": 0, "seconds": 616.5, "max_memory_gb": 27.92, "pair": 0.3174068033695221, "sign": 0.2126213563606143, "keep": 0.058322254568338394, "leak": 0.8283763825893402, "replacement": 0.6599727272987366}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 16.897533416748047, "loss_scale": 4.0, "retries": 0, "seconds": 618.1, "max_memory_gb": 27.92, "pair": 0.7408503144979477, "sign": 0.6949002277106047, "keep": 0.6823169179260731, "leak": 0.07398803939577192, "replacement": 1.4479558393359184}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 9.85591983795166, "loss_scale": 4.0, "retries": 0, "seconds": 619.5, "max_memory_gb": 27.92, "pair": 0.3148846998810768, "sign": 0.28112874273210764, "keep": 0.13670655293390155, "leak": 2.4035495221614838, "replacement": 1.124689757823944}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 6.449481964111328, "loss_scale": 4.0, "retries": 0, "seconds": 621.0, "max_memory_gb": 27.92, "pair": 0.22610293701291084, "sign": 0.12306415382772684, "keep": 0.01079979381756857, "leak": 0.4106830870732665, "replacement": 0.3957056775689125}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 5.444093227386475, "loss_scale": 4.0, "retries": 0, "seconds": 622.5, "max_memory_gb": 27.92, "pair": 0.23984486982226372, "sign": 0.08757573296315968, "keep": 0.02710355940507725, "leak": 0.37150130048394203, "replacement": 0.39005984738469124}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 6.147617816925049, "loss_scale": 4.0, "retries": 0, "seconds": 623.9, "max_memory_gb": 27.92, "pair": 0.3494912460446358, "sign": 0.304658479988575, "keep": 0.2617910101544112, "leak": 0.5437254691496491, "replacement": 0.7686473429203033}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 16.798648834228516, "loss_scale": 4.0, "retries": 0, "seconds": 625.2, "max_memory_gb": 27.92, "pair": 0.150031179189682, "sign": 0.07307009492069483, "keep": 0.08651071257190779, "leak": 1.6508980756625533, "replacement": 0.6425461024045944}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 8.027122497558594, "loss_scale": 4.0, "retries": 0, "seconds": 626.6, "max_memory_gb": 27.92, "pair": 0.2273635733872652, "sign": 0.2460484839975834, "keep": 0.1017562028428074, "leak": 0.44400956528261304, "replacement": 0.5122682973742485}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 6.196936130523682, "loss_scale": 4.0, "retries": 0, "seconds": 628.0, "max_memory_gb": 27.92, "pair": 0.21109294146299362, "sign": 0.26683060824871063, "keep": 0.09368549333885312, "leak": 1.249392256140709, "replacement": 0.7036990523338318}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 7.934420108795166, "loss_scale": 4.0, "retries": 0, "seconds": 629.5, "max_memory_gb": 27.92, "pair": 0.21050062030553818, "sign": 0.14888147078454494, "keep": 0.012697001569904387, "leak": 0.11923199752345681, "replacement": 0.32109785825014114}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 16.83290672302246, "loss_scale": 4.0, "retries": 0, "seconds": 630.9, "max_memory_gb": 27.92, "pair": 0.435694795101881, "sign": 0.43750160187482834, "keep": 0.4763876357465051, "leak": 0.6615211116150022, "replacement": 1.0580196715891361}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 9.86208438873291, "loss_scale": 4.0, "retries": 0, "seconds": 632.4, "max_memory_gb": 27.92, "pair": 0.27369350008666515, "sign": 0.30474089458584785, "keep": 0.15535419061779976, "leak": 1.2884850353002548, "replacement": 0.8258622884750366}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 29.710304260253906, "loss_scale": 4.0, "retries": 0, "seconds": 633.8, "max_memory_gb": 27.92, "pair": 0.3954453654587269, "sign": 0.5538792116567492, "keep": 0.6307775003369898, "leak": 1.4128689281642437, "replacement": 1.3409909158945084}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 6.335437297821045, "loss_scale": 4.0, "retries": 0, "seconds": 635.3, "max_memory_gb": 27.92, "pair": 0.13714097626507282, "sign": 0.09222328709438443, "keep": 0.11210165033116937, "leak": 0.8269635029137135, "replacement": 0.44604431465268135}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 7.260511875152588, "loss_scale": 4.0, "retries": 0, "seconds": 636.7, "max_memory_gb": 27.92, "pair": 0.4308262802660465, "sign": 0.44024524465203285, "keep": 0.3654042137786746, "leak": 0.2339091620524414, "replacement": 0.8921282738447189}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 6.780208587646484, "loss_scale": 4.0, "retries": 0, "seconds": 638.2, "max_memory_gb": 27.92, "pair": 0.24148524925112724, "sign": 0.3244093433022499, "keep": 0.017757285968400538, "leak": 0.6926320157945156, "replacement": 0.5857265591621399}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 14.440526008605957, "loss_scale": 4.0, "retries": 0, "seconds": 639.6, "max_memory_gb": 27.92, "pair": 0.6259948387742043, "sign": 0.7048286758363247, "keep": 0.9250164269469678, "leak": 0.7373323077335954, "replacement": 1.6252503246068954}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 9.174952507019043, "loss_scale": 4.0, "retries": 0, "seconds": 641.0, "max_memory_gb": 27.92, "pair": 0.406276348978281, "sign": 0.4111297056078911, "keep": 0.13927779346704483, "leak": 1.9435726990923285, "replacement": 1.167373239994049}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 7.781917095184326, "loss_scale": 4.0, "retries": 0, "seconds": 642.5, "max_memory_gb": 27.92, "pair": 0.25300055369734764, "sign": 0.3174873776733875, "keep": 0.06098069658037275, "leak": 0.7980940863490105, "replacement": 0.6417580917477608}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 8.667790412902832, "loss_scale": 4.0, "retries": 0, "seconds": 643.9, "max_memory_gb": 27.92, "pair": 0.3002834524959326, "sign": 0.4246323904953897, "keep": 0.20716495858505368, "leak": 0.287254024296999, "replacement": 0.687995620071888}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 8.22046947479248, "loss_scale": 4.0, "retries": 0, "seconds": 645.4, "max_memory_gb": 27.92, "pair": 0.6395321264863014, "sign": 0.802932501770556, "keep": 0.9081360345007852, "leak": 1.1025203245571902, "replacement": 1.770696483552456}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 6.380627155303955, "loss_scale": 4.0, "retries": 0, "seconds": 646.8, "max_memory_gb": 27.92, "pair": 0.2414630763232708, "sign": 0.23579067550599575, "keep": 0.2433928992250003, "leak": 0.5837497070897371, "replacement": 0.626992292702198}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 8.668191909790039, "loss_scale": 4.0, "retries": 0, "seconds": 648.3, "max_memory_gb": 27.92, "pair": 0.43717724084854126, "sign": 0.555744506418705, "keep": 0.4351006129581947, "leak": 1.483448324717756, "replacement": 1.3034618720412254}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 7.876980304718018, "loss_scale": 4.0, "retries": 0, "seconds": 649.7, "max_memory_gb": 27.92, "pair": 0.1970847174525261, "sign": 0.2558154333382845, "keep": 0.2799692515982315, "leak": 0.7775119766592979, "replacement": 0.6593550592660904}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 8.444778442382812, "loss_scale": 4.0, "retries": 0, "seconds": 651.2, "max_memory_gb": 27.92, "pair": 0.1761674154549837, "sign": 0.15251697599887848, "keep": 0.05266479909187183, "leak": 1.4538534651510417, "replacement": 0.642221674323082}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 7.507588863372803, "loss_scale": 4.0, "retries": 0, "seconds": 652.6, "max_memory_gb": 27.92, "pair": 0.28471988439559937, "sign": 0.15112508786842227, "keep": 0.264160513645038, "leak": 0.8563452865928411, "replacement": 0.7064490094780922}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 11.814642906188965, "loss_scale": 4.0, "retries": 0, "seconds": 654.0, "max_memory_gb": 27.92, "pair": 0.24477020278573036, "sign": 0.1440356131643057, "keep": 0.1257795626297593, "leak": 0.6655295975506306, "replacement": 0.5460601858794689}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 11.222421646118164, "loss_scale": 4.0, "retries": 0, "seconds": 655.4, "max_memory_gb": 27.92, "pair": 0.3919374532997608, "sign": 0.36069904267787933, "keep": 0.43859110958874226, "leak": 0.9027217694092542, "replacement": 1.017262950539589}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 13.212492942810059, "loss_scale": 4.0, "retries": 0, "seconds": 656.9, "max_memory_gb": 27.92, "pair": 0.40444472432136536, "sign": 0.3068559458479285, "keep": 0.09322342788800597, "leak": 1.0352086529601365, "replacement": 0.8632865473628044}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 10.81180477142334, "loss_scale": 4.0, "retries": 0, "seconds": 658.4, "max_memory_gb": 27.92, "pair": 0.369805172085762, "sign": 0.2564567122608423, "keep": 0.2652072347700596, "leak": 0.8618733808398247, "replacement": 0.8461054936051369}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 7.2698750495910645, "loss_scale": 4.0, "retries": 0, "seconds": 659.8, "max_memory_gb": 27.92, "pair": 0.21438205987215042, "sign": 0.25522248819470406, "keep": 0.26993035385385156, "leak": 1.5851861238479614, "replacement": 0.8732550144195557}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 10.375927925109863, "loss_scale": 4.0, "retries": 0, "seconds": 661.3, "max_memory_gb": 27.92, "pair": 0.3837893642485142, "sign": 0.6375708552077413, "keep": 0.07916041929274797, "leak": 0.6562272263690829, "replacement": 0.9062118157744408}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 6.8702712059021, "loss_scale": 4.0, "retries": 0, "seconds": 662.7, "max_memory_gb": 27.92, "pair": 0.2032202873378992, "sign": 0.27633069083094597, "keep": 0.04318361391779035, "leak": 0.3558198595419526, "replacement": 0.4519324116408825}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 23.6076602935791, "loss_scale": 4.0, "retries": 0, "seconds": 664.2, "max_memory_gb": 27.92, "pair": 0.4232805594801903, "sign": 0.5342975594103336, "keep": 0.5193024948239326, "leak": 1.868142526189331, "replacement": 1.4171162247657776}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 5.651373863220215, "loss_scale": 4.0, "retries": 0, "seconds": 665.6, "max_memory_gb": 27.92, "pair": 0.30211570858955383, "sign": 0.3121243566274643, "keep": 0.05714265536516905, "leak": 0.09402742283418775, "replacement": 0.5102560520172119}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 7.6846184730529785, "loss_scale": 4.0, "retries": 0, "seconds": 667.1, "max_memory_gb": 27.92, "pair": 0.249866864643991, "sign": 0.2811107384040952, "keep": 0.04599922482157126, "leak": 0.5460972945438698, "replacement": 0.5499461889266968}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 15.263572692871094, "loss_scale": 4.0, "retries": 0, "seconds": 668.6, "max_memory_gb": 27.92, "pair": 0.3355907676741481, "sign": 0.6385445334017277, "keep": 0.7364078261889517, "leak": 1.8278149142861366, "replacement": 1.4800207018852234}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 9.40631103515625, "loss_scale": 4.0, "retries": 0, "seconds": 670.1, "max_memory_gb": 27.92, "pair": 0.2238803692162037, "sign": 0.2403246285393834, "keep": 0.03420775046106428, "leak": 0.4891736740246415, "replacement": 0.48343996703624725}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 12.8873872756958, "loss_scale": 4.0, "retries": 0, "seconds": 671.5, "max_memory_gb": 27.92, "pair": 0.23161482252180576, "sign": 0.19499332457780838, "keep": 0.022474885918200016, "leak": 0.2529785712249577, "replacement": 0.4035935699939728}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 15.617311477661133, "loss_scale": 4.0, "retries": 0, "seconds": 673.0, "max_memory_gb": 27.92, "pair": 0.13935552164912224, "sign": 0.12938499823212624, "keep": 0.00779805111233145, "leak": 2.633545361459255, "replacement": 0.8663334101438522}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 18.16985511779785, "loss_scale": 4.0, "retries": 0, "seconds": 674.5, "max_memory_gb": 27.92, "pair": 0.09992059879004955, "sign": 0.15340944891795516, "keep": 0.14885990100447088, "leak": 2.2381260618567467, "replacement": 0.8105867877602577}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 26.35985565185547, "loss_scale": 4.0, "retries": 0, "seconds": 675.9, "max_memory_gb": 27.92, "pair": 0.1559832552447915, "sign": 0.25794547726400197, "keep": 0.06871356815099716, "leak": 2.652053112629801, "replacement": 0.9823260605335236}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 10.33820629119873, "loss_scale": 4.0, "retries": 0, "seconds": 677.4, "max_memory_gb": 27.92, "pair": 0.10036532953381538, "sign": 0.10581525322049856, "keep": 0.021167809376493096, "leak": 1.134923130273819, "replacement": 0.44758763164281845}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 4.2824201583862305, "loss_scale": 4.0, "retries": 0, "seconds": 678.8, "max_memory_gb": 27.92, "pair": 0.14157451130449772, "sign": 0.1446097856387496, "keep": 0.021026540780439973, "leak": 0.1544637493789196, "replacement": 0.26300860941410065}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 11.949345588684082, "loss_scale": 4.0, "retries": 0, "seconds": 680.3, "max_memory_gb": 27.92, "pair": 0.38895805552601814, "sign": 0.15766911394894123, "keep": 0.4571002737211529, "leak": 1.277097712183604, "replacement": 1.0156171768903732}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 18.50185203552246, "loss_scale": 4.0, "retries": 0, "seconds": 681.7, "max_memory_gb": 27.92, "pair": 0.13549006544053555, "sign": 0.09048958821222186, "keep": 0.026360323536209762, "leak": 0.23082472453825176, "replacement": 0.25162120908498764}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 18.317161560058594, "loss_scale": 4.0, "retries": 0, "seconds": 683.2, "max_memory_gb": 27.92, "pair": 0.42558827344328165, "sign": 0.5248462986201048, "keep": 0.5091211614198983, "leak": 1.4109730476047844, "replacement": 1.2953153029084206}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 19.134658813476562, "loss_scale": 4.0, "retries": 0, "seconds": 684.6, "max_memory_gb": 27.92, "pair": 0.1982386102899909, "sign": 0.26647560950368643, "keep": 0.05863875232171267, "leak": 1.5389028489589691, "replacement": 0.7455215007066727}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 14.932242393493652, "loss_scale": 4.0, "retries": 0, "seconds": 686.1, "max_memory_gb": 27.92, "pair": 0.18427039496600628, "sign": 0.1268965657800436, "keep": 0.2264028414792847, "leak": 0.2956217595938142, "replacement": 0.4348255395889282}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 14.876733779907227, "loss_scale": 4.0, "retries": 0, "seconds": 687.5, "max_memory_gb": 27.92, "pair": 0.3017891589552164, "sign": 0.4148579053580761, "keep": 0.31098052560992073, "leak": 0.5258341506123543, "replacement": 0.7961669713258743}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 12.322726249694824, "loss_scale": 4.0, "retries": 0, "seconds": 689.0, "max_memory_gb": 27.92, "pair": 0.1851580236107111, "sign": 0.3287607282400131, "keep": 0.2659931983653223, "leak": 0.5794075950980186, "replacement": 0.6273868530988693}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 10.412138938903809, "loss_scale": 4.0, "retries": 0, "seconds": 690.4, "max_memory_gb": 27.92, "pair": 0.3682195758447051, "sign": 0.2027408666908741, "keep": 0.10111358857830055, "leak": 1.7199876457452774, "replacement": 0.9501436948776245}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 15.311429977416992, "loss_scale": 4.0, "retries": 0, "seconds": 691.8, "max_memory_gb": 27.92, "pair": 0.25265980139374733, "sign": 0.4630748708732426, "keep": 0.020524672479950823, "leak": 0.8666430786252022, "replacement": 0.7111203446984291}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 12.775747299194336, "loss_scale": 4.0, "retries": 0, "seconds": 693.3, "max_memory_gb": 27.92, "pair": 0.2705269977450371, "sign": 0.22715575224719942, "keep": 0.0720767987659201, "leak": 0.7939837424273719, "replacement": 0.6186392232775688}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 25.25005531311035, "loss_scale": 4.0, "retries": 0, "seconds": 694.7, "max_memory_gb": 27.92, "pair": 0.5008400324732065, "sign": 0.5536596570163965, "keep": 0.8761459293309599, "leak": 0.6204161904752254, "replacement": 1.3708468228578568}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 11.060965538024902, "loss_scale": 4.0, "retries": 0, "seconds": 696.2, "max_memory_gb": 27.92, "pair": 0.5607902780175209, "sign": 0.573856420814991, "keep": 0.4281070102006197, "leak": 0.45733894931618124, "replacement": 1.176106721162796}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 26.996915817260742, "loss_scale": 4.0, "retries": 0, "seconds": 697.7, "max_memory_gb": 27.92, "pair": 0.1925786454230547, "sign": 0.1468260190449655, "keep": 0.0690531037398614, "leak": 0.9681205302476883, "replacement": 0.542548343539238}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 16.48469352722168, "loss_scale": 4.0, "retries": 0, "seconds": 699.1, "max_memory_gb": 27.92, "pair": 0.37633573077619076, "sign": 0.5088146515190601, "keep": 0.369580359372776, "leak": 0.5710673909634352, "replacement": 0.9583001434803009}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 31.814279556274414, "loss_scale": 4.0, "retries": 0, "seconds": 700.6, "max_memory_gb": 27.92, "pair": 0.6052185222506523, "sign": 0.8896436989307404, "keep": 0.7594106141477823, "leak": 0.30008091661147773, "replacement": 1.5047659277915955}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 21.418594360351562, "loss_scale": 4.0, "retries": 0, "seconds": 702.0, "max_memory_gb": 27.92, "pair": 0.6579601801931858, "sign": 0.9297993145883083, "keep": 1.2316426550969481, "leak": 0.3526683261152357, "replacement": 1.8268482312560081}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 11.536251068115234, "loss_scale": 4.0, "retries": 0, "seconds": 703.5, "max_memory_gb": 27.92, "pair": 0.24821991380304098, "sign": 0.46368248015642166, "keep": 0.02826778031885624, "leak": 1.083270475268364, "replacement": 0.7650126665830612}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 14.42547607421875, "loss_scale": 4.0, "retries": 0, "seconds": 704.9, "max_memory_gb": 27.92, "pair": 0.4752764627337456, "sign": 0.6356127597391605, "keep": 0.7050316252280027, "leak": 0.6607199908667098, "replacement": 1.3107786551117897}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 5.649924278259277, "loss_scale": 4.0, "retries": 0, "seconds": 706.3, "max_memory_gb": 27.92, "pair": 0.19455905444920063, "sign": 0.1567375618033111, "keep": 0.051963033620268106, "leak": 0.673473946750164, "replacement": 0.4672778397798538}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 9.83288860321045, "loss_scale": 4.0, "retries": 0, "seconds": 707.8, "max_memory_gb": 27.92, "pair": 0.3443928249180317, "sign": 0.353885505348444, "keep": 0.2780202468857169, "leak": 2.5162658374756575, "replacement": 1.2894121706485748}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 11.959896087646484, "loss_scale": 4.0, "retries": 0, "seconds": 709.2, "max_memory_gb": 27.92, "pair": 0.1730162613093853, "sign": 0.34644652530550957, "keep": 0.73357262522768, "leak": 1.9925583880394697, "replacement": 1.2111654318869114}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 11.304604530334473, "loss_scale": 4.0, "retries": 0, "seconds": 710.6, "max_memory_gb": 27.92, "pair": 0.22566949017345905, "sign": 0.28355997521430254, "keep": 0.035096648265607655, "leak": 0.849303524941206, "replacement": 0.5973236784338951}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 14.729534149169922, "loss_scale": 4.0, "retries": 0, "seconds": 712.1, "max_memory_gb": 27.92, "pair": 0.32912931498140097, "sign": 0.42044613882899284, "keep": 0.21906879886228126, "leak": 1.2263649813830853, "replacement": 0.9554779902100563}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 7.530716896057129, "loss_scale": 4.0, "retries": 0, "seconds": 713.5, "max_memory_gb": 27.92, "pair": 0.23462851345539093, "sign": 0.1861338224261999, "keep": 0.02596763346809894, "leak": 0.7739923298358917, "replacement": 0.5341773182153702}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 13.405702590942383, "loss_scale": 4.0, "retries": 0, "seconds": 715.0, "max_memory_gb": 27.92, "pair": 0.5792633816599846, "sign": 0.8433196544647217, "keep": 0.33377606797148474, "leak": 0.5742932893335819, "replacement": 1.3113846331834793}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 15.363490104675293, "loss_scale": 4.0, "retries": 0, "seconds": 716.5, "max_memory_gb": 27.92, "pair": 0.17833831906318665, "sign": 0.24512622226029634, "keep": 0.06924238079227507, "leak": 1.1693495959043503, "replacement": 0.6278600245714188}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 8.891860008239746, "loss_scale": 4.0, "retries": 0, "seconds": 717.9, "max_memory_gb": 27.92, "pair": 0.17982018180191517, "sign": 0.25568351289257407, "keep": 0.02305247145704925, "leak": 1.4090701155364513, "replacement": 0.6714556962251663}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 6.3678975105285645, "loss_scale": 4.0, "retries": 0, "seconds": 719.4, "max_memory_gb": 27.92, "pair": 0.19307811558246613, "sign": 0.06578305084258318, "keep": 0.07794011238729581, "leak": 0.7733383132144809, "replacement": 0.458274282515049}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 32.403255462646484, "loss_scale": 4.0, "retries": 0, "seconds": 720.7, "max_memory_gb": 27.92, "pair": 0.3722143042832613, "sign": 0.6007706103846431, "keep": 0.9242499237880111, "leak": 1.201514108106494, "replacement": 1.4351030886173248}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 8.24357795715332, "loss_scale": 4.0, "retries": 0, "seconds": 722.1, "max_memory_gb": 27.92, "pair": 0.43588101491332054, "sign": 0.38913610484451056, "keep": 0.2552012938540429, "leak": 1.5503490989794955, "replacement": 1.1456370130181313}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 10.481541633605957, "loss_scale": 4.0, "retries": 0, "seconds": 723.6, "max_memory_gb": 27.92, "pair": 0.43627277202904224, "sign": 0.4073451035656035, "keep": 0.37005850905552506, "leak": 0.846523939166218, "replacement": 1.0366055741906166}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 17.094045639038086, "loss_scale": 4.0, "retries": 0, "seconds": 725.0, "max_memory_gb": 27.92, "pair": 0.45516038686037064, "sign": 0.5675951899029315, "keep": 0.7046660264313687, "leak": 0.6083999536931515, "replacement": 1.243390940129757}
diff --git a/runs/single_llava/replacement/L06_s13/adapter/README.md b/runs/single_llava/replacement/L06_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L06_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L06_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L06_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..8d247afd5bb967b0587a4de556ebef4faa5a0043
--- /dev/null
+++ b/runs/single_llava/replacement/L06_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.6.mlp.gate_proj",
+    "model.language_model.layers.6.self_attn.k_proj",
+    "model.language_model.layers.6.self_attn.v_proj",
+    "model.language_model.layers.6.self_attn.q_proj",
+    "model.language_model.layers.6.self_attn.o_proj",
+    "model.language_model.layers.6.mlp.up_proj",
+    "model.language_model.layers.6.mlp.down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L06_s13/config.json b/runs/single_llava/replacement/L06_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..1ccfb4ae4c459fa93975d26febf010b0e1cf9881
--- /dev/null
+++ b/runs/single_llava/replacement/L06_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 6,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L06_s13/steps.jsonl b/runs/single_llava/replacement/L06_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..40e590e895841f2a81763a662d2179da9d8bf7f9
--- /dev/null
+++ b/runs/single_llava/replacement/L06_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.45989036560058594, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 26.04, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.6765344142913818, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 26.05, "pair": 1.09890416264534, "sign": 0.536851430311799, "keep": 5.374710360309109, "leak": 1.5101139183570922e-05, "replacement": 4.054688900709152}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.0464361906051636, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 26.15, "pair": 1.0982759892940521, "sign": 0.5420565977692604, "keep": 10.80023717880249, "leak": 6.496091606322807e-05, "replacement": 6.769439220428467}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.8760121464729309, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 26.17, "pair": 1.0920778810977936, "sign": 0.4712378429248929, "keep": 7.671447755768895, "leak": 9.126372566470309e-05, "replacement": 5.163443475961685}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.4747414290904999, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 26.17, "pair": 1.0982564687728882, "sign": 0.7763587534427643, "keep": 0.15896913781762123, "leak": 1.6662930431010636e-06, "replacement": 1.5659208297729492}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.49109822511672974, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 26.17, "pair": 1.093490719795227, "sign": 0.7359645739197731, "keep": 2.9301770627498627, "leak": 0.0001415431811437884, "replacement": 2.9265968799591064}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.9979747533798218, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 26.17, "pair": 1.0754530131816864, "sign": 0.5464815692976117, "keep": 6.870533170178533, "leak": 0.0005607988132396713, "replacement": 4.7841004729270935}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.9100481271743774, "loss_scale": 4.0, "retries": 0, "seconds": 10.6, "max_memory_gb": 26.17, "pair": 1.0787595808506012, "sign": 0.5843470767140388, "keep": 7.380683153867722, "leak": 0.00031677577817390556, "replacement": 5.061353802680969}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.8733254075050354, "loss_scale": 4.0, "retries": 0, "seconds": 11.9, "max_memory_gb": 26.17, "pair": 1.073350429534912, "sign": 0.6176950903609395, "keep": 5.70336938649416, "leak": 0.0011713278728677778, "replacement": 4.234175443649292}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.6392775774002075, "loss_scale": 4.0, "retries": 0, "seconds": 13.2, "max_memory_gb": 26.17, "pair": 1.081462413072586, "sign": 0.7797357551753521, "keep": 3.2712866812944412, "leak": 0.00040910443022212917, "replacement": 3.107075870037079}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 1.0799610614776611, "loss_scale": 4.0, "retries": 0, "seconds": 14.5, "max_memory_gb": 26.17, "pair": 1.0571471452713013, "sign": 0.45202929340302944, "keep": 5.256211645901203, "leak": 0.0023050711429277726, "replacement": 3.911843866109848}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.8966634273529053, "loss_scale": 4.0, "retries": 0, "seconds": 15.8, "max_memory_gb": 26.17, "pair": 1.0655721426010132, "sign": 0.6087011136114597, "keep": 3.046318910084665, "leak": 0.0030152685803876977, "replacement": 2.8938359320163727}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.8090197443962097, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 26.17, "pair": 1.0707924962043762, "sign": 0.5068861804902554, "keep": 5.344017997384071, "leak": 0.0018962762930954113, "replacement": 3.996718555688858}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.3630805015563965, "loss_scale": 4.0, "retries": 0, "seconds": 18.3, "max_memory_gb": 26.17, "pair": 1.0327130854129791, "sign": 0.3939954526722431, "keep": 7.771806851029396, "leak": 0.0038140946762723615, "replacement": 5.116567760705948}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.8655910491943359, "loss_scale": 4.0, "retries": 0, "seconds": 19.6, "max_memory_gb": 26.17, "pair": 1.058187022805214, "sign": 0.6877204291522503, "keep": 5.162570386193693, "leak": 0.002815462203216157, "replacement": 3.984036237001419}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.8596871495246887, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 26.17, "pair": 1.0721271634101868, "sign": 0.7802236825227737, "keep": 2.7516398613806814, "leak": 0.0011217712108191336, "replacement": 2.838339388370514}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.7055597901344299, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 26.17, "pair": 1.0612436532974243, "sign": 0.7056663930416107, "keep": 3.1759117329493165, "leak": 0.00479465849718963, "replacement": 3.0032314360141754}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.227708339691162, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 26.24, "pair": 1.046991616487503, "sign": 0.5300454832613468, "keep": 7.1534361988306046, "leak": 0.003327992439153604, "replacement": 4.889564484357834}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.817690134048462, "loss_scale": 4.0, "retries": 0, "seconds": 24.8, "max_memory_gb": 26.24, "pair": 0.9906478822231293, "sign": 0.5229155477136374, "keep": 10.6692955493927, "leak": 0.013328899862244725, "replacement": 6.590085744857788}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.5826730728149414, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 26.24, "pair": 1.081843912601471, "sign": 0.4285959005355835, "keep": 0.02329439902678132, "leak": 0.00021200634728302248, "replacement": 1.3078420758247375}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.0824040174484253, "loss_scale": 4.0, "retries": 0, "seconds": 27.4, "max_memory_gb": 26.24, "pair": 1.0351103246212006, "sign": 0.3706974424421787, "keep": 5.52883867919445, "leak": 0.003162539447657764, "replacement": 3.985669046640396}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.1416711807250977, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 26.24, "pair": 1.0162553787231445, "sign": 0.5666443035006523, "keep": 5.917574726045132, "leak": 0.018197935076386784, "replacement": 4.262914270162582}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.2560433149337769, "loss_scale": 4.0, "retries": 0, "seconds": 30.1, "max_memory_gb": 26.24, "pair": 0.9819339662790298, "sign": 0.24293902143836021, "keep": 5.305342623498291, "leak": 0.01754576520761475, "replacement": 3.760461211204529}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.0451858043670654, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 26.24, "pair": 0.9525516182184219, "sign": 0.24729623086750507, "keep": 10.004666566848755, "leak": 0.01390937389805913, "replacement": 6.082010388374329}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.6781518459320068, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 26.24, "pair": 0.9553306400775909, "sign": 0.3827568106353283, "keep": 8.426583029329777, "leak": 0.022080675129473093, "replacement": 5.365520864725113}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.0886754989624023, "loss_scale": 4.0, "retries": 0, "seconds": 34.0, "max_memory_gb": 26.24, "pair": 0.915023535490036, "sign": 0.6567376125603914, "keep": 10.454968094825745, "leak": 0.03825511655304581, "replacement": 6.480440020561218}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.6459357738494873, "loss_scale": 4.0, "retries": 0, "seconds": 35.3, "max_memory_gb": 26.24, "pair": 0.961297869682312, "sign": 0.578699316829443, "keep": 8.373142041265965, "leak": 0.026050121894513723, "replacement": 5.4437311589717865}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.2354787588119507, "loss_scale": 4.0, "retries": 0, "seconds": 36.6, "max_memory_gb": 26.24, "pair": 1.0212089121341705, "sign": 0.7679182328283787, "keep": 5.6001405864953995, "leak": 0.016040881047956645, "replacement": 4.2092485427856445}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.0369244813919067, "loss_scale": 4.0, "retries": 0, "seconds": 37.9, "max_memory_gb": 26.24, "pair": 1.0330918729305267, "sign": 0.8410519994795322, "keep": 5.199819220229983, "leak": 0.011516992324018815, "replacement": 4.056406676769257}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.4107723236083984, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 26.24, "pair": 0.9290031492710114, "sign": 0.34260603226721287, "keep": 5.487176734954119, "leak": 0.04929414944672317, "replacement": 3.856218010187149}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.3403222560882568, "loss_scale": 4.0, "retries": 0, "seconds": 40.5, "max_memory_gb": 26.24, "pair": 0.9652465730905533, "sign": 0.799534747377038, "keep": 5.659519135951996, "leak": 0.05288169598134118, "replacement": 4.207993865013123}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.353015422821045, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 26.24, "pair": 0.9075288325548172, "sign": 0.37013695016503334, "keep": 10.68119192123413, "leak": 0.05432000569999218, "replacement": 6.446773290634155}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.4034758806228638, "loss_scale": 4.0, "retries": 0, "seconds": 43.1, "max_memory_gb": 26.24, "pair": 0.9697510749101639, "sign": 0.6126157678663731, "keep": 4.865131134167314, "leak": 0.07552437519734667, "replacement": 3.727505683898926}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.4576473236083984, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 26.24, "pair": 0.9613554775714874, "sign": 0.5100995860993862, "keep": 5.251725897192955, "leak": 0.01954861691046972, "replacement": 3.8471553325653076}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.6895910501480103, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 26.24, "pair": 0.8903408497571945, "sign": 0.42212498001754284, "keep": 5.424521127715707, "leak": 0.06475819159823004, "replacement": 3.8298535346984863}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.5230289697647095, "loss_scale": 4.0, "retries": 0, "seconds": 47.1, "max_memory_gb": 26.24, "pair": 0.8897119164466858, "sign": 0.4239121852442622, "keep": 5.121568441390991, "leak": 0.05366144507934223, "replacement": 3.6758674681186676}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.6311699151992798, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 26.24, "pair": 0.9065506160259247, "sign": 0.5746775381267071, "keep": 4.874717473983765, "leak": 0.07927726042180439, "replacement": 3.6510675251483917}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 2.8599119186401367, "loss_scale": 4.0, "retries": 0, "seconds": 49.7, "max_memory_gb": 26.24, "pair": 0.8412971645593643, "sign": 0.7423112839460373, "keep": 9.349916458129883, "leak": 0.06513270933646709, "replacement": 5.903694152832031}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.237828254699707, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 26.24, "pair": 0.98200723528862, "sign": 0.5550199262797832, "keep": 2.8254664819687605, "leak": 0.06973698182991939, "replacement": 2.689684718847275}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.7975845336914062, "loss_scale": 4.0, "retries": 0, "seconds": 52.2, "max_memory_gb": 26.24, "pair": 0.9168369472026825, "sign": 0.48153940215706825, "keep": 5.451653316617012, "leak": 0.039928277634317055, "replacement": 3.893415331840515}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.8132663369178772, "loss_scale": 4.0, "retries": 0, "seconds": 53.5, "max_memory_gb": 26.24, "pair": 1.0430399477481842, "sign": 1.0166129022836685, "keep": 2.9452489763498306, "leak": 0.02213802751066396, "replacement": 3.0295053720474243}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.9154109954833984, "loss_scale": 4.0, "retries": 0, "seconds": 54.8, "max_memory_gb": 26.24, "pair": 0.9546625465154648, "sign": 0.39224419742822647, "keep": 4.564478258602321, "leak": 0.013530797557905316, "replacement": 3.4364064037799835}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.042510986328125, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 26.24, "pair": 0.9303853809833527, "sign": 0.5019699670374393, "keep": 4.753880531527102, "leak": 0.0254519122390775, "replacement": 3.564673662185669}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.2226648330688477, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 26.24, "pair": 0.9508914351463318, "sign": 0.5639284811913967, "keep": 2.73963293293491, "leak": 0.03805184579687193, "replacement": 2.612185001373291}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.155699372291565, "loss_scale": 4.0, "retries": 0, "seconds": 58.6, "max_memory_gb": 26.24, "pair": 1.0293048471212387, "sign": 0.8853859901428223, "keep": 2.7483887746930122, "leak": 0.01348212277298444, "replacement": 2.8495627343654633}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.0271589756011963, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 26.24, "pair": 0.9564754366874695, "sign": 0.7061262801289558, "keep": 4.480883374810219, "leak": 0.05515562775167382, "replacement": 3.5637693405151367}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.088067054748535, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 26.24, "pair": 0.8448833674192429, "sign": 0.6242450214922428, "keep": 3.984362667426467, "leak": 0.06199198486865498, "replacement": 3.1646851897239685}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 3.784099817276001, "loss_scale": 4.0, "retries": 0, "seconds": 62.5, "max_memory_gb": 26.24, "pair": 0.7853203862905502, "sign": 0.19218554999679327, "keep": 7.277070850133896, "leak": 0.11511808552313596, "replacement": 4.5487280786037445}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.452826738357544, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 26.24, "pair": 0.7589915320277214, "sign": 0.5804063202813268, "keep": 6.904181298799813, "leak": 0.15025866229552776, "replacement": 4.538850158452988}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 3.4376609325408936, "loss_scale": 4.0, "retries": 0, "seconds": 65.1, "max_memory_gb": 26.24, "pair": 0.890412300825119, "sign": 0.42356543242931366, "keep": 6.895865090191364, "leak": 0.06001718062816508, "replacement": 4.56513187289238}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.4187822341918945, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 26.24, "pair": 0.9472899436950684, "sign": 0.5689862854778767, "keep": 2.6231398275122046, "leak": 0.06889942538509786, "replacement": 2.560577869415283}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 2.7167959213256836, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 26.24, "pair": 0.84185741096735, "sign": 0.6738711036741734, "keep": 4.479057863354683, "leak": 0.1100383680750383, "replacement": 3.445831537246704}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 3.6179428100585938, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 26.24, "pair": 0.8921713531017303, "sign": 0.7111057341098785, "keep": 7.103083223104477, "leak": 0.10688797588591115, "replacement": 4.825987875461578}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 4.62586784362793, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 26.24, "pair": 0.767401397228241, "sign": 0.5929005183279514, "keep": 7.448542147874832, "leak": 0.16035034589003772, "replacement": 4.828210353851318}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 3.316765546798706, "loss_scale": 4.0, "retries": 0, "seconds": 71.5, "max_memory_gb": 26.24, "pair": 0.8541930913925171, "sign": 0.4041463080793619, "keep": 4.473435965832323, "leak": 0.11171466472296743, "replacement": 3.320912778377533}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 6.734230995178223, "loss_scale": 4.0, "retries": 0, "seconds": 72.8, "max_memory_gb": 26.24, "pair": 0.6027266755700111, "sign": 0.1544315218925476, "keep": 8.515976667404175, "leak": 0.19276490062475204, "replacement": 4.986122012138367}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.605574369430542, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 26.24, "pair": 0.967841163277626, "sign": 0.7853552848100662, "keep": 2.3114646300673485, "leak": 0.01703963364707306, "replacement": 2.520511031150818}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 4.256528854370117, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 26.24, "pair": 0.886822521686554, "sign": 0.27109055407345295, "keep": 4.303032236173749, "leak": 0.18631922954227775, "replacement": 3.220463752746582}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.8228118419647217, "loss_scale": 4.0, "retries": 0, "seconds": 76.7, "max_memory_gb": 26.24, "pair": 0.9709933698177338, "sign": 0.6523961499333382, "keep": 2.6742199137806892, "leak": 0.006140356716059614, "replacement": 2.6358364522457123}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 4.09442663192749, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 26.24, "pair": 0.7905889302492142, "sign": 0.6108608283102512, "keep": 4.321577820926905, "leak": 0.07707651820965111, "replacement": 3.276077389717102}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 7.828238010406494, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 26.24, "pair": 0.8583802878856659, "sign": 0.6658555250614882, "keep": 5.596278965473175, "leak": 0.17435806929279352, "replacement": 4.033037006855011}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 2.8645412921905518, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 26.24, "pair": 0.94722481071949, "sign": 0.6977633535861969, "keep": 2.386086044833064, "leak": 0.01804576604263275, "replacement": 2.493661016225815}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 10.098966598510742, "loss_scale": 4.0, "retries": 0, "seconds": 81.8, "max_memory_gb": 26.24, "pair": 0.8681772947311401, "sign": 0.5711559243500233, "keep": 5.944955317303538, "leak": 0.07344758386170724, "replacement": 4.144594758749008}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 9.20711898803711, "loss_scale": 4.0, "retries": 0, "seconds": 83.1, "max_memory_gb": 26.28, "pair": 0.7100175023078918, "sign": 0.42031703144311905, "keep": 5.737823490053415, "leak": 0.26473865512525663, "replacement": 3.8552724421024323}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 9.57674789428711, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 26.28, "pair": 0.8650800734758377, "sign": 0.5402820780873299, "keep": 5.708104491233826, "leak": 0.12901008162589278, "replacement": 4.021525979042053}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 2.0674540996551514, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 26.28, "pair": 1.078211486339569, "sign": 0.6351999938488007, "keep": 1.6412533142138273, "leak": 0.024634502866234698, "replacement": 2.2225968539714813}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 10.932002067565918, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 26.28, "pair": 0.7880852818489075, "sign": 0.3779271226376295, "keep": 4.141229957342148, "leak": 0.3558943306852598, "replacement": 3.1366374492645264}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 3.60221004486084, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 26.28, "pair": 0.9969815909862518, "sign": 0.5078891701996326, "keep": 1.939588223118335, "leak": 0.05720963105704868, "replacement": 2.235022723674774}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 9.186724662780762, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 26.28, "pair": 0.8855605274438858, "sign": 0.6928990464657545, "keep": 2.902377961203456, "leak": 0.08729104575468227, "replacement": 2.7050216794013977}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 5.887152194976807, "loss_scale": 4.0, "retries": 0, "seconds": 91.0, "max_memory_gb": 26.28, "pair": 1.0864536315202713, "sign": 0.6018466651439667, "keep": 2.560335504822433, "leak": 0.4549506386683788, "replacement": 2.7812823355197906}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 11.1754150390625, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 26.28, "pair": 0.884275272488594, "sign": 0.17197940312325954, "keep": 3.8394983764737844, "leak": 0.2944657215557527, "replacement": 2.9636306166648865}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 12.041654586791992, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 26.28, "pair": 0.7516232579946518, "sign": 0.24547307565808296, "keep": 4.355483174324036, "leak": 0.42745141685009, "replacement": 3.158964157104492}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.3923426866531372, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 26.28, "pair": 1.0246122777462006, "sign": 0.8004675805568695, "keep": 0.43701935838907957, "leak": 0.0078083657717797905, "replacement": 1.6453078985214233}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 5.886599063873291, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 26.28, "pair": 1.0895183235406876, "sign": 0.5064669568091631, "keep": 1.756326887756586, "leak": 0.1472210648989858, "replacement": 2.257720470428467}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.2571070194244385, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 26.28, "pair": 0.9127620533108711, "sign": 0.896573681384325, "keep": 2.2746195942163467, "leak": 0.14168464901013067, "replacement": 2.533779889345169}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 6.23856782913208, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 26.28, "pair": 0.9019467234611511, "sign": 0.4521471615880728, "keep": 1.6370246764272451, "leak": 0.07629265554714948, "replacement": 1.9656058251857758}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 9.659318923950195, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 26.28, "pair": 0.9630685299634933, "sign": 0.4843645300716162, "keep": 3.032422512769699, "leak": 0.4361243681050837, "replacement": 2.830493152141571}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 4.543478965759277, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 26.28, "pair": 1.016059324145317, "sign": 0.7287628129124641, "keep": 1.6887209918349981, "leak": 0.44634662255703006, "replacement": 2.336387872695923}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 3.650350570678711, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 26.28, "pair": 0.9598175436258316, "sign": 0.5835633650422096, "keep": 0.668402710929513, "leak": 0.008025233808439225, "replacement": 1.5878068804740906}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 6.079843044281006, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 26.28, "pair": 1.1541147381067276, "sign": 0.4308520257472992, "keep": 1.10908461548388, "leak": 0.12775821710238233, "replacement": 1.9560226500034332}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 8.233105659484863, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 26.28, "pair": 0.8068059682846069, "sign": 0.4751175716519356, "keep": 1.6979510188102722, "leak": 0.7072193069616333, "replacement": 2.0701451003551483}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 8.045095443725586, "loss_scale": 4.0, "retries": 0, "seconds": 106.4, "max_memory_gb": 26.28, "pair": 1.0959770232439041, "sign": 0.42705693282186985, "keep": 1.541482761502266, "leak": 0.10471698488254333, "replacement": 2.10642609000206}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 7.938604354858398, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 26.28, "pair": 0.830189436674118, "sign": 0.4048917852342129, "keep": 1.6631451547145844, "leak": 0.6452835545642301, "replacement": 2.025528848171234}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 5.102238178253174, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 26.28, "pair": 1.0795790553092957, "sign": 0.7457985579967499, "keep": 0.6993849715217948, "leak": 0.5827333485940471, "replacement": 1.9478541910648346}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 3.769973039627075, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 26.28, "pair": 1.0831912755966187, "sign": 0.5874879136681557, "keep": 0.1467712577432394, "leak": 0.014066603063838556, "replacement": 1.453837513923645}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 7.881194591522217, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 26.28, "pair": 1.0279593765735626, "sign": 0.669013038277626, "keep": 1.2020038813352585, "leak": 0.08229487887729192, "replacement": 1.9840416610240936}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 6.555176258087158, "loss_scale": 4.0, "retries": 0, "seconds": 113.0, "max_memory_gb": 26.28, "pair": 0.7422901019454002, "sign": 0.3952004984021187, "keep": 0.9008968137204647, "leak": 1.028897616546601, "replacement": 1.6475632190704346}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 4.781548500061035, "loss_scale": 4.0, "retries": 0, "seconds": 114.4, "max_memory_gb": 26.28, "pair": 0.9906463027000427, "sign": 0.6513114906847477, "keep": 0.8308197697624564, "leak": 0.09689455328771146, "replacement": 1.7559355795383453}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 3.425685405731201, "loss_scale": 4.0, "retries": 0, "seconds": 115.7, "max_memory_gb": 26.28, "pair": 1.1218043267726898, "sign": 0.713973194360733, "keep": 0.23934369115158916, "leak": 0.013202682486735284, "replacement": 1.6017634570598602}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 4.380420684814453, "loss_scale": 4.0, "retries": 0, "seconds": 117.0, "max_memory_gb": 26.28, "pair": 0.7829633355140686, "sign": 0.5425708517432213, "keep": 0.7418160364031792, "leak": 0.3718793768202886, "replacement": 1.5181266069412231}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.3918722867965698, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 26.28, "pair": 0.9988702535629272, "sign": 0.9923316836357117, "keep": 0.6324958354234695, "leak": 0.013892650575144216, "replacement": 1.8147571384906769}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.597949266433716, "loss_scale": 4.0, "retries": 0, "seconds": 119.8, "max_memory_gb": 26.28, "pair": 0.9881996810436249, "sign": 0.5439524278044701, "keep": 0.3365656426176429, "leak": 1.265946980798617, "replacement": 1.7449454367160797}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 4.709416389465332, "loss_scale": 4.0, "retries": 0, "seconds": 121.0, "max_memory_gb": 26.28, "pair": 0.8346296697854996, "sign": 0.3045900668948889, "keep": 0.6842200406827033, "leak": 0.022779468483349774, "replacement": 1.3347295820713043}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 7.842708587646484, "loss_scale": 4.0, "retries": 0, "seconds": 122.3, "max_memory_gb": 26.28, "pair": 0.6158931329846382, "sign": 0.366134125739336, "keep": 1.1485817581415176, "leak": 0.5308401199436048, "replacement": 1.5059611201286316}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 2.8795173168182373, "loss_scale": 4.0, "retries": 0, "seconds": 123.6, "max_memory_gb": 26.28, "pair": 0.6774954348802567, "sign": 0.4634505547583103, "keep": 0.8314544074237347, "leak": 0.027296643937006593, "replacement": 1.3317721486091614}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 6.804643154144287, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 26.28, "pair": 0.8323866948485374, "sign": 0.42078988812863827, "keep": 0.6208122372627258, "leak": 2.646253564962535, "replacement": 2.014751225709915}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.3252344131469727, "loss_scale": 4.0, "retries": 0, "seconds": 126.2, "max_memory_gb": 26.28, "pair": 0.8632584512233734, "sign": 0.5125108473002911, "keep": 0.4980766586959362, "leak": 0.30509512266144156, "replacement": 1.4448260366916656}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 2.786180019378662, "loss_scale": 4.0, "retries": 0, "seconds": 127.5, "max_memory_gb": 26.28, "pair": 0.763786893337965, "sign": 0.4999072141945362, "keep": 0.28198765870183706, "leak": 0.7497025050688535, "replacement": 1.342159926891327}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 1.9015005826950073, "loss_scale": 4.0, "retries": 0, "seconds": 128.8, "max_memory_gb": 26.28, "pair": 1.0089227110147476, "sign": 0.8782386407256126, "keep": 0.38865082152187824, "leak": 0.017365177045576274, "replacement": 1.6467088162899017}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 3.226382255554199, "loss_scale": 4.0, "retries": 0, "seconds": 130.1, "max_memory_gb": 26.28, "pair": 0.681635670363903, "sign": 0.6077931374311447, "keep": 0.6480204313993454, "leak": 1.3047690604871605, "replacement": 1.6357347071170807}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 4.350858211517334, "loss_scale": 4.0, "retries": 0, "seconds": 131.4, "max_memory_gb": 26.28, "pair": 0.5696258880198002, "sign": 0.1704951897263527, "keep": 0.4511105418205261, "leak": 0.22670080442912877, "replacement": 0.9371039867401123}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 3.4938714504241943, "loss_scale": 4.0, "retries": 0, "seconds": 132.7, "max_memory_gb": 26.28, "pair": 0.6290733814239502, "sign": 0.3190812058746815, "keep": 1.112064503133297, "leak": 0.8228604774922132, "replacement": 1.5503613650798798}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 10.321208000183105, "loss_scale": 4.0, "retries": 0, "seconds": 134.0, "max_memory_gb": 26.28, "pair": 0.683315709233284, "sign": 0.3509153239428997, "keep": 0.3336226618848741, "leak": 1.465833568945527, "replacement": 1.392043113708496}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.780162811279297, "loss_scale": 4.0, "retries": 0, "seconds": 135.3, "max_memory_gb": 26.28, "pair": 0.9179143607616425, "sign": 0.9149934761226177, "keep": 0.1531065459130332, "leak": 0.08497742447070777, "replacement": 1.473208725452423}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 2.0069165229797363, "loss_scale": 4.0, "retries": 0, "seconds": 136.6, "max_memory_gb": 26.28, "pair": 0.5857428908348083, "sign": 0.38136699236929417, "keep": 0.3452357202768326, "leak": 2.2164857847383246, "replacement": 1.5031656920909882}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 2.4084455966949463, "loss_scale": 4.0, "retries": 0, "seconds": 137.9, "max_memory_gb": 26.28, "pair": 0.5161693543195724, "sign": 0.42997327260673046, "keep": 0.342244029045105, "leak": 2.3522993996739388, "replacement": 1.4903528690338135}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.000548839569092, "loss_scale": 4.0, "retries": 0, "seconds": 139.2, "max_memory_gb": 26.28, "pair": 0.8946164846420288, "sign": 0.7256549596786499, "keep": 0.30875128880143166, "leak": 0.0400329266267363, "replacement": 1.4218278527259827}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 1.5637624263763428, "loss_scale": 4.0, "retries": 0, "seconds": 140.5, "max_memory_gb": 26.28, "pair": 0.8425109535455704, "sign": 0.7031848579645157, "keep": 0.2568186251446605, "leak": 1.6489533103667782, "replacement": 1.7347509860992432}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 2.732335329055786, "loss_scale": 4.0, "retries": 0, "seconds": 141.8, "max_memory_gb": 26.28, "pair": 0.44945361092686653, "sign": 0.4085944937542081, "keep": 0.5295266583561897, "leak": 1.9689613743685186, "replacement": 1.410754531621933}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 2.17925763130188, "loss_scale": 4.0, "retries": 0, "seconds": 143.1, "max_memory_gb": 26.28, "pair": 0.6513842213898897, "sign": 0.5181365869939327, "keep": 0.4232609197497368, "leak": 0.8914252930117073, "replacement": 1.3449393510818481}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 1.434227466583252, "loss_scale": 4.0, "retries": 0, "seconds": 144.4, "max_memory_gb": 26.28, "pair": 0.5451282486319542, "sign": 0.41458010813221335, "keep": 0.15853668842464685, "leak": 0.2409735657274723, "replacement": 0.8919300585985184}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 1.5558897256851196, "loss_scale": 4.0, "retries": 0, "seconds": 145.7, "max_memory_gb": 26.28, "pair": 0.634321466088295, "sign": 0.43174146860837936, "keep": 0.25457928888499737, "leak": 0.429432631470263, "replacement": 1.084840014576912}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 1.1798021793365479, "loss_scale": 4.0, "retries": 0, "seconds": 147.0, "max_memory_gb": 26.28, "pair": 0.6137360036373138, "sign": 0.5109672285616398, "keep": 0.22805781289935112, "leak": 0.49187434517079964, "replacement": 1.1062171086668968}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 2.243992567062378, "loss_scale": 4.0, "retries": 0, "seconds": 148.3, "max_memory_gb": 26.28, "pair": 0.4771809056401253, "sign": 0.4643605099990964, "keep": 0.6160763651132584, "leak": 1.5521125216037035, "replacement": 1.4054274559020996}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 1.479713797569275, "loss_scale": 4.0, "retries": 0, "seconds": 149.6, "max_memory_gb": 26.28, "pair": 0.6865103915333748, "sign": 0.5629608295857906, "keep": 0.38540604896843433, "leak": 1.1695355709325668, "replacement": 1.4530777633190155}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 1.53684663772583, "loss_scale": 4.0, "retries": 0, "seconds": 150.9, "max_memory_gb": 26.28, "pair": 0.835921123623848, "sign": 0.6598930656909943, "keep": 0.21545881684869528, "leak": 0.07818147311627399, "replacement": 1.293142393231392}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 2.0759646892547607, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 26.28, "pair": 0.6266291067004204, "sign": 0.41581564117223024, "keep": 0.5276559616904706, "leak": 1.0831583850085735, "replacement": 1.3691545873880386}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 1.0671976804733276, "loss_scale": 4.0, "retries": 0, "seconds": 153.5, "max_memory_gb": 26.28, "pair": 0.73345997184515, "sign": 0.5250876918435097, "keep": 0.287065704818815, "leak": 0.2432485816971166, "replacement": 1.2003488093614578}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 1.407050371170044, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 26.28, "pair": 0.854147806763649, "sign": 0.7864150926470757, "keep": 0.21392295509576797, "leak": 0.2294278598856181, "replacement": 1.4116738140583038}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 1.6823527812957764, "loss_scale": 4.0, "retries": 0, "seconds": 156.1, "max_memory_gb": 26.28, "pair": 0.5587362237274647, "sign": 0.47389326244592667, "keep": 0.6298736119642854, "leak": 0.8701012771343812, "replacement": 1.3281449601054192}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 1.5314908027648926, "loss_scale": 4.0, "retries": 0, "seconds": 157.4, "max_memory_gb": 26.28, "pair": 0.7841966673731804, "sign": 0.6817876845598221, "keep": 0.5604234240017831, "leak": 2.186377266421914, "replacement": 1.951896607875824}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 1.595555067062378, "loss_scale": 4.0, "retries": 0, "seconds": 158.7, "max_memory_gb": 26.28, "pair": 0.6424619033932686, "sign": 0.6250224355608225, "keep": 0.22082282858900726, "leak": 0.4337728620347727, "replacement": 1.1738277897238731}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.401887059211731, "loss_scale": 4.0, "retries": 0, "seconds": 160.0, "max_memory_gb": 26.28, "pair": 0.7448896169662476, "sign": 0.7645842088386416, "keep": 0.8493914026767015, "leak": 0.5045858476078138, "replacement": 1.678023874759674}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 1.6765859127044678, "loss_scale": 4.0, "retries": 0, "seconds": 161.3, "max_memory_gb": 26.28, "pair": 0.9156660586595535, "sign": 1.0399855822324753, "keep": 0.4457110408693552, "leak": 0.05568512686295435, "replacement": 1.6724356710910797}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 1.3844189643859863, "loss_scale": 4.0, "retries": 0, "seconds": 162.6, "max_memory_gb": 26.28, "pair": 0.8351786434650421, "sign": 0.8171327039599419, "keep": 0.24580371752381325, "leak": 0.1820507359225303, "replacement": 1.412159502506256}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 0.9098389744758606, "loss_scale": 4.0, "retries": 0, "seconds": 163.8, "max_memory_gb": 26.28, "pair": 0.6239129900932312, "sign": 0.6033070622943342, "keep": 0.573647141456604, "leak": 0.36221634995308705, "replacement": 1.30294419080019}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 1.2709945440292358, "loss_scale": 4.0, "retries": 0, "seconds": 165.1, "max_memory_gb": 26.28, "pair": 0.6498848870396614, "sign": 0.3624341916292906, "keep": 0.09981091343797743, "leak": 0.29943105019629, "replacement": 0.9558651894330978}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.834505558013916, "loss_scale": 4.0, "retries": 0, "seconds": 166.4, "max_memory_gb": 26.28, "pair": 0.4763994626700878, "sign": 0.30667515238747, "keep": 0.21360777132213116, "leak": 1.1988670115824789, "replacement": 1.0362576991319656}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 1.4446061849594116, "loss_scale": 4.0, "retries": 0, "seconds": 167.7, "max_memory_gb": 26.28, "pair": 0.3327387310564518, "sign": 0.1768538192845881, "keep": 0.13253262918442488, "leak": 2.2040976700372994, "replacement": 1.038456380367279}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 2.341017246246338, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 26.28, "pair": 0.7095905877649784, "sign": 0.8411542382091284, "keep": 0.5656975265592337, "leak": 0.5656663803383708, "replacement": 1.554433062672615}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 1.287072777748108, "loss_scale": 4.0, "retries": 0, "seconds": 170.3, "max_memory_gb": 26.28, "pair": 0.7699546664953232, "sign": 0.6940069310367107, "keep": 0.4529806524515152, "leak": 0.21037667179559838, "replacement": 1.3960426449775696}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 1.8756364583969116, "loss_scale": 4.0, "retries": 0, "seconds": 171.6, "max_memory_gb": 26.28, "pair": 0.3829289432615042, "sign": 0.12193380435928702, "keep": 0.7878087162971497, "leak": 0.7642883509397507, "replacement": 1.0288722813129425}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 1.685745120048523, "loss_scale": 4.0, "retries": 0, "seconds": 172.9, "max_memory_gb": 26.28, "pair": 0.4206215888261795, "sign": 0.38254317827522755, "keep": 0.48589410446584225, "leak": 2.498449087550398, "replacement": 1.4794525131583214}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 1.9483473300933838, "loss_scale": 4.0, "retries": 0, "seconds": 174.2, "max_memory_gb": 26.28, "pair": 0.799943745136261, "sign": 0.6754163727164268, "keep": 0.34379058750346303, "leak": 0.16674796701408923, "replacement": 1.3512341529130936}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.199195623397827, "loss_scale": 4.0, "retries": 0, "seconds": 175.5, "max_memory_gb": 26.28, "pair": 0.4062645062804222, "sign": 0.3917856593616307, "keep": 0.24434961937367916, "leak": 3.4992526816786267, "replacement": 1.5991453230381012}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 2.4506781101226807, "loss_scale": 4.0, "retries": 0, "seconds": 176.8, "max_memory_gb": 26.28, "pair": 0.6021690145134926, "sign": 0.4655250683426857, "keep": 0.46472291834652424, "leak": 0.21281418576836586, "replacement": 1.1204965561628342}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 3.9439215660095215, "loss_scale": 4.0, "retries": 0, "seconds": 178.1, "max_memory_gb": 26.28, "pair": 0.5719684809446335, "sign": 0.45466811303049326, "keep": 0.21139795565977693, "leak": 3.4859390603378415, "replacement": 1.7764862477779388}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 1.4526786804199219, "loss_scale": 4.0, "retries": 0, "seconds": 179.4, "max_memory_gb": 26.28, "pair": 0.48675331473350525, "sign": 0.28134636767208576, "keep": 0.20667701121419668, "leak": 0.16937348051578738, "replacement": 0.7731083631515503}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 3.135385036468506, "loss_scale": 4.0, "retries": 0, "seconds": 180.8, "max_memory_gb": 26.28, "pair": 0.5416441187262535, "sign": 0.5017457157373428, "keep": 0.34442760795354843, "leak": 0.9365443917922676, "replacement": 1.198866829276085}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 2.219248056411743, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 26.28, "pair": 0.4807816110551357, "sign": 0.4684967575594783, "keep": 0.1407792279496789, "leak": 1.019122902303934, "replacement": 1.040200337767601}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 1.7112716436386108, "loss_scale": 4.0, "retries": 0, "seconds": 183.4, "max_memory_gb": 26.28, "pair": 0.5623053461313248, "sign": 0.49366394989192486, "keep": 0.3086505183018744, "leak": 1.6327698673121631, "replacement": 1.3716550171375275}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 1.392767071723938, "loss_scale": 4.0, "retries": 0, "seconds": 184.7, "max_memory_gb": 26.28, "pair": 0.5314321294426918, "sign": 0.39588592760264874, "keep": 0.21612591249868274, "leak": 0.38453164891689084, "replacement": 0.933570921421051}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 1.548409342765808, "loss_scale": 4.0, "retries": 0, "seconds": 186.0, "max_memory_gb": 26.28, "pair": 0.8950827717781067, "sign": 1.0085053741931915, "keep": 0.22699469106737524, "leak": 0.1420137146487832, "replacement": 1.5483362674713135}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 1.9404168128967285, "loss_scale": 4.0, "retries": 0, "seconds": 187.3, "max_memory_gb": 26.28, "pair": 0.4894108362495899, "sign": 0.3598102703690529, "keep": 0.2567413169890642, "leak": 1.4503488996997476, "replacement": 1.1602738797664642}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 2.0269083976745605, "loss_scale": 4.0, "retries": 0, "seconds": 188.6, "max_memory_gb": 26.28, "pair": 0.589532133191824, "sign": 0.3082914175465703, "keep": 0.07100203447043896, "leak": 0.9681924358010292, "replacement": 1.0212269872426987}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.43837571144104, "loss_scale": 4.0, "retries": 0, "seconds": 189.9, "max_memory_gb": 26.28, "pair": 0.5232813823968172, "sign": 0.5375840030610561, "keep": 0.2220409568399191, "leak": 1.0647713124053553, "replacement": 1.169286698102951}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 1.459299921989441, "loss_scale": 4.0, "retries": 0, "seconds": 191.2, "max_memory_gb": 26.28, "pair": 0.5910373702645302, "sign": 0.5137101020663977, "keep": 0.4241157879587263, "leak": 0.7511243284679949, "replacement": 1.2477314472198486}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.7600682973861694, "loss_scale": 4.0, "retries": 0, "seconds": 192.5, "max_memory_gb": 26.28, "pair": 0.6281110644340515, "sign": 0.3199024572968483, "keep": 0.03554299275856465, "leak": 0.20368332043290138, "replacement": 0.8567546010017395}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 1.5086348056793213, "loss_scale": 4.0, "retries": 0, "seconds": 193.8, "max_memory_gb": 26.28, "pair": 0.5797283388674259, "sign": 0.459690323099494, "keep": 0.16478510689921677, "leak": 0.14543216329911957, "replacement": 0.9283240884542465}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 1.8655604124069214, "loss_scale": 4.0, "retries": 0, "seconds": 195.1, "max_memory_gb": 26.28, "pair": 0.39140673726797104, "sign": 0.32979446463286877, "keep": 0.41695269756019115, "leak": 2.650240143761039, "replacement": 1.4273403584957123}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 1.5999248027801514, "loss_scale": 4.0, "retries": 0, "seconds": 196.5, "max_memory_gb": 26.28, "pair": 0.5057023838162422, "sign": 0.42265911772847176, "keep": 0.2728613328654319, "leak": 0.3716084733605385, "replacement": 0.9463647305965424}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 2.6026201248168945, "loss_scale": 4.0, "retries": 0, "seconds": 197.8, "max_memory_gb": 26.28, "pair": 0.295576861128211, "sign": 0.20987162087112665, "keep": 0.1861059619113803, "leak": 2.45265556871891, "replacement": 1.1067295372486115}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 1.772740125656128, "loss_scale": 4.0, "retries": 0, "seconds": 199.1, "max_memory_gb": 26.28, "pair": 0.4111120067536831, "sign": 0.258577112108469, "keep": 0.07717307144775987, "leak": 1.716275256127119, "replacement": 1.0080559253692627}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.203242778778076, "loss_scale": 4.0, "retries": 0, "seconds": 200.4, "max_memory_gb": 26.28, "pair": 0.33458771370351315, "sign": 0.1961483098566532, "keep": 0.7397934840992093, "leak": 0.8641058467328548, "replacement": 1.0185850635170937}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 1.7969093322753906, "loss_scale": 4.0, "retries": 0, "seconds": 201.8, "max_memory_gb": 26.28, "pair": 0.48136358708143234, "sign": 0.3966544419527054, "keep": 0.8902145328465849, "leak": 2.634773114696145, "replacement": 1.7834913581609726}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.3637731075286865, "loss_scale": 4.0, "retries": 0, "seconds": 203.0, "max_memory_gb": 26.28, "pair": 0.6210272125899792, "sign": 0.5281086172908545, "keep": 0.24067135096993297, "leak": 0.4346567213651724, "replacement": 1.1140813678503036}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 1.8030234575271606, "loss_scale": 4.0, "retries": 0, "seconds": 204.3, "max_memory_gb": 26.28, "pair": 0.7969665974378586, "sign": 1.1367955803871155, "keep": 0.2609065165743232, "leak": 0.16175032609316986, "replacement": 1.5362551808357239}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.9493885040283203, "loss_scale": 4.0, "retries": 0, "seconds": 205.6, "max_memory_gb": 26.28, "pair": 0.6475505121052265, "sign": 0.6676731631159782, "keep": 0.4817018862813711, "leak": 0.2682897476479411, "replacement": 1.2893104553222656}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 2.10819149017334, "loss_scale": 4.0, "retries": 0, "seconds": 206.9, "max_memory_gb": 26.28, "pair": 0.31763802096247673, "sign": 0.17686053179204464, "keep": 0.10945925023406744, "leak": 2.1306699588894844, "replacement": 0.9934653639793396}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 2.808594226837158, "loss_scale": 4.0, "retries": 0, "seconds": 208.2, "max_memory_gb": 26.28, "pair": 0.468379121273756, "sign": 0.3462362417485565, "keep": 0.24579384177923203, "leak": 2.0803846959024668, "replacement": 1.284490317106247}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 1.4405018091201782, "loss_scale": 4.0, "retries": 0, "seconds": 209.5, "max_memory_gb": 26.28, "pair": 0.6499036084860563, "sign": 0.6020582895725965, "keep": 0.4090424180030823, "leak": 0.5241418373771012, "replacement": 1.2864893674850464}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 2.3493223190307617, "loss_scale": 4.0, "retries": 0, "seconds": 210.8, "max_memory_gb": 26.28, "pair": 0.6412020493298769, "sign": 0.612035364843905, "keep": 0.13600393990054727, "leak": 0.3025529936130624, "replacement": 1.090859942138195}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 1.7415398359298706, "loss_scale": 4.0, "retries": 0, "seconds": 212.0, "max_memory_gb": 26.28, "pair": 0.43718211352825165, "sign": 0.3115138257853687, "keep": 0.09525698842480779, "leak": 0.9447533711791039, "replacement": 0.8767558485269547}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 2.2389369010925293, "loss_scale": 4.0, "retries": 0, "seconds": 213.3, "max_memory_gb": 26.28, "pair": 0.7083521634340286, "sign": 0.9100967971608043, "keep": 0.4618617072701454, "leak": 0.6352609017048962, "replacement": 1.5531466752290726}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 2.3607218265533447, "loss_scale": 4.0, "retries": 0, "seconds": 214.7, "max_memory_gb": 26.28, "pair": 0.5180612206459045, "sign": 0.5563726350665092, "keep": 0.19432740239426494, "leak": 1.5223089675419033, "replacement": 1.273988425731659}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 1.70037841796875, "loss_scale": 4.0, "retries": 0, "seconds": 216.0, "max_memory_gb": 26.28, "pair": 0.5018337070941925, "sign": 0.5200373120605946, "keep": 0.28828637255355716, "leak": 1.2002033787102846, "replacement": 1.2060463577508926}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 2.4452288150787354, "loss_scale": 4.0, "retries": 0, "seconds": 217.3, "max_memory_gb": 26.28, "pair": 0.3574913078919053, "sign": 0.38231352902948856, "keep": 0.5771776884794235, "leak": 1.3817367086885497, "replacement": 1.1826710551977158}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.1851067543029785, "loss_scale": 4.0, "retries": 0, "seconds": 218.6, "max_memory_gb": 26.28, "pair": 0.45074405148625374, "sign": 0.4250815073028207, "keep": 0.14453065674751997, "leak": 1.7641930803656578, "replacement": 1.1765984147787094}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 3.3060455322265625, "loss_scale": 4.0, "retries": 0, "seconds": 219.9, "max_memory_gb": 26.28, "pair": 0.4487251304090023, "sign": 0.41394155845046043, "keep": 0.7405282445251942, "leak": 2.1969690583646297, "replacement": 1.575202316045761}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 1.2632709741592407, "loss_scale": 4.0, "retries": 0, "seconds": 221.2, "max_memory_gb": 26.28, "pair": 0.29764965921640396, "sign": 0.10359812155365944, "keep": 0.056231893599033356, "leak": 1.1017877236008644, "replacement": 0.6530115827918053}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 3.143449068069458, "loss_scale": 4.0, "retries": 0, "seconds": 222.5, "max_memory_gb": 26.28, "pair": 0.5093683302402496, "sign": 0.40561402775347233, "keep": 0.11364666512235999, "leak": 1.8456861013546586, "replacement": 1.2304202169179916}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 2.7436375617980957, "loss_scale": 4.0, "retries": 0, "seconds": 223.9, "max_memory_gb": 26.28, "pair": 0.24711227789521217, "sign": 0.08119794633239508, "keep": 0.2585061192512512, "leak": 4.519988775951788, "replacement": 1.5469615012407303}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 3.9552223682403564, "loss_scale": 4.0, "retries": 0, "seconds": 225.1, "max_memory_gb": 26.28, "pair": 0.2967883348464966, "sign": 0.4011512631550431, "keep": 0.32629474997520447, "leak": 1.7604416459798813, "replacement": 1.1006217449903488}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 2.2373909950256348, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 26.28, "pair": 0.43231165409088135, "sign": 0.26426422595977783, "keep": 0.20835980540141463, "leak": 1.1344203352928162, "replacement": 0.9522287398576736}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 1.8313285112380981, "loss_scale": 4.0, "retries": 0, "seconds": 227.7, "max_memory_gb": 26.28, "pair": 0.5043192356824875, "sign": 0.5306191686540842, "keep": 0.2712798975408077, "leak": 1.5437520602717996, "replacement": 1.291206806898117}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 2.3106207847595215, "loss_scale": 4.0, "retries": 0, "seconds": 229.1, "max_memory_gb": 26.28, "pair": 0.3705260530114174, "sign": 0.13081822078675032, "keep": 0.19902560114860535, "leak": 2.1021715570241213, "replacement": 1.0609908550977707}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 4.363550662994385, "loss_scale": 4.0, "retries": 0, "seconds": 230.3, "max_memory_gb": 26.28, "pair": 0.5223667845129967, "sign": 0.34388440754264593, "keep": 0.34572343807667494, "leak": 0.6119161032183911, "replacement": 1.0201497077941895}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 1.7965046167373657, "loss_scale": 4.0, "retries": 0, "seconds": 231.6, "max_memory_gb": 26.28, "pair": 0.5815845429897308, "sign": 0.5495638754218817, "keep": 0.24782701581716537, "leak": 1.492035983595997, "replacement": 1.3532889634370804}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 1.3196412324905396, "loss_scale": 4.0, "retries": 0, "seconds": 232.9, "max_memory_gb": 26.28, "pair": 0.7300626672804356, "sign": 0.7677294705063105, "keep": 0.5997968683950603, "leak": 0.6333117734466214, "replacement": 1.5721537619829178}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 7.055172920227051, "loss_scale": 4.0, "retries": 0, "seconds": 234.2, "max_memory_gb": 26.28, "pair": 0.7223009839653969, "sign": 0.60124721378088, "keep": 0.8807626701891422, "leak": 0.3182043003325816, "replacement": 1.5428570061922073}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 2.213841438293457, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 26.28, "pair": 0.321810957044363, "sign": 0.1325590922497213, "keep": 0.1817077249288559, "leak": 2.778307645319728, "replacement": 1.1735212877392769}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 3.2929911613464355, "loss_scale": 4.0, "retries": 0, "seconds": 236.8, "max_memory_gb": 26.28, "pair": 0.6424781456589699, "sign": 0.6023090546950698, "keep": 0.5873486786149442, "leak": 0.6648549241945148, "replacement": 1.4035207629203796}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.563148260116577, "loss_scale": 4.0, "retries": 0, "seconds": 238.1, "max_memory_gb": 26.28, "pair": 0.24996141716837883, "sign": 0.14035805547609925, "keep": 0.052985430927947164, "leak": 1.3700955510139465, "replacement": 0.6891570389270782}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 3.603119134902954, "loss_scale": 4.0, "retries": 0, "seconds": 239.4, "max_memory_gb": 26.28, "pair": 0.5040777213871479, "sign": 0.33630717173218727, "keep": 0.15222404338419437, "leak": 0.8725447095930576, "replacement": 0.9664794951677322}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 3.347978353500366, "loss_scale": 4.0, "retries": 0, "seconds": 240.7, "max_memory_gb": 26.28, "pair": 0.3954494297504425, "sign": 0.2548888921737671, "keep": 0.10867577977478504, "leak": 3.3465014453977346, "replacement": 1.4138571321964264}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 2.0771420001983643, "loss_scale": 4.0, "retries": 0, "seconds": 242.0, "max_memory_gb": 26.28, "pair": 0.38810741901397705, "sign": 0.23989183362573385, "keep": 0.16534356190823019, "leak": 0.8614366762340069, "replacement": 0.8060842901468277}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 2.513688564300537, "loss_scale": 4.0, "retries": 0, "seconds": 243.4, "max_memory_gb": 26.28, "pair": 0.34105599485337734, "sign": 0.07365063671022654, "keep": 0.17370589077472687, "leak": 1.5443743541836739, "replacement": 0.8508278578519821}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 1.9268544912338257, "loss_scale": 4.0, "retries": 0, "seconds": 244.7, "max_memory_gb": 26.28, "pair": 0.5967252999544144, "sign": 0.5719348005950451, "keep": 0.24734684987924993, "leak": 0.3335828837589361, "replacement": 1.0897618681192398}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.3640427589416504, "loss_scale": 4.0, "retries": 0, "seconds": 246.0, "max_memory_gb": 26.28, "pair": 0.4047287106513977, "sign": 0.22576406225562096, "keep": 0.10773387318477035, "leak": 1.040002271067351, "replacement": 0.8314782455563545}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 2.111743211746216, "loss_scale": 4.0, "retries": 0, "seconds": 247.2, "max_memory_gb": 26.28, "pair": 0.31795232743024826, "sign": 0.2017816212028265, "keep": 0.10127279348671436, "leak": 1.346378207206726, "replacement": 0.8060740977525711}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 3.0252676010131836, "loss_scale": 4.0, "retries": 0, "seconds": 248.5, "max_memory_gb": 26.28, "pair": 0.6265082508325577, "sign": 0.6244697785004973, "keep": 0.6675779558718204, "leak": 0.42347750067710876, "replacement": 1.3784014880657196}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 1.2220884561538696, "loss_scale": 4.0, "retries": 0, "seconds": 249.8, "max_memory_gb": 26.28, "pair": 0.6148173958063126, "sign": 0.49297237768769264, "keep": 0.35406431183218956, "leak": 0.6083134917653297, "replacement": 1.1904140710830688}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 2.722086191177368, "loss_scale": 4.0, "retries": 0, "seconds": 251.0, "max_memory_gb": 26.28, "pair": 0.4139615371823311, "sign": 0.19752614200115204, "keep": 0.15237647911999375, "leak": 0.4713117368519306, "replacement": 0.7067407816648483}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 3.9994938373565674, "loss_scale": 4.0, "retries": 0, "seconds": 252.3, "max_memory_gb": 26.28, "pair": 0.692620012909174, "sign": 0.8048259019851685, "keep": 0.8707877155393362, "leak": 0.9816816374659538, "replacement": 1.775847315788269}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 4.650151252746582, "loss_scale": 4.0, "retries": 0, "seconds": 253.7, "max_memory_gb": 26.28, "pair": 0.5789522081613541, "sign": 0.3729273658245802, "keep": 0.3337630648165941, "leak": 2.393197898287326, "replacement": 1.5305968821048737}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 1.9631726741790771, "loss_scale": 4.0, "retries": 0, "seconds": 255.0, "max_memory_gb": 26.28, "pair": 0.21742932684719563, "sign": 0.10382849304005504, "keep": 0.06951993785332888, "leak": 1.0368184745311737, "replacement": 0.5633081719279289}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 2.7904787063598633, "loss_scale": 4.0, "retries": 0, "seconds": 256.3, "max_memory_gb": 26.28, "pair": 0.543378159403801, "sign": 0.5423808395862579, "keep": 0.16322279535233974, "leak": 0.7254170025698841, "replacement": 1.077534258365631}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 2.1775500774383545, "loss_scale": 4.0, "retries": 0, "seconds": 257.6, "max_memory_gb": 26.28, "pair": 0.5629597306251526, "sign": 0.5817769318819046, "keep": 0.23775105737149715, "leak": 0.8189310701563954, "replacement": 1.1774564683437347}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.5166356563568115, "loss_scale": 4.0, "retries": 0, "seconds": 258.9, "max_memory_gb": 26.28, "pair": 0.2627143934369087, "sign": 0.12781028635799885, "keep": 0.2716677349526435, "leak": 2.509871542453766, "replacement": 1.0899212807416916}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 1.6698530912399292, "loss_scale": 4.0, "retries": 0, "seconds": 260.2, "max_memory_gb": 26.28, "pair": 0.5734127201139927, "sign": 0.6435970403254032, "keep": 0.41860457230359316, "leak": 0.6128539610654116, "replacement": 1.2577269971370697}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 2.1556951999664307, "loss_scale": 4.0, "retries": 0, "seconds": 261.5, "max_memory_gb": 26.28, "pair": 0.49772050231695175, "sign": 0.6115021975710988, "keep": 0.2296861121430993, "leak": 0.7277619279921055, "replacement": 1.1002551764249802}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 2.9749181270599365, "loss_scale": 4.0, "retries": 0, "seconds": 262.8, "max_memory_gb": 26.28, "pair": 0.5279745012521744, "sign": 0.3537350073456764, "keep": 0.008467731182463467, "leak": 0.6264950707554817, "replacement": 0.8656996190547943}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 2.6553335189819336, "loss_scale": 4.0, "retries": 0, "seconds": 264.1, "max_memory_gb": 26.28, "pair": 0.5521912425756454, "sign": 0.48671372421085835, "keep": 0.293638052418828, "leak": 1.3427998302795459, "replacement": 1.2780670896172523}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 2.4646856784820557, "loss_scale": 4.0, "retries": 0, "seconds": 265.4, "max_memory_gb": 26.28, "pair": 0.5305869542062283, "sign": 0.6036070808768272, "keep": 0.28955723950639367, "leak": 1.3602087423205376, "replacement": 1.3172212839126587}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 2.8787248134613037, "loss_scale": 4.0, "retries": 0, "seconds": 266.8, "max_memory_gb": 26.28, "pair": 0.3238924443721771, "sign": 0.3267969563603401, "keep": 0.12920119403861463, "leak": 1.8165715262293816, "replacement": 1.0060343742370605}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 2.5759902000427246, "loss_scale": 4.0, "retries": 0, "seconds": 268.1, "max_memory_gb": 26.28, "pair": 0.43994617834687233, "sign": 0.4040711931884289, "keep": 0.15950645372504368, "leak": 2.291957047127653, "replacement": 1.294724240899086}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 5.267763614654541, "loss_scale": 4.0, "retries": 0, "seconds": 269.4, "max_memory_gb": 26.28, "pair": 0.8299584612250328, "sign": 0.9475086778402328, "keep": 1.4675685509573668, "leak": 0.2879057553363964, "replacement": 2.1094735264778137}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 1.8176063299179077, "loss_scale": 4.0, "retries": 0, "seconds": 270.7, "max_memory_gb": 26.28, "pair": 0.28383295238018036, "sign": 0.08393328916281462, "keep": 0.03640171582810581, "leak": 1.471438642591238, "replacement": 0.7118600979447365}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 3.3269007205963135, "loss_scale": 4.0, "retries": 0, "seconds": 272.1, "max_memory_gb": 26.28, "pair": 0.5520353224128485, "sign": 0.5904732178896666, "keep": 0.3481556698679924, "leak": 1.3676508624266717, "replacement": 1.3632624596357346}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 1.530110478401184, "loss_scale": 4.0, "retries": 0, "seconds": 273.4, "max_memory_gb": 26.28, "pair": 0.46333908289670944, "sign": 0.48938165977597237, "keep": 0.49655148247256875, "leak": 1.622350296471268, "replacement": 1.3618932664394379}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 2.5029211044311523, "loss_scale": 4.0, "retries": 0, "seconds": 274.7, "max_memory_gb": 26.28, "pair": 0.6119367387145758, "sign": 0.5692461845465004, "keep": 0.32073291204869747, "leak": 0.6139255302259699, "replacement": 1.210407704114914}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 3.4138431549072266, "loss_scale": 4.0, "retries": 0, "seconds": 276.0, "max_memory_gb": 26.28, "pair": 0.48455386608839035, "sign": 0.6037803292274475, "keep": 0.5142033025622368, "leak": 0.7763868346810341, "replacement": 1.2376424223184586}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 3.0793581008911133, "loss_scale": 4.0, "retries": 0, "seconds": 277.3, "max_memory_gb": 26.28, "pair": 0.39839249290525913, "sign": 0.4586512017995119, "keep": 0.7393652466125786, "leak": 1.7458354458212852, "replacement": 1.4338595643639565}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 1.707430362701416, "loss_scale": 4.0, "retries": 0, "seconds": 278.6, "max_memory_gb": 26.28, "pair": 0.6325684487819672, "sign": 0.6212562099099159, "keep": 0.5400938222883269, "leak": 0.2510046542447526, "replacement": 1.2759946286678314}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 3.4861555099487305, "loss_scale": 4.0, "retries": 0, "seconds": 279.9, "max_memory_gb": 26.28, "pair": 0.8219452500343323, "sign": 0.9654982388019562, "keep": 1.214791803038679, "leak": 0.35477145900949836, "replacement": 2.0007831156253815}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.704588532447815, "loss_scale": 4.0, "retries": 0, "seconds": 281.2, "max_memory_gb": 26.28, "pair": 0.367681335657835, "sign": 0.31331065855920315, "keep": 0.31389080395456403, "leak": 0.5844786465167999, "replacement": 0.8274017125368118}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 4.0561723709106445, "loss_scale": 4.0, "retries": 0, "seconds": 282.5, "max_memory_gb": 26.28, "pair": 0.2911657691001892, "sign": 0.2360944813117385, "keep": 0.06748727196827531, "leak": 1.6361575424671173, "replacement": 0.8519960343837738}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 2.4149832725524902, "loss_scale": 4.0, "retries": 0, "seconds": 283.8, "max_memory_gb": 26.28, "pair": 0.855005607008934, "sign": 1.0447376668453217, "keep": 0.3411389000248164, "leak": 0.16690331732388586, "replacement": 1.5896696895360947}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 3.4259231090545654, "loss_scale": 4.0, "retries": 0, "seconds": 285.1, "max_memory_gb": 26.28, "pair": 0.5006448924541473, "sign": 0.5646352879703045, "keep": 0.19957893388345838, "leak": 0.4891172908246517, "replacement": 1.005031332373619}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 2.5250368118286133, "loss_scale": 4.0, "retries": 0, "seconds": 286.4, "max_memory_gb": 26.28, "pair": 0.44055982679128647, "sign": 0.48982299864292145, "keep": 0.2521519451402128, "leak": 0.6717947870492935, "replacement": 0.9794959872961044}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 2.2780954837799072, "loss_scale": 4.0, "retries": 0, "seconds": 287.7, "max_memory_gb": 26.28, "pair": 0.35701433196663857, "sign": 0.2679335717111826, "keep": 0.19524545595049858, "leak": 0.8215594225330278, "replacement": 0.7939936928451061}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 5.401108264923096, "loss_scale": 4.0, "retries": 0, "seconds": 289.0, "max_memory_gb": 26.28, "pair": 0.43159787356853485, "sign": 0.238052386790514, "keep": 0.6146839912980795, "leak": 1.3897224286338314, "replacement": 1.2053967118263245}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 3.4924283027648926, "loss_scale": 4.0, "retries": 0, "seconds": 290.3, "max_memory_gb": 26.28, "pair": 0.6040220484137535, "sign": 0.6601313725113869, "keep": 0.32407660013996065, "leak": 0.16138164792209864, "replacement": 1.136471450328827}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 2.7320291996002197, "loss_scale": 4.0, "retries": 0, "seconds": 291.6, "max_memory_gb": 26.28, "pair": 0.37736535631120205, "sign": 0.3927594441920519, "keep": 0.13117947056889534, "leak": 1.4461120255291462, "replacement": 1.0008628517389297}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 3.1902763843536377, "loss_scale": 4.0, "retries": 0, "seconds": 292.9, "max_memory_gb": 26.28, "pair": 0.5193029511719942, "sign": 0.554868146777153, "keep": 0.12576173804700375, "leak": 1.0125394249334931, "replacement": 1.112752690911293}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 3.628389596939087, "loss_scale": 4.0, "retries": 0, "seconds": 294.2, "max_memory_gb": 26.28, "pair": 0.33346647769212723, "sign": 0.30723824352025986, "keep": 0.25887094950303435, "leak": 0.8920789062976837, "replacement": 0.8395408242940903}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 1.7159699201583862, "loss_scale": 4.0, "retries": 0, "seconds": 295.5, "max_memory_gb": 26.28, "pair": 0.4572897031903267, "sign": 0.2111234748736024, "keep": 0.03433330962434411, "leak": 0.8591374196112156, "replacement": 0.7948024421930313}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 2.4229328632354736, "loss_scale": 4.0, "retries": 0, "seconds": 296.8, "max_memory_gb": 26.28, "pair": 0.3519870862364769, "sign": 0.39127081632614136, "keep": 0.0481715789064765, "leak": 0.7072700187563896, "replacement": 0.7485257983207703}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 3.0299015045166016, "loss_scale": 4.0, "retries": 0, "seconds": 298.1, "max_memory_gb": 26.28, "pair": 0.5613609328866005, "sign": 0.6666292427107692, "keep": 0.45890933042392135, "leak": 0.6565493550151587, "replacement": 1.288267582654953}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 3.195633888244629, "loss_scale": 4.0, "retries": 0, "seconds": 299.3, "max_memory_gb": 26.28, "pair": 0.6798281669616699, "sign": 0.7321294248104095, "keep": 0.5878389184363186, "leak": 0.39057897962629795, "replacement": 1.4374571293592453}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 3.253105640411377, "loss_scale": 4.0, "retries": 0, "seconds": 300.6, "max_memory_gb": 26.28, "pair": 0.7011895254254341, "sign": 0.7946744188666344, "keep": 0.5681758628925309, "leak": 0.2638660594820976, "replacement": 1.4485811591148376}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 3.547053337097168, "loss_scale": 4.0, "retries": 0, "seconds": 301.9, "max_memory_gb": 26.28, "pair": 0.30029259249567986, "sign": 0.33918867632746696, "keep": 0.05376039072871208, "leak": 0.5634903088212013, "replacement": 0.6376397050917149}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 2.835118055343628, "loss_scale": 4.0, "retries": 0, "seconds": 303.2, "max_memory_gb": 26.28, "pair": 0.34195293486118317, "sign": 0.17687859013676643, "keep": 0.27992954291403294, "leak": 0.931414045393467, "replacement": 0.8032105416059494}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 3.1070756912231445, "loss_scale": 4.0, "retries": 0, "seconds": 304.5, "max_memory_gb": 26.28, "pair": 0.43295231461524963, "sign": 0.4533762400969863, "keep": 0.2575557384407148, "leak": 1.5512196253985167, "replacement": 1.1762232184410095}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 3.715174436569214, "loss_scale": 4.0, "retries": 0, "seconds": 305.8, "max_memory_gb": 26.28, "pair": 0.5960096195340157, "sign": 0.6592043116688728, "keep": 0.7287254882976413, "leak": 0.5823767059482634, "replacement": 1.4355687201023102}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 2.9422874450683594, "loss_scale": 4.0, "retries": 0, "seconds": 307.0, "max_memory_gb": 26.28, "pair": 0.5811757594347, "sign": 0.4928160086274147, "keep": 0.18683529482223094, "leak": 0.2792224729200825, "replacement": 0.9908070266246796}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.50229549407959, "loss_scale": 4.0, "retries": 0, "seconds": 308.3, "max_memory_gb": 26.28, "pair": 0.501407116651535, "sign": 0.45249943248927593, "keep": 0.3965206858702004, "leak": 0.630089147215898, "replacement": 1.0834394097328186}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 2.8531875610351562, "loss_scale": 4.0, "retries": 0, "seconds": 309.6, "max_memory_gb": 26.28, "pair": 0.2608906924724579, "sign": 0.1551429806277156, "keep": 0.09072774648666382, "leak": 1.4075000584125519, "replacement": 0.7357010468840599}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 4.335728168487549, "loss_scale": 4.0, "retries": 0, "seconds": 310.9, "max_memory_gb": 26.28, "pair": 0.458274032920599, "sign": 0.5524041373282671, "keep": 0.32562970370054245, "leak": 3.5463051917031407, "replacement": 1.7838672399520874}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.0188655853271484, "loss_scale": 4.0, "retries": 0, "seconds": 312.2, "max_memory_gb": 26.28, "pair": 0.3505169562995434, "sign": 0.3023643009364605, "keep": 0.07247178419493139, "leak": 0.7041968693956733, "replacement": 0.7139842212200165}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 3.040816068649292, "loss_scale": 4.0, "retries": 0, "seconds": 313.5, "max_memory_gb": 26.28, "pair": 0.5473600551486015, "sign": 0.6370020061731339, "keep": 0.19214480184018612, "leak": 0.3642278863117099, "replacement": 1.0529904514551163}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 1.7826783657073975, "loss_scale": 4.0, "retries": 0, "seconds": 314.8, "max_memory_gb": 26.28, "pair": 0.5047830566763878, "sign": 0.5838065100833774, "keep": 0.23743533610831946, "leak": 0.5145317614078522, "replacement": 1.0440369099378586}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 2.76865553855896, "loss_scale": 4.0, "retries": 0, "seconds": 316.1, "max_memory_gb": 26.28, "pair": 0.28197595849633217, "sign": 0.2262813337147236, "keep": 0.05887055245693773, "leak": 2.9902291633188725, "replacement": 1.172109179198742}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 6.625775337219238, "loss_scale": 4.0, "retries": 0, "seconds": 317.5, "max_memory_gb": 26.28, "pair": 0.7282654009759426, "sign": 0.8963673263788223, "keep": 0.9116145335137844, "leak": 0.7712837979197502, "replacement": 1.8250773176550865}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 2.630620002746582, "loss_scale": 4.0, "retries": 0, "seconds": 318.7, "max_memory_gb": 26.28, "pair": 0.8781705796718597, "sign": 1.0188577249646187, "keep": 0.7199827367439866, "leak": 0.45384201407432556, "replacement": 1.8610513508319855}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 3.3182950019836426, "loss_scale": 4.0, "retries": 0, "seconds": 320.0, "max_memory_gb": 26.28, "pair": 0.20560236182063818, "sign": 0.19146537873893976, "keep": 0.05458603904116899, "leak": 1.477620616555214, "replacement": 0.6980332285165787}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 2.938232421875, "loss_scale": 4.0, "retries": 0, "seconds": 321.3, "max_memory_gb": 26.28, "pair": 0.29777538031339645, "sign": 0.30327222123742104, "keep": 0.1753926078963559, "leak": 0.5987439565360546, "replacement": 0.6867937743663788}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.6847522258758545, "loss_scale": 4.0, "retries": 0, "seconds": 322.6, "max_memory_gb": 26.28, "pair": 0.21996014937758446, "sign": 0.1281699505634606, "keep": 0.14450380951166153, "leak": 0.6090436017839238, "replacement": 0.5085579231381416}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 5.540674209594727, "loss_scale": 4.0, "retries": 0, "seconds": 323.9, "max_memory_gb": 26.28, "pair": 0.5269577950239182, "sign": 0.6807775050401688, "keep": 0.11100432835519314, "leak": 0.7997604641132057, "replacement": 1.1227888464927673}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 3.4566123485565186, "loss_scale": 4.0, "retries": 0, "seconds": 325.2, "max_memory_gb": 26.28, "pair": 0.256425853818655, "sign": 0.22401565499603748, "keep": 0.0653341039433144, "leak": 1.3389102593064308, "replacement": 0.7358282953500748}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 3.9885170459747314, "loss_scale": 4.0, "retries": 0, "seconds": 326.5, "max_memory_gb": 26.28, "pair": 0.42435572296380997, "sign": 0.44288243539631367, "keep": 0.1824988427106291, "leak": 2.01975851503812, "replacement": 1.2419859915971756}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 3.6083016395568848, "loss_scale": 4.0, "retries": 0, "seconds": 327.8, "max_memory_gb": 26.28, "pair": 0.2714550271630287, "sign": 0.15991753386333585, "keep": 0.027770008193328977, "leak": 1.2436531037092209, "replacement": 0.6762120723724365}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 6.351988315582275, "loss_scale": 4.0, "retries": 0, "seconds": 329.1, "max_memory_gb": 26.28, "pair": 0.2758710850030184, "sign": 0.3107905751094222, "keep": 0.20039415446808562, "leak": 1.955227730795741, "replacement": 1.020270362496376}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 4.852778911590576, "loss_scale": 4.0, "retries": 0, "seconds": 330.4, "max_memory_gb": 26.28, "pair": 0.22959017381072044, "sign": 0.3064318150281906, "keep": 0.33287687599658966, "leak": 1.762931264936924, "replacement": 0.9899773299694061}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 2.898458957672119, "loss_scale": 4.0, "retries": 0, "seconds": 331.7, "max_memory_gb": 26.28, "pair": 0.5236681662499905, "sign": 0.6784739308059216, "keep": 0.14106350485235453, "leak": 0.5844939965754747, "replacement": 1.0795603841543198}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 3.6545979976654053, "loss_scale": 4.0, "retries": 0, "seconds": 333.0, "max_memory_gb": 26.28, "pair": 0.454457214102149, "sign": 0.5447525335475802, "keep": 0.11315734451636672, "leak": 1.2783337868750095, "replacement": 1.102995604276657}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 5.0201096534729, "loss_scale": 4.0, "retries": 0, "seconds": 334.2, "max_memory_gb": 26.28, "pair": 0.4558066166937351, "sign": 0.49397224839776754, "keep": 0.13112212606938556, "leak": 0.966928927693516, "replacement": 1.0100860595703125}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 2.6817805767059326, "loss_scale": 4.0, "retries": 0, "seconds": 335.6, "max_memory_gb": 26.28, "pair": 0.19221749342978, "sign": 0.1464646328240633, "keep": 0.23419275658670813, "leak": 1.4581146240234375, "replacement": 0.7470748275518417}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.571699380874634, "loss_scale": 4.0, "retries": 0, "seconds": 336.9, "max_memory_gb": 26.28, "pair": 0.25607383996248245, "sign": 0.22218695655465126, "keep": 0.009205742971971631, "leak": 1.1065887212753296, "replacement": 0.6484173685312271}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 4.228734493255615, "loss_scale": 4.0, "retries": 0, "seconds": 338.1, "max_memory_gb": 26.28, "pair": 0.2931514382362366, "sign": 0.3574081901460886, "keep": 0.22071160236373544, "leak": 1.0676676332950592, "replacement": 0.8491282314062119}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.5845236778259277, "loss_scale": 4.0, "retries": 0, "seconds": 339.4, "max_memory_gb": 26.28, "pair": 0.3906309138983488, "sign": 0.3324133139103651, "keep": 0.27021792344748974, "leak": 0.8851058788131922, "replacement": 0.9132230207324028}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 4.641022682189941, "loss_scale": 4.0, "retries": 0, "seconds": 340.7, "max_memory_gb": 26.28, "pair": 0.5030317157506943, "sign": 0.5814527105540037, "keep": 0.5353404083289206, "leak": 0.5409387722611427, "replacement": 1.1966629475355148}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 4.464531421661377, "loss_scale": 4.0, "retries": 0, "seconds": 342.0, "max_memory_gb": 26.28, "pair": 0.3147229142487049, "sign": 0.3366196872666478, "keep": 0.308342631906271, "leak": 1.8643275164067745, "replacement": 1.103285938501358}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 6.334976673126221, "loss_scale": 4.0, "retries": 0, "seconds": 343.3, "max_memory_gb": 26.28, "pair": 0.2922547645866871, "sign": 0.14283425733447075, "keep": 0.6782089122571051, "leak": 2.8015555404126644, "replacement": 1.4031652063131332}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 3.761213779449463, "loss_scale": 4.0, "retries": 0, "seconds": 344.6, "max_memory_gb": 26.28, "pair": 0.3815375529229641, "sign": 0.3341216892004013, "keep": 0.12112351797986776, "leak": 1.2136289223562926, "replacement": 0.9125673770904541}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 6.631618976593018, "loss_scale": 4.0, "retries": 0, "seconds": 345.9, "max_memory_gb": 26.28, "pair": 0.5940005257725716, "sign": 0.5849457429721951, "keep": 1.0307388193905354, "leak": 0.3940505236387253, "replacement": 1.5003554672002792}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.6831419467926025, "loss_scale": 4.0, "retries": 0, "seconds": 347.2, "max_memory_gb": 26.28, "pair": 0.46285879611968994, "sign": 0.4633999727666378, "keep": 0.3277564737945795, "leak": 0.6654530018568039, "replacement": 1.0248002484440804}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 2.654627561569214, "loss_scale": 4.0, "retries": 0, "seconds": 348.5, "max_memory_gb": 26.28, "pair": 0.5835181511938572, "sign": 0.7200837582349777, "keep": 0.3913499814225361, "leak": 0.4439960829913616, "replacement": 1.2502340748906136}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 5.5406012535095215, "loss_scale": 4.0, "retries": 0, "seconds": 349.8, "max_memory_gb": 26.28, "pair": 0.3306147828698158, "sign": 0.42846381943672895, "keep": 1.281212495174259, "leak": 0.7899216488003731, "replacement": 1.3829333186149597}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 4.642188549041748, "loss_scale": 4.0, "retries": 0, "seconds": 351.1, "max_memory_gb": 26.28, "pair": 0.36011921986937523, "sign": 0.4198454078286886, "keep": 0.4790792092680931, "leak": 0.9070713333785534, "replacement": 1.0363493971526623}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 4.447934627532959, "loss_scale": 4.0, "retries": 0, "seconds": 352.4, "max_memory_gb": 26.28, "pair": 0.21884387359023094, "sign": 0.2552651911973953, "keep": 0.08299261704087257, "leak": 2.577467016875744, "replacement": 1.03233952075243}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 3.334745407104492, "loss_scale": 4.0, "retries": 0, "seconds": 353.7, "max_memory_gb": 26.28, "pair": 0.14310252107679844, "sign": 0.052151011768728495, "keep": 0.05976849142462015, "leak": 0.7381822299212217, "replacement": 0.38360781967639923}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 5.0893964767456055, "loss_scale": 4.0, "retries": 0, "seconds": 355.0, "max_memory_gb": 26.28, "pair": 0.37949417531490326, "sign": 0.18440093472599983, "keep": 0.06580981344450265, "leak": 0.19115994500134548, "replacement": 0.552389532327652}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 7.130953788757324, "loss_scale": 4.0, "retries": 0, "seconds": 356.3, "max_memory_gb": 26.28, "pair": 0.5941399186849594, "sign": 0.8206704258918762, "keep": 0.165408443659544, "leak": 0.3199012152526848, "replacement": 1.167154684662819}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 2.7527010440826416, "loss_scale": 4.0, "retries": 0, "seconds": 357.6, "max_memory_gb": 26.28, "pair": 0.4204840511083603, "sign": 0.4808829156681895, "keep": 1.3270059563219547, "leak": 0.30319352832157165, "replacement": 1.400226902216673}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 2.8269803524017334, "loss_scale": 4.0, "retries": 0, "seconds": 358.9, "max_memory_gb": 26.28, "pair": 0.183952571824193, "sign": 0.10979661205783486, "keep": 0.07517359149642289, "leak": 0.5464939088560641, "replacement": 0.41306115314364433}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 4.2006306648254395, "loss_scale": 4.0, "retries": 0, "seconds": 360.2, "max_memory_gb": 26.28, "pair": 0.31978655979037285, "sign": 0.2921579144895077, "keep": 0.044394032564014196, "leak": 0.6750066173262894, "replacement": 0.6568141728639603}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 6.837794303894043, "loss_scale": 4.0, "retries": 0, "seconds": 361.5, "max_memory_gb": 26.28, "pair": 0.45833832025527954, "sign": 0.5607704967260361, "keep": 0.5093005397357047, "leak": 0.3893183034379035, "replacement": 1.0907034054398537}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 4.258673191070557, "loss_scale": 4.0, "retries": 0, "seconds": 362.8, "max_memory_gb": 26.28, "pair": 0.4889962822198868, "sign": 0.2959722988307476, "keep": 0.16171306907199323, "leak": 1.1470288001000881, "replacement": 1.0045962035655975}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 5.025125026702881, "loss_scale": 4.0, "retries": 0, "seconds": 364.1, "max_memory_gb": 26.28, "pair": 0.2856047563254833, "sign": 0.14771402534097433, "keep": 0.0513226673938334, "leak": 0.6376534812152386, "replacement": 0.5445364713668823}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 3.1032354831695557, "loss_scale": 4.0, "retries": 0, "seconds": 365.3, "max_memory_gb": 26.28, "pair": 0.2826486174017191, "sign": 0.27950573340058327, "keep": 0.32888458319939673, "leak": 0.6592217807192355, "replacement": 0.7516491934657097}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.1652779579162598, "loss_scale": 4.0, "retries": 0, "seconds": 366.6, "max_memory_gb": 26.28, "pair": 0.28047108836472034, "sign": 0.29384233616292477, "keep": 0.12976423546206206, "leak": 0.367221686989069, "replacement": 0.5840797908604145}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 4.886368274688721, "loss_scale": 4.0, "retries": 0, "seconds": 368.0, "max_memory_gb": 26.28, "pair": 0.18791769072413445, "sign": 0.08873564936220646, "keep": 0.15428641438484192, "leak": 3.0554932057857513, "replacement": 1.0733020305633545}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 4.552684307098389, "loss_scale": 4.0, "retries": 0, "seconds": 369.3, "max_memory_gb": 26.28, "pair": 0.2336425967514515, "sign": 0.16220812685787678, "keep": 0.09834805480204523, "leak": 0.968125231564045, "replacement": 0.6059519797563553}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 6.428516387939453, "loss_scale": 4.0, "retries": 0, "seconds": 370.6, "max_memory_gb": 26.28, "pair": 0.3680577799677849, "sign": 0.523755956441164, "keep": 0.16394046461209655, "leak": 2.283173888707097, "replacement": 1.2826994508504868}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 6.287977695465088, "loss_scale": 4.0, "retries": 0, "seconds": 372.0, "max_memory_gb": 26.28, "pair": 0.3709326982498169, "sign": 0.5460733883082867, "keep": 0.047466100309975445, "leak": 2.363657236099243, "replacement": 1.2586167752742767}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 8.423908233642578, "loss_scale": 4.0, "retries": 0, "seconds": 373.3, "max_memory_gb": 26.28, "pair": 0.4102085269987583, "sign": 0.4453002968803048, "keep": 0.5318315215408802, "leak": 0.46780499536544085, "replacement": 1.0157257169485092}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 5.928073406219482, "loss_scale": 4.0, "retries": 0, "seconds": 374.6, "max_memory_gb": 26.28, "pair": 0.3142262641340494, "sign": 0.3053482919931412, "keep": 0.0694177383556962, "leak": 0.6079764738678932, "replacement": 0.6536034010350704}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 4.153594017028809, "loss_scale": 4.0, "retries": 0, "seconds": 375.9, "max_memory_gb": 26.28, "pair": 0.29149847850203514, "sign": 0.20166299678385258, "keep": 0.08826383762061596, "leak": 1.2419511452317238, "replacement": 0.7469496726989746}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 6.059169769287109, "loss_scale": 4.0, "retries": 0, "seconds": 377.3, "max_memory_gb": 26.28, "pair": 0.34602193906903267, "sign": 0.43580102850683033, "keep": 0.3078574842074886, "leak": 1.0563404224812984, "replacement": 0.9819363057613373}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 6.063729763031006, "loss_scale": 4.0, "retries": 0, "seconds": 378.6, "max_memory_gb": 26.28, "pair": 0.10066110640764236, "sign": 0.051363639533519745, "keep": 0.02863371721468866, "leak": 2.3160077929496765, "replacement": 0.719661720097065}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 3.479158401489258, "loss_scale": 4.0, "retries": 0, "seconds": 379.9, "max_memory_gb": 26.28, "pair": 0.18157697282731533, "sign": 0.1617343369871378, "keep": 0.043348005041480064, "leak": 0.8195027858018875, "replacement": 0.4889938533306122}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 3.750551462173462, "loss_scale": 4.0, "retries": 0, "seconds": 381.2, "max_memory_gb": 26.28, "pair": 0.31227007415145636, "sign": 0.2972481166943908, "keep": 0.07343234424479306, "leak": 0.6684896796941757, "replacement": 0.6647327467799187}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 4.5527663230896, "loss_scale": 4.0, "retries": 0, "seconds": 382.5, "max_memory_gb": 26.28, "pair": 0.20924624800682068, "sign": 0.1524280495941639, "keep": 0.24651631526648998, "leak": 1.1050937343388796, "replacement": 0.6849918887019157}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 6.988991737365723, "loss_scale": 4.0, "retries": 0, "seconds": 383.7, "max_memory_gb": 26.28, "pair": 0.25399003364145756, "sign": 0.3834196627140045, "keep": 0.3776492935139686, "leak": 2.5681968070566654, "replacement": 1.2765737175941467}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 7.222944736480713, "loss_scale": 4.0, "retries": 0, "seconds": 385.0, "max_memory_gb": 26.28, "pair": 0.4730225447565317, "sign": 0.70333231985569, "keep": 0.638113115914166, "leak": 1.401159230619669, "replacement": 1.4940351098775864}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 5.341195106506348, "loss_scale": 4.0, "retries": 0, "seconds": 386.3, "max_memory_gb": 26.28, "pair": 0.4839217010885477, "sign": 0.7378065399825573, "keep": 0.4499297284637578, "leak": 0.7782530300319195, "replacement": 1.2723531052470207}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 3.8834176063537598, "loss_scale": 4.0, "retries": 0, "seconds": 387.6, "max_memory_gb": 26.28, "pair": 0.5403875485062599, "sign": 0.6675370596349239, "keep": 0.7294961703009903, "leak": 0.4121560679268441, "replacement": 1.3419431671500206}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 6.294761657714844, "loss_scale": 4.0, "retries": 0, "seconds": 389.0, "max_memory_gb": 26.28, "pair": 0.3329409398138523, "sign": 0.12424168735742569, "keep": 0.06505957245826721, "leak": 1.711918868124485, "replacement": 0.8555712997913361}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 7.824833869934082, "loss_scale": 4.0, "retries": 0, "seconds": 390.3, "max_memory_gb": 26.28, "pair": 0.6375373229384422, "sign": 0.8403306379914284, "keep": 0.5258131623268127, "leak": 1.9705609012594323, "replacement": 1.8132494390010834}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 4.1005377769470215, "loss_scale": 4.0, "retries": 0, "seconds": 391.7, "max_memory_gb": 26.28, "pair": 0.25790590792894363, "sign": 0.33168646367266774, "keep": 0.09005471970885992, "leak": 0.750338105019182, "replacement": 0.6563610136508942}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 3.4515910148620605, "loss_scale": 4.0, "retries": 0, "seconds": 393.0, "max_memory_gb": 26.28, "pair": 0.16987064108252525, "sign": 0.05825222237035632, "keep": 0.12413279002066702, "leak": 0.7175537720322609, "replacement": 0.44045158475637436}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 7.270634651184082, "loss_scale": 4.0, "retries": 0, "seconds": 394.4, "max_memory_gb": 26.28, "pair": 0.3938830681145191, "sign": 0.28054936323314905, "keep": 0.1041885064332746, "leak": 1.863066228106618, "replacement": 1.05201855301857}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 8.620565414428711, "loss_scale": 4.0, "retries": 0, "seconds": 395.8, "max_memory_gb": 26.28, "pair": 0.4922259971499443, "sign": 0.4278589952737093, "keep": 0.06881369650363922, "leak": 1.6355322748422623, "replacement": 1.1494453996419907}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 8.435729026794434, "loss_scale": 4.0, "retries": 0, "seconds": 397.1, "max_memory_gb": 26.28, "pair": 0.34438611567020416, "sign": 0.197639518417418, "keep": 0.09457629173994064, "leak": 1.5275719759811182, "replacement": 0.8723869994282722}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 5.970269203186035, "loss_scale": 4.0, "retries": 0, "seconds": 398.4, "max_memory_gb": 26.28, "pair": 0.2736920602619648, "sign": 0.05299515277147293, "keep": 0.09822061657905579, "leak": 0.8459226228296757, "replacement": 0.5607805997133255}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 9.029330253601074, "loss_scale": 4.0, "retries": 0, "seconds": 399.7, "max_memory_gb": 26.28, "pair": 0.44184470921754837, "sign": 0.4557946715503931, "keep": 0.5307384869083762, "leak": 1.022283460944891, "replacement": 1.1906821690499783}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 6.133781909942627, "loss_scale": 4.0, "retries": 0, "seconds": 401.0, "max_memory_gb": 26.28, "pair": 0.2755205426365137, "sign": 0.4145143125206232, "keep": 0.11446087784133852, "leak": 1.2308088690042496, "replacement": 0.8477103486657143}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 11.845840454101562, "loss_scale": 4.0, "retries": 0, "seconds": 402.3, "max_memory_gb": 26.28, "pair": 0.5903499983251095, "sign": 0.693325201747939, "keep": 0.42814183607697487, "leak": 0.6846907488070428, "replacement": 1.3222562074661255}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 3.5749547481536865, "loss_scale": 4.0, "retries": 0, "seconds": 403.6, "max_memory_gb": 26.28, "pair": 0.34628959745168686, "sign": 0.12745163030922413, "keep": 0.05794799968134612, "leak": 1.1078679785132408, "replacement": 0.715956412255764}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 3.7535812854766846, "loss_scale": 4.0, "retries": 0, "seconds": 404.8, "max_memory_gb": 26.28, "pair": 0.3560906518250704, "sign": 0.41910291463136673, "keep": 0.2508232198888436, "leak": 0.5132353752851486, "replacement": 0.8193625435233116}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 7.324298858642578, "loss_scale": 4.0, "retries": 0, "seconds": 406.1, "max_memory_gb": 26.28, "pair": 0.3241329453885555, "sign": 0.1925880666822195, "keep": 0.0586615955689922, "leak": 0.30750842671841383, "replacement": 0.5266348868608475}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 11.052878379821777, "loss_scale": 4.0, "retries": 0, "seconds": 407.4, "max_memory_gb": 26.28, "pair": 0.48894689977169037, "sign": 0.5900615574792027, "keep": 0.6936493162065744, "leak": 0.6852208822965622, "replacement": 1.302107535302639}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 9.85405445098877, "loss_scale": 4.0, "retries": 0, "seconds": 408.8, "max_memory_gb": 26.28, "pair": 0.4433561973273754, "sign": 0.43646751530468464, "keep": 0.5257077509886585, "leak": 1.125782459974289, "replacement": 1.2058894038200378}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 7.754857063293457, "loss_scale": 4.0, "retries": 0, "seconds": 410.1, "max_memory_gb": 26.28, "pair": 0.24756036885082722, "sign": 0.22019612276926637, "keep": 0.08234359510242939, "leak": 1.2787895891815424, "replacement": 0.7185276001691818}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 4.404874801635742, "loss_scale": 4.0, "retries": 0, "seconds": 411.4, "max_memory_gb": 26.28, "pair": 0.649418368935585, "sign": 0.7900401661172509, "keep": 0.3948552729561925, "leak": 0.5805733698070981, "replacement": 1.3870094269514084}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 5.697738170623779, "loss_scale": 4.0, "retries": 0, "seconds": 412.7, "max_memory_gb": 26.28, "pair": 0.31678799353539944, "sign": 0.383983813226223, "keep": 0.22288341633975506, "leak": 1.406183511018753, "replacement": 0.9717675000429153}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 4.496990203857422, "loss_scale": 4.0, "retries": 0, "seconds": 413.9, "max_memory_gb": 26.28, "pair": 0.3342009913176298, "sign": 0.24897808767855167, "keep": 0.10697164433076978, "leak": 0.7802872024476528, "replacement": 0.7072476595640182}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 5.141599655151367, "loss_scale": 4.0, "retries": 0, "seconds": 415.2, "max_memory_gb": 26.28, "pair": 0.3981301262974739, "sign": 0.3501129765063524, "keep": 0.17252502404153347, "leak": 0.32742578350007534, "replacement": 0.7413055785000324}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 5.40775728225708, "loss_scale": 4.0, "retries": 0, "seconds": 416.5, "max_memory_gb": 26.28, "pair": 0.40070217102766037, "sign": 0.42020659800618887, "keep": 0.3835011908086017, "leak": 0.5051716603338718, "replacement": 0.9288489893078804}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 4.917027950286865, "loss_scale": 4.0, "retries": 0, "seconds": 417.8, "max_memory_gb": 26.28, "pair": 0.3258587494492531, "sign": 0.21978645026683807, "keep": 0.050860548159107566, "leak": 0.7397473265336885, "replacement": 0.6461190730333328}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 8.435751914978027, "loss_scale": 4.0, "retries": 0, "seconds": 419.1, "max_memory_gb": 26.28, "pair": 0.5463436581194401, "sign": 0.44992876052856445, "keep": 0.0709080696105957, "leak": 0.1761895983681825, "replacement": 0.8508094549179077}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 10.024483680725098, "loss_scale": 4.0, "retries": 0, "seconds": 420.4, "max_memory_gb": 26.28, "pair": 0.10765334032475948, "sign": 0.09680389892309904, "keep": 0.05542846443131566, "leak": 3.728690952062607, "replacement": 1.1159422919154167}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 4.266790390014648, "loss_scale": 4.0, "retries": 0, "seconds": 421.7, "max_memory_gb": 26.28, "pair": 0.1549359792843461, "sign": 0.07671087328344584, "keep": 0.07545788818970323, "leak": 0.8935454017482698, "replacement": 0.4544067233800888}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 8.586048126220703, "loss_scale": 4.0, "retries": 0, "seconds": 423.0, "max_memory_gb": 26.28, "pair": 0.7330804765224457, "sign": 0.6955029498785734, "keep": 0.6124046505428851, "leak": 0.10256290505640209, "replacement": 1.4126749783754349}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 7.224076747894287, "loss_scale": 4.0, "retries": 0, "seconds": 424.3, "max_memory_gb": 26.28, "pair": 0.4601197615265846, "sign": 0.5111595876514912, "keep": 0.3088050263468176, "leak": 0.8986105628428049, "replacement": 1.0947546660900116}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 11.436760902404785, "loss_scale": 4.0, "retries": 0, "seconds": 425.6, "max_memory_gb": 26.28, "pair": 0.2522417027503252, "sign": 0.1368772303685546, "keep": 0.5439972272142768, "leak": 0.6143020170275122, "replacement": 0.746254451572895}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 6.614604473114014, "loss_scale": 4.0, "retries": 0, "seconds": 426.9, "max_memory_gb": 26.28, "pair": 0.5326447188854218, "sign": 0.6446960866451263, "keep": 0.16754601104184985, "leak": 0.20963301043957472, "replacement": 0.9911739677190781}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 7.296014785766602, "loss_scale": 4.0, "retries": 0, "seconds": 428.2, "max_memory_gb": 26.28, "pair": 0.298422968480736, "sign": 0.4164656684733927, "keep": 0.37614803900942206, "leak": 1.320306308567524, "replacement": 1.0248064324259758}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 5.3470234870910645, "loss_scale": 4.0, "retries": 0, "seconds": 429.5, "max_memory_gb": 26.28, "pair": 0.5237623583525419, "sign": 0.5458583259023726, "keep": 0.30843655799981207, "leak": 0.545802986198396, "replacement": 1.0873605906963348}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 7.640434265136719, "loss_scale": 4.0, "retries": 0, "seconds": 430.8, "max_memory_gb": 26.28, "pair": 0.5940733179450035, "sign": 0.6397431567311287, "keep": 0.6270705605857074, "leak": 1.6483596004545689, "replacement": 1.6395700424909592}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 6.275918960571289, "loss_scale": 4.0, "retries": 0, "seconds": 432.2, "max_memory_gb": 26.28, "pair": 0.3557666018605232, "sign": 0.3434256576001644, "keep": 0.29838654852937907, "leak": 0.7177691720426083, "replacement": 0.856114998459816}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 5.48766565322876, "loss_scale": 4.0, "retries": 0, "seconds": 433.4, "max_memory_gb": 26.28, "pair": 0.6230908390134573, "sign": 0.7865026332437992, "keep": 0.6492396332323551, "leak": 0.5291291655739769, "replacement": 1.4732443168759346}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 6.343233585357666, "loss_scale": 4.0, "retries": 0, "seconds": 434.8, "max_memory_gb": 26.28, "pair": 0.43678785488009453, "sign": 0.5097968522459269, "keep": 0.4377123531885445, "leak": 1.890641191144823, "replacement": 1.383202813565731}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 4.632945537567139, "loss_scale": 4.0, "retries": 0, "seconds": 436.1, "max_memory_gb": 26.28, "pair": 0.5063445419073105, "sign": 0.5694450959563255, "keep": 0.6976488209911622, "leak": 0.31322601553983986, "replacement": 1.2181979790329933}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 7.223863124847412, "loss_scale": 4.0, "retries": 0, "seconds": 437.4, "max_memory_gb": 26.28, "pair": 0.22753887251019478, "sign": 0.12985021620988846, "keep": 0.05050769814988598, "leak": 1.0253879874944687, "replacement": 0.5740648210048676}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 16.571414947509766, "loss_scale": 4.0, "retries": 0, "seconds": 438.7, "max_memory_gb": 26.28, "pair": 0.6466581076383591, "sign": 0.7269439119845629, "keep": 0.4151166118681431, "leak": 1.269940851489082, "replacement": 1.5351735949516296}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 6.3000264167785645, "loss_scale": 4.0, "retries": 0, "seconds": 440.1, "max_memory_gb": 26.28, "pair": 0.4023016858845949, "sign": 0.5200504628010094, "keep": 0.42202258796896785, "leak": 0.8731347043067217, "replacement": 1.0916219055652618}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 6.325904846191406, "loss_scale": 4.0, "retries": 0, "seconds": 441.4, "max_memory_gb": 26.28, "pair": 0.4345661625266075, "sign": 0.7045548195019364, "keep": 0.2292402705643326, "leak": 1.7965243770740926, "replacement": 1.3505948185920715}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 5.80228328704834, "loss_scale": 4.0, "retries": 0, "seconds": 442.7, "max_memory_gb": 26.28, "pair": 0.3482934460043907, "sign": 0.3727236483246088, "keep": 0.7469192181015387, "leak": 0.5699638817459345, "replacement": 1.0506058521568775}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 6.868765830993652, "loss_scale": 4.0, "retries": 0, "seconds": 444.0, "max_memory_gb": 26.28, "pair": 0.25703986920416355, "sign": 0.29560118122026324, "keep": 0.2979776702122763, "leak": 1.6964797116816044, "replacement": 0.9779492318630219}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 6.938694477081299, "loss_scale": 4.0, "retries": 0, "seconds": 445.3, "max_memory_gb": 26.28, "pair": 0.6382362879812717, "sign": 0.7101507876068354, "keep": 0.6348501825705171, "leak": 0.8546710666269064, "replacement": 1.5244045406579971}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 4.749020099639893, "loss_scale": 4.0, "retries": 0, "seconds": 446.6, "max_memory_gb": 26.28, "pair": 0.45541493222117424, "sign": 0.5837020892649889, "keep": 0.02474020980298519, "leak": 0.25874463003128767, "replacement": 0.8243222609162331}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 10.618902206420898, "loss_scale": 4.0, "retries": 0, "seconds": 447.8, "max_memory_gb": 26.28, "pair": 0.243335728533566, "sign": 0.2134344894438982, "keep": 0.05782548262504861, "leak": 1.7428544908761978, "replacement": 0.814679354429245}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 5.884861946105957, "loss_scale": 4.0, "retries": 0, "seconds": 449.1, "max_memory_gb": 26.28, "pair": 0.33641698211431503, "sign": 0.43401009356603026, "keep": 0.35444832406938076, "leak": 0.6510431198403239, "replacement": 0.8934069573879242}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 7.001111030578613, "loss_scale": 4.0, "retries": 0, "seconds": 450.4, "max_memory_gb": 26.28, "pair": 0.313461109995842, "sign": 0.4436953440308571, "keep": 0.5083010658272542, "leak": 1.9201797395944595, "replacement": 1.2695042192935944}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 3.0589847564697266, "loss_scale": 4.0, "retries": 0, "seconds": 451.7, "max_memory_gb": 26.28, "pair": 0.34352948144078255, "sign": 0.37417106330394745, "keep": 0.2791533409617841, "leak": 0.7751436526887119, "replacement": 0.8639775887131691}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 4.204613208770752, "loss_scale": 4.0, "retries": 0, "seconds": 453.0, "max_memory_gb": 26.28, "pair": 0.30296516604721546, "sign": 0.1511630266904831, "keep": 0.02792476548347622, "leak": 1.8848814982920885, "replacement": 0.8637294471263885}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 7.155147075653076, "loss_scale": 4.0, "retries": 0, "seconds": 454.4, "max_memory_gb": 26.28, "pair": 0.35663075745105743, "sign": 0.38338254392147064, "keep": 0.16899983820621856, "leak": 1.8806741485896055, "replacement": 1.1029904633760452}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 7.023723125457764, "loss_scale": 4.0, "retries": 0, "seconds": 455.7, "max_memory_gb": 26.28, "pair": 0.45084916055202484, "sign": 0.6949665546417236, "keep": 0.24721440766006708, "leak": 0.5633167661726475, "replacement": 1.0627688318490982}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 2.833270788192749, "loss_scale": 4.0, "retries": 0, "seconds": 457.0, "max_memory_gb": 26.28, "pair": 0.1879094522446394, "sign": 0.09101543808355927, "keep": 0.04942207969725132, "leak": 0.49851948278956115, "replacement": 0.3827580725774169}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 7.590593338012695, "loss_scale": 4.0, "retries": 0, "seconds": 458.3, "max_memory_gb": 26.28, "pair": 0.31726166047155857, "sign": 0.5143218487501144, "keep": 0.40579164831433445, "leak": 1.1807911023497581, "replacement": 1.0725162029266357}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 5.6741042137146, "loss_scale": 4.0, "retries": 0, "seconds": 459.6, "max_memory_gb": 26.28, "pair": 0.4211025722324848, "sign": 0.48323512077331543, "keep": 0.5174025192391127, "leak": 0.6926178298890591, "replacement": 1.0945758670568466}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 3.2454612255096436, "loss_scale": 4.0, "retries": 0, "seconds": 460.8, "max_memory_gb": 26.28, "pair": 0.42600848712027073, "sign": 0.44309762120246887, "keep": 0.31609255488729104, "leak": 0.39907107409089804, "replacement": 0.905371330678463}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 4.71329402923584, "loss_scale": 4.0, "retries": 0, "seconds": 462.1, "max_memory_gb": 26.28, "pair": 0.4393872618675232, "sign": 0.42054102034308016, "keep": 0.252798093482852, "leak": 1.7904588817618787, "replacement": 1.2236715629696846}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 5.438689708709717, "loss_scale": 4.0, "retries": 0, "seconds": 463.5, "max_memory_gb": 26.28, "pair": 0.2002606801688671, "sign": 0.17551834508776665, "keep": 0.03387544094584882, "leak": 2.3781399726867676, "replacement": 0.8994925916194916}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 9.378976821899414, "loss_scale": 4.0, "retries": 0, "seconds": 464.8, "max_memory_gb": 26.28, "pair": 0.22767010144889355, "sign": 0.09421892929822206, "keep": 0.13292899308726192, "leak": 1.7637457698583603, "replacement": 0.782180517911911}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 6.109248638153076, "loss_scale": 4.0, "retries": 0, "seconds": 466.1, "max_memory_gb": 26.28, "pair": 0.5571452304720879, "sign": 0.7173533700406551, "keep": 0.5474718627519906, "leak": 0.3644232889637351, "replacement": 1.2806636542081833}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 3.560394048690796, "loss_scale": 4.0, "retries": 0, "seconds": 467.5, "max_memory_gb": 26.28, "pair": 0.3575534801930189, "sign": 0.5286333747208118, "keep": 0.05706683034077287, "leak": 1.0951890172436833, "replacement": 0.9242008477449417}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 6.304558753967285, "loss_scale": 4.0, "retries": 0, "seconds": 468.8, "max_memory_gb": 26.28, "pair": 0.1812609825283289, "sign": 0.182802377268672, "keep": 0.032889697584323585, "leak": 2.0244171023368835, "replacement": 0.7952113002538681}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 11.140544891357422, "loss_scale": 4.0, "retries": 0, "seconds": 470.1, "max_memory_gb": 26.28, "pair": 0.295977795496583, "sign": 0.5210405364632607, "keep": 0.02942440181504935, "leak": 1.3415794000029564, "replacement": 0.9066050946712494}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 6.457849025726318, "loss_scale": 4.0, "retries": 0, "seconds": 471.4, "max_memory_gb": 26.28, "pair": 0.4274008348584175, "sign": 0.43522691540420055, "keep": 1.000748822465539, "leak": 0.20858793193474412, "replacement": 1.1975356861948967}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 5.893736362457275, "loss_scale": 4.0, "retries": 0, "seconds": 472.6, "max_memory_gb": 26.28, "pair": 0.45158838480710983, "sign": 0.7561903223395348, "keep": 0.053465142380446196, "leak": 0.27961388416588306, "replacement": 0.9263195917010307}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 4.413305282592773, "loss_scale": 4.0, "retries": 0, "seconds": 474.0, "max_memory_gb": 26.28, "pair": 0.26801867224276066, "sign": 0.055524983908981085, "keep": 0.032713623251765966, "leak": 1.1496155247441493, "replacement": 0.5995418541133404}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 5.6621270179748535, "loss_scale": 4.0, "retries": 0, "seconds": 475.3, "max_memory_gb": 26.28, "pair": 0.4826826676726341, "sign": 0.4391771610826254, "keep": 0.2816549986600876, "leak": 0.07266179588623345, "replacement": 0.861264169216156}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 7.5109405517578125, "loss_scale": 4.0, "retries": 0, "seconds": 476.6, "max_memory_gb": 26.28, "pair": 0.3231014497578144, "sign": 0.3077436722815037, "keep": 0.08484400529414415, "leak": 2.029481053352356, "replacement": 1.0267655551433563}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 5.18032693862915, "loss_scale": 4.0, "retries": 0, "seconds": 477.9, "max_memory_gb": 26.28, "pair": 0.42433571442961693, "sign": 0.43383312225341797, "keep": 0.4197043066378683, "leak": 0.8615209795534611, "replacement": 1.0664846897125244}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 5.667147159576416, "loss_scale": 4.0, "retries": 0, "seconds": 479.1, "max_memory_gb": 26.28, "pair": 0.2625122703611851, "sign": 0.13327789679169655, "keep": 0.04081479262094945, "leak": 0.45193168707191944, "replacement": 0.46254153549671173}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 6.201188564300537, "loss_scale": 4.0, "retries": 0, "seconds": 480.5, "max_memory_gb": 26.28, "pair": 0.43588325567543507, "sign": 0.3777623730711639, "keep": 0.07561336862272583, "leak": 0.9050004651071504, "replacement": 0.8888212293386459}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 10.845788955688477, "loss_scale": 4.0, "retries": 0, "seconds": 481.8, "max_memory_gb": 26.28, "pair": 0.2559942863881588, "sign": 0.2927146013826132, "keep": 0.9306011162698269, "leak": 0.5234292484819889, "replacement": 0.998509481549263}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 6.912563323974609, "loss_scale": 4.0, "retries": 0, "seconds": 483.1, "max_memory_gb": 26.28, "pair": 0.15926032699644566, "sign": 0.06593660172075033, "keep": 0.2236478654667735, "leak": 1.282734863460064, "replacement": 0.6247362717986107}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 9.098976135253906, "loss_scale": 4.0, "retries": 0, "seconds": 484.4, "max_memory_gb": 26.28, "pair": 0.3607639688998461, "sign": 0.33199435845017433, "keep": 0.2071241462836042, "leak": 0.43032532185316086, "replacement": 0.7379045635461807}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 3.6203525066375732, "loss_scale": 4.0, "retries": 0, "seconds": 485.7, "max_memory_gb": 26.28, "pair": 0.545188695192337, "sign": 0.7044843137264252, "keep": 0.06518759997561574, "leak": 0.23600684222765267, "replacement": 0.989026352763176}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 8.725631713867188, "loss_scale": 4.0, "retries": 0, "seconds": 487.0, "max_memory_gb": 26.28, "pair": 0.3628728426992893, "sign": 0.3368691671639681, "keep": 0.5662325652083382, "leak": 0.5994729767553508, "replacement": 0.9642919525504112}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 5.454737186431885, "loss_scale": 4.0, "retries": 0, "seconds": 488.3, "max_memory_gb": 26.28, "pair": 0.4038461856544018, "sign": 0.4776020459830761, "keep": 0.50375146639999, "leak": 0.5686049126088619, "replacement": 1.036674179136753}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.5793955326080322, "loss_scale": 4.0, "retries": 0, "seconds": 489.6, "max_memory_gb": 26.28, "pair": 0.22219730913639069, "sign": 0.1935320757329464, "keep": 0.08304438833147287, "leak": 0.17047634348273277, "replacement": 0.40310461074113846}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 4.1396026611328125, "loss_scale": 4.0, "retries": 0, "seconds": 490.9, "max_memory_gb": 26.28, "pair": 0.3154195621609688, "sign": 0.33694721199572086, "keep": 0.06213492061942816, "leak": 0.49639955069869757, "replacement": 0.6390605121850967}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 6.065846920013428, "loss_scale": 4.0, "retries": 0, "seconds": 492.2, "max_memory_gb": 26.28, "pair": 0.26458962447941303, "sign": 0.2074395464733243, "keep": 0.11090549593791366, "leak": 0.8054832769557834, "replacement": 0.6251329556107521}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 10.025118827819824, "loss_scale": 4.0, "retries": 0, "seconds": 493.5, "max_memory_gb": 26.28, "pair": 0.27668760530650616, "sign": 0.355357825756073, "keep": 0.24326993757858872, "leak": 0.14583525341004133, "replacement": 0.6124603115022182}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 11.898134231567383, "loss_scale": 4.0, "retries": 0, "seconds": 494.8, "max_memory_gb": 26.28, "pair": 0.2799694207496941, "sign": 0.3620891124010086, "keep": 0.09797789435833693, "leak": 2.5487520694732666, "replacement": 1.1471909433603287}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 4.617804050445557, "loss_scale": 4.0, "retries": 0, "seconds": 496.1, "max_memory_gb": 26.28, "pair": 0.5656637847423553, "sign": 0.7780040353536606, "keep": 0.3946724981069565, "leak": 0.5349340737448074, "replacement": 1.2857355922460556}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 7.928899765014648, "loss_scale": 4.0, "retries": 0, "seconds": 497.4, "max_memory_gb": 26.28, "pair": 0.47089796513319016, "sign": 0.5755954310297966, "keep": 0.9814409278333187, "leak": 0.08688614005222917, "replacement": 1.2711376436054707}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 5.608850955963135, "loss_scale": 4.0, "retries": 0, "seconds": 498.7, "max_memory_gb": 26.28, "pair": 0.250204686075449, "sign": 0.2730231899768114, "keep": 0.11823465023189783, "leak": 0.9822554142738227, "replacement": 0.691397450864315}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 7.501232624053955, "loss_scale": 4.0, "retries": 0, "seconds": 500.0, "max_memory_gb": 26.28, "pair": 0.08983028819784522, "sign": 0.06189147708937526, "keep": 0.01984443888068199, "leak": 0.6618038034066558, "replacement": 0.29614919051527977}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 3.5892696380615234, "loss_scale": 4.0, "retries": 0, "seconds": 501.3, "max_memory_gb": 26.28, "pair": 0.40592763014137745, "sign": 0.475383585318923, "keep": 0.2333904046099633, "leak": 0.5577301217999775, "replacement": 0.8997471034526825}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 8.280488967895508, "loss_scale": 4.0, "retries": 0, "seconds": 502.7, "max_memory_gb": 26.28, "pair": 0.24038026481866837, "sign": 0.18097978876903653, "keep": 0.03674642078112811, "leak": 0.571783296307558, "replacement": 0.4921891763806343}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 10.174174308776855, "loss_scale": 4.0, "retries": 0, "seconds": 504.0, "max_memory_gb": 26.28, "pair": 0.2413214799016714, "sign": 0.4189876616001129, "keep": 0.07823368808021769, "leak": 1.3985896605881862, "replacement": 0.8395795524120331}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 7.385549545288086, "loss_scale": 4.0, "retries": 0, "seconds": 505.3, "max_memory_gb": 26.28, "pair": 0.23294942267239094, "sign": 0.2417778465896845, "keep": 0.11485252744751051, "leak": 1.0156964818015695, "replacement": 0.6651887446641922}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 4.971746921539307, "loss_scale": 4.0, "retries": 0, "seconds": 506.7, "max_memory_gb": 26.28, "pair": 0.220969095826149, "sign": 0.11017977353185415, "keep": 0.12748036812990904, "leak": 2.1939454656094313, "replacement": 0.8882855474948883}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 11.363710403442383, "loss_scale": 4.0, "retries": 0, "seconds": 508.0, "max_memory_gb": 26.28, "pair": 0.41314563155174255, "sign": 0.46555698476731777, "keep": 0.2106036152690649, "leak": 0.8347156144609471, "replacement": 0.9599048495292664}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 4.978890419006348, "loss_scale": 4.0, "retries": 0, "seconds": 509.3, "max_memory_gb": 26.28, "pair": 0.5517375897616148, "sign": 0.7413363512605429, "keep": 0.5395323783159256, "leak": 0.5673414701595902, "replacement": 1.3340073451399803}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 3.637488603591919, "loss_scale": 4.0, "retries": 0, "seconds": 510.6, "max_memory_gb": 26.28, "pair": 0.14035158418118954, "sign": 0.08517303224653006, "keep": 0.030443876748904586, "leak": 0.18287864373996854, "replacement": 0.24387969821691513}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 4.922802448272705, "loss_scale": 4.0, "retries": 0, "seconds": 511.9, "max_memory_gb": 26.28, "pair": 0.09425082709640265, "sign": 0.14549996610730886, "keep": 0.07574057416059077, "leak": 0.7344832504168153, "replacement": 0.3884919211268425}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 16.803848266601562, "loss_scale": 4.0, "retries": 0, "seconds": 513.3, "max_memory_gb": 26.28, "pair": 0.4673023037612438, "sign": 0.5144241936504841, "keep": 0.2761184419505298, "leak": 0.7104261298663914, "replacement": 1.0401801280677319}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 9.01704216003418, "loss_scale": 4.0, "retries": 0, "seconds": 514.6, "max_memory_gb": 26.28, "pair": 0.14530288241803646, "sign": 0.13419063948094845, "keep": 0.02393352671060711, "leak": 2.616337329149246, "replacement": 0.8784492909908295}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 7.304894924163818, "loss_scale": 4.0, "retries": 0, "seconds": 515.8, "max_memory_gb": 26.28, "pair": 0.21709940116852522, "sign": 0.07879334036260843, "keep": 0.12365358043462038, "leak": 0.32858665846288204, "replacement": 0.400469534099102}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 7.0263800621032715, "loss_scale": 4.0, "retries": 0, "seconds": 517.1, "max_memory_gb": 26.28, "pair": 0.3367229513823986, "sign": 0.4937247522175312, "keep": 0.23180286353453994, "leak": 1.0073691010475159, "replacement": 0.9513290822505951}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 6.792325496673584, "loss_scale": 4.0, "retries": 0, "seconds": 518.5, "max_memory_gb": 26.28, "pair": 0.3259736970067024, "sign": 0.3548004552721977, "keep": 0.16865959600545466, "leak": 0.9195219669491053, "replacement": 0.817584216594696}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 15.213200569152832, "loss_scale": 4.0, "retries": 0, "seconds": 519.8, "max_memory_gb": 26.28, "pair": 0.46216265484690666, "sign": 0.760008336044848, "keep": 0.9128334559500217, "leak": 0.7813944723457098, "replacement": 1.493932168930769}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 8.449938774108887, "loss_scale": 4.0, "retries": 0, "seconds": 521.1, "max_memory_gb": 26.28, "pair": 0.33361322432756424, "sign": 0.28345293179154396, "keep": 0.11026245402172208, "leak": 1.855886097997427, "replacement": 0.9944424703717232}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 19.40794563293457, "loss_scale": 4.0, "retries": 0, "seconds": 522.5, "max_memory_gb": 26.28, "pair": 0.3811732307076454, "sign": 0.4831916643306613, "keep": 0.4292537549044937, "leak": 0.9028188828378916, "replacement": 1.0631006695330143}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 10.591527938842773, "loss_scale": 4.0, "retries": 0, "seconds": 523.8, "max_memory_gb": 26.28, "pair": 0.7021204382181168, "sign": 0.8115069130435586, "keep": 0.34173568338155746, "leak": 0.3348203608766198, "replacement": 1.3624468445777893}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 6.609903812408447, "loss_scale": 4.0, "retries": 0, "seconds": 525.2, "max_memory_gb": 26.28, "pair": 0.17775314673781395, "sign": 0.11966606602072716, "keep": 0.011606100015342236, "leak": 1.4741320349276066, "replacement": 0.6119222417473793}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 15.216272354125977, "loss_scale": 4.0, "retries": 0, "seconds": 526.5, "max_memory_gb": 26.28, "pair": 0.3123920373618603, "sign": 0.5075776572339237, "keep": 0.4883821806870401, "leak": 0.9246297394856811, "replacement": 1.041529357433319}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 5.555199146270752, "loss_scale": 4.0, "retries": 0, "seconds": 527.8, "max_memory_gb": 26.28, "pair": 0.12602103129029274, "sign": 0.12365337181836367, "keep": 0.035687983327079564, "leak": 0.8824672922492027, "replacement": 0.4263085275888443}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 5.821077346801758, "loss_scale": 4.0, "retries": 0, "seconds": 529.1, "max_memory_gb": 26.28, "pair": 0.5664740763604641, "sign": 0.7111744601279497, "keep": 0.27359625429380685, "leak": 0.5592684326693416, "replacement": 1.1986765712499619}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 11.851696014404297, "loss_scale": 4.0, "retries": 0, "seconds": 530.5, "max_memory_gb": 26.28, "pair": 0.666274718940258, "sign": 0.8917595539242029, "keep": 2.2422378985211253, "leak": 0.32899954193271697, "replacement": 2.3155232667922974}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 5.747216701507568, "loss_scale": 4.0, "retries": 0, "seconds": 531.8, "max_memory_gb": 26.28, "pair": 0.45281522534787655, "sign": 0.5290680080652237, "keep": 0.4721298476215452, "leak": 0.506682489067316, "replacement": 1.0800848305225372}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 8.424934387207031, "loss_scale": 4.0, "retries": 0, "seconds": 533.1, "max_memory_gb": 26.28, "pair": 0.3740863371640444, "sign": 0.41301641426980495, "keep": 0.3321486446657218, "leak": 1.3023238344721904, "replacement": 1.0722498632967472}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 7.382655143737793, "loss_scale": 4.0, "retries": 0, "seconds": 534.4, "max_memory_gb": 26.28, "pair": 0.12800925597548485, "sign": 0.09490220341831446, "keep": 0.3988987730117515, "leak": 1.6515666097402573, "replacement": 0.7878014147281647}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 9.0015287399292, "loss_scale": 4.0, "retries": 0, "seconds": 535.8, "max_memory_gb": 26.28, "pair": 0.4128141328692436, "sign": 0.49746783822774887, "keep": 0.7329600697266869, "leak": 1.5409744915668853, "replacement": 1.4132717326283455}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 19.31804656982422, "loss_scale": 4.0, "retries": 0, "seconds": 537.1, "max_memory_gb": 26.28, "pair": 0.610160406678915, "sign": 0.7978007607161999, "keep": 0.4955239089904353, "leak": 0.34835768630728126, "replacement": 1.3439121842384338}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 5.271955490112305, "loss_scale": 4.0, "retries": 0, "seconds": 538.3, "max_memory_gb": 26.28, "pair": 0.1993299163877964, "sign": 0.13521066308021545, "keep": 0.03183375240769237, "leak": 1.4290058463811874, "replacement": 0.6401035785675049}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 10.176575660705566, "loss_scale": 4.0, "retries": 0, "seconds": 539.6, "max_memory_gb": 26.28, "pair": 0.3767487481236458, "sign": 0.3827452873811126, "keep": 0.7890815977007151, "leak": 0.2991774659603834, "replacement": 1.0374565869569778}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 10.947842597961426, "loss_scale": 4.0, "retries": 0, "seconds": 540.9, "max_memory_gb": 26.28, "pair": 0.3636547261849046, "sign": 0.38055075518786907, "keep": 1.2611850146204233, "leak": 0.6072207360994071, "replacement": 1.3363277986645699}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 4.446962356567383, "loss_scale": 4.0, "retries": 0, "seconds": 542.2, "max_memory_gb": 26.28, "pair": 0.24084034748375416, "sign": 0.14799228869378567, "keep": 0.0639508482418023, "leak": 0.3859446353453677, "replacement": 0.4432980753481388}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 13.536805152893066, "loss_scale": 4.0, "retries": 0, "seconds": 543.5, "max_memory_gb": 26.28, "pair": 0.8072037696838379, "sign": 1.1445667520165443, "keep": 1.3165212494786829, "leak": 0.5888275243341923, "replacement": 2.1849546879529953}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 6.119967937469482, "loss_scale": 4.0, "retries": 0, "seconds": 544.7, "max_memory_gb": 26.28, "pair": 0.17335558962076902, "sign": 0.1897957157343626, "keep": 0.03180460829753429, "leak": 0.7254021117696539, "replacement": 0.46550627797842026}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 10.224531173706055, "loss_scale": 4.0, "retries": 0, "seconds": 546.0, "max_memory_gb": 26.28, "pair": 0.4491526857018471, "sign": 0.6927931383252144, "keep": 0.07168973924126476, "leak": 1.6518531953915954, "replacement": 1.2443574517965317}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 11.802199363708496, "loss_scale": 4.0, "retries": 0, "seconds": 547.3, "max_memory_gb": 26.28, "pair": 0.21928647719323635, "sign": 0.387827951926738, "keep": 0.5893551036715508, "leak": 2.3191303461790085, "replacement": 1.2876605540513992}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 8.548897743225098, "loss_scale": 4.0, "retries": 0, "seconds": 548.6, "max_memory_gb": 26.28, "pair": 0.20774144679307938, "sign": 0.2695747744292021, "keep": 0.06321610097074881, "leak": 1.8987049227580428, "replacement": 0.8488131314516068}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 10.234172821044922, "loss_scale": 4.0, "retries": 0, "seconds": 549.9, "max_memory_gb": 26.28, "pair": 0.6306062377989292, "sign": 0.5716808084398508, "keep": 0.5349254412576556, "leak": 0.11965483048697934, "replacement": 1.2138231322169304}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 7.970039367675781, "loss_scale": 4.0, "retries": 0, "seconds": 551.2, "max_memory_gb": 26.28, "pair": 0.5810185074806213, "sign": 0.793942503631115, "keep": 0.9396864082664251, "leak": 0.7394230728968978, "replacement": 1.6326887756586075}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 4.840392112731934, "loss_scale": 4.0, "retries": 0, "seconds": 552.5, "max_memory_gb": 26.28, "pair": 0.12656983453780413, "sign": 0.06579420249909163, "keep": 0.044513145461678505, "leak": 1.0850582420825958, "replacement": 0.45298807322978973}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 9.848251342773438, "loss_scale": 4.0, "retries": 0, "seconds": 553.8, "max_memory_gb": 26.28, "pair": 0.3401553574949503, "sign": 0.20636871363967657, "keep": 0.055012308061122894, "leak": 0.7954418137669563, "replacement": 0.6697063073515892}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 5.761717796325684, "loss_scale": 4.0, "retries": 0, "seconds": 555.2, "max_memory_gb": 26.28, "pair": 0.5597714111208916, "sign": 0.6178694041445851, "keep": 0.378213869407773, "leak": 0.46427871473133564, "replacement": 1.1738827526569366}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 7.435377597808838, "loss_scale": 4.0, "retries": 0, "seconds": 556.5, "max_memory_gb": 26.28, "pair": 0.28087219782173634, "sign": 0.23398897051811218, "keep": 0.14097213465720415, "leak": 2.562594630802778, "replacement": 1.109001412987709}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 3.0010151863098145, "loss_scale": 4.0, "retries": 0, "seconds": 557.9, "max_memory_gb": 26.28, "pair": 0.19152753986418247, "sign": 0.10329900309443474, "keep": 0.012255370718776248, "leak": 0.522828979883343, "replacement": 0.3800119608640671}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 7.041054725646973, "loss_scale": 4.0, "retries": 0, "seconds": 559.3, "max_memory_gb": 26.28, "pair": 0.2562513742595911, "sign": 0.09817138314247131, "keep": 0.04306231773807667, "leak": 0.3265749476850033, "replacement": 0.4085119441151619}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 4.6188273429870605, "loss_scale": 4.0, "retries": 0, "seconds": 560.6, "max_memory_gb": 26.28, "pair": 0.26760114543139935, "sign": 0.2674037045799196, "keep": 0.24393170699477196, "leak": 0.7353970278054476, "replacement": 0.7071181088685989}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 8.434913635253906, "loss_scale": 4.0, "retries": 0, "seconds": 562.0, "max_memory_gb": 26.28, "pair": 0.15094349533319473, "sign": 0.06507474603131413, "keep": 0.08561798758455552, "leak": 1.3155503328889608, "replacement": 0.5551774427294731}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 4.526900291442871, "loss_scale": 4.0, "retries": 0, "seconds": 563.3, "max_memory_gb": 26.28, "pair": 0.18333834409713745, "sign": 0.21889393590390682, "keep": 0.0939147649332881, "leak": 0.2849678590137046, "replacement": 0.41098465397953987}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 3.9764158725738525, "loss_scale": 4.0, "retries": 0, "seconds": 564.6, "max_memory_gb": 26.28, "pair": 0.1692131720483303, "sign": 0.22408235259354115, "keep": 0.06358315888792276, "leak": 1.0544072836637497, "replacement": 0.5766477659344673}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 3.571425676345825, "loss_scale": 4.0, "retries": 0, "seconds": 565.9, "max_memory_gb": 26.28, "pair": 0.1694699078798294, "sign": 0.13486817106604576, "keep": 0.017367758642649278, "leak": 0.11631233734078705, "replacement": 0.2746659517288208}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 12.813505172729492, "loss_scale": 4.0, "retries": 0, "seconds": 567.2, "max_memory_gb": 26.28, "pair": 0.4230106263421476, "sign": 0.42635585088282824, "keep": 0.48878558399155736, "leak": 0.8042298071086407, "replacement": 1.0816388055682182}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 11.154669761657715, "loss_scale": 4.0, "retries": 0, "seconds": 568.4, "max_memory_gb": 26.28, "pair": 0.25460514053702354, "sign": 0.28068175725638866, "keep": 0.16408448852598667, "leak": 1.3937496449798346, "replacement": 0.8254256695508957}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 18.25314712524414, "loss_scale": 4.0, "retries": 0, "seconds": 569.7, "max_memory_gb": 26.28, "pair": 0.34052870236337185, "sign": 0.5352831711061299, "keep": 0.6560650644823909, "leak": 1.4067255556583405, "replacement": 1.2878842055797577}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 5.246870994567871, "loss_scale": 4.0, "retries": 0, "seconds": 571.0, "max_memory_gb": 26.28, "pair": 0.10058006085455418, "sign": 0.07555960025638342, "keep": 0.07230125361820683, "leak": 0.436239360133186, "replacement": 0.2835703305900097}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 6.198949337005615, "loss_scale": 4.0, "retries": 0, "seconds": 572.3, "max_memory_gb": 26.28, "pair": 0.3996739648282528, "sign": 0.4392273807898164, "keep": 0.3792448863387108, "leak": 0.1546380054205656, "replacement": 0.8475695848464966}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 5.900665283203125, "loss_scale": 4.0, "retries": 0, "seconds": 573.6, "max_memory_gb": 26.28, "pair": 0.2966356910765171, "sign": 0.38144106417894363, "keep": 0.026521043037064373, "leak": 0.4095393238385441, "replacement": 0.6030015647411346}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 18.47895050048828, "loss_scale": 4.0, "retries": 0, "seconds": 574.9, "max_memory_gb": 26.28, "pair": 0.7230513337999582, "sign": 0.8413011599332094, "keep": 1.2454376176465303, "leak": 1.3230423294007778, "replacement": 2.0971813425421715}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 10.342041969299316, "loss_scale": 4.0, "retries": 0, "seconds": 576.2, "max_memory_gb": 26.28, "pair": 0.3924682717770338, "sign": 0.41999361850321293, "keep": 0.11250113870482892, "leak": 2.0227901935577393, "replacement": 1.1644132137298584}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 33.626712799072266, "loss_scale": 4.0, "retries": 0, "seconds": 577.5, "max_memory_gb": 26.28, "pair": 0.33461276069283485, "sign": 0.38489479571580887, "keep": 0.1291172259952873, "leak": 1.0173488026484847, "replacement": 0.8459559977054596}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 9.010651588439941, "loss_scale": 4.0, "retries": 0, "seconds": 578.8, "max_memory_gb": 26.28, "pair": 0.30834028124809265, "sign": 0.4489407502114773, "keep": 0.306563803460449, "leak": 0.30671497993171215, "replacement": 0.7627712786197662}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 11.2263765335083, "loss_scale": 4.0, "retries": 0, "seconds": 580.1, "max_memory_gb": 26.28, "pair": 0.6571054570376873, "sign": 0.8280255291610956, "keep": 0.9219173279125243, "leak": 1.1210081038880162, "replacement": 1.8123289197683334}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 7.111232280731201, "loss_scale": 4.0, "retries": 0, "seconds": 581.4, "max_memory_gb": 26.28, "pair": 0.18610743433237076, "sign": 0.2034807913005352, "keep": 0.17431457480415702, "leak": 0.7922426667064428, "replacement": 0.5730657875537872}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 9.249542236328125, "loss_scale": 4.0, "retries": 0, "seconds": 582.7, "max_memory_gb": 26.28, "pair": 0.45267580822110176, "sign": 0.5862888135015965, "keep": 0.5294302707770839, "leak": 1.8145470800536714, "replacement": 1.464172087609768}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 7.197469234466553, "loss_scale": 4.0, "retries": 0, "seconds": 583.9, "max_memory_gb": 26.28, "pair": 0.23581946454942226, "sign": 0.25089193135499954, "keep": 0.3513607223285362, "leak": 0.4333551326417364, "replacement": 0.6452845558524132}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 18.235742568969727, "loss_scale": 4.0, "retries": 0, "seconds": 585.2, "max_memory_gb": 26.28, "pair": 0.18734959047287703, "sign": 0.17044763267040253, "keep": 0.0335251335054636, "leak": 1.5846466198563576, "replacement": 0.6854976117610931}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 12.996333122253418, "loss_scale": 4.0, "retries": 0, "seconds": 586.5, "max_memory_gb": 26.28, "pair": 0.23098550643771887, "sign": 0.1351159899495542, "keep": 0.3101126704714261, "leak": 1.5651921974495053, "replacement": 0.8448978587985039}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 6.968667984008789, "loss_scale": 4.0, "retries": 0, "seconds": 587.8, "max_memory_gb": 26.28, "pair": 0.18267452716827393, "sign": 0.126413868740201, "keep": 0.01622999261599034, "leak": 1.2064475491642952, "replacement": 0.5556083396077156}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 7.1997880935668945, "loss_scale": 4.0, "retries": 0, "seconds": 589.1, "max_memory_gb": 26.28, "pair": 0.385379396378994, "sign": 0.3489246126264334, "keep": 0.421898583881557, "leak": 1.3279570192098618, "replacement": 1.1027802526950836}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 19.882991790771484, "loss_scale": 4.0, "retries": 0, "seconds": 590.4, "max_memory_gb": 26.28, "pair": 0.3815876208245754, "sign": 0.3081627395004034, "keep": 0.05751890642568469, "leak": 1.195680633685697, "replacement": 0.863348588347435}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 10.690994262695312, "loss_scale": 4.0, "retries": 0, "seconds": 591.7, "max_memory_gb": 26.28, "pair": 0.4495713710784912, "sign": 0.2943497244268656, "keep": 0.39004652854055166, "leak": 0.8497427646070719, "replacement": 1.0042051747441292}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 6.891809463500977, "loss_scale": 4.0, "retries": 0, "seconds": 593.0, "max_memory_gb": 26.28, "pair": 0.12169767916202545, "sign": 0.15493313781917095, "keep": 0.13853297382593155, "leak": 2.4478750824928284, "replacement": 0.8803995102643967}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 6.954499244689941, "loss_scale": 4.0, "retries": 0, "seconds": 594.3, "max_memory_gb": 26.28, "pair": 0.40743155777454376, "sign": 0.7186728995293379, "keep": 0.052722674794495106, "leak": 0.5819910152931698, "replacement": 0.9386270567774773}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 10.173850059509277, "loss_scale": 4.0, "retries": 0, "seconds": 595.6, "max_memory_gb": 26.28, "pair": 0.13428630772978067, "sign": 0.23023291304707527, "keep": 0.021472725857165642, "leak": 1.0273355096578598, "replacement": 0.5169729962944984}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 10.58419418334961, "loss_scale": 4.0, "retries": 0, "seconds": 596.9, "max_memory_gb": 26.28, "pair": 0.32265205308794975, "sign": 0.4516526833176613, "keep": 0.288440584205091, "leak": 1.4497442151186988, "replacement": 1.055134728550911}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 9.541271209716797, "loss_scale": 4.0, "retries": 0, "seconds": 598.2, "max_memory_gb": 26.28, "pair": 0.273958601988852, "sign": 0.28696307726204395, "keep": 0.05593646550551057, "leak": 0.5602934961098072, "replacement": 0.585481733083725}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 5.161281585693359, "loss_scale": 4.0, "retries": 0, "seconds": 599.6, "max_memory_gb": 26.28, "pair": 0.2594893639907241, "sign": 0.2741676811128855, "keep": 0.03442355839069933, "leak": 0.40781038627028465, "replacement": 0.5157375633716583}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 9.601218223571777, "loss_scale": 4.0, "retries": 0, "seconds": 600.9, "max_memory_gb": 26.28, "pair": 0.25995370745658875, "sign": 0.5841922163963318, "keep": 0.6205310802906752, "leak": 1.572106957435608, "replacement": 1.255342110991478}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 6.747931003570557, "loss_scale": 4.0, "retries": 0, "seconds": 602.2, "max_memory_gb": 26.28, "pair": 0.15452684182673693, "sign": 0.19234922900795937, "keep": 0.021733716639573686, "leak": 0.9786642640829086, "replacement": 0.5062343776226044}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 9.091506004333496, "loss_scale": 4.0, "retries": 0, "seconds": 603.5, "max_memory_gb": 26.28, "pair": 0.240519305691123, "sign": 0.1977213155478239, "keep": 0.0330569283105433, "leak": 0.20509278507961426, "replacement": 0.40718161314725876}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 6.622665882110596, "loss_scale": 4.0, "retries": 0, "seconds": 604.8, "max_memory_gb": 26.28, "pair": 0.1764922197908163, "sign": 0.11207959987223148, "keep": 0.009544780681608245, "leak": 2.265606999397278, "replacement": 0.803706131875515}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 5.125246524810791, "loss_scale": 4.0, "retries": 0, "seconds": 606.1, "max_memory_gb": 26.28, "pair": 0.15877778083086014, "sign": 0.15959169575944543, "keep": 0.10638096125330776, "leak": 1.4223203733563423, "replacement": 0.647344209253788}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 6.75496244430542, "loss_scale": 4.0, "retries": 0, "seconds": 607.4, "max_memory_gb": 26.28, "pair": 0.32742196694016457, "sign": 0.2513120393268764, "keep": 0.08062028372660279, "leak": 0.5426611233597214, "replacement": 0.6290534287691116}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 4.646917819976807, "loss_scale": 4.0, "retries": 0, "seconds": 608.7, "max_memory_gb": 26.28, "pair": 0.19350736774504185, "sign": 0.13433637004345655, "keep": 0.06681476486846805, "leak": 0.6289344225078821, "replacement": 0.4513165354728699}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 14.631518363952637, "loss_scale": 4.0, "retries": 0, "seconds": 610.0, "max_memory_gb": 26.28, "pair": 0.30854297801852226, "sign": 0.20158926350995898, "keep": 0.05862705083563924, "leak": 0.033815975417383015, "replacement": 0.44710513204336166}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 14.216326713562012, "loss_scale": 4.0, "retries": 0, "seconds": 611.3, "max_memory_gb": 26.28, "pair": 0.5409332159906626, "sign": 0.21034610085189342, "keep": 0.5402548727579415, "leak": 1.3272989597171545, "replacement": 1.2480584383010864}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 19.668296813964844, "loss_scale": 4.0, "retries": 0, "seconds": 612.6, "max_memory_gb": 26.28, "pair": 0.31188710778951645, "sign": 0.11957514099776745, "keep": 0.024863683618605137, "leak": 0.08969359815819189, "replacement": 0.40652991086244583}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 8.323275566101074, "loss_scale": 4.0, "retries": 0, "seconds": 613.9, "max_memory_gb": 26.28, "pair": 0.5054820524528623, "sign": 0.5013595744967461, "keep": 0.3971579000353813, "leak": 1.0159931844100356, "replacement": 1.2087390571832657}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 13.414410591125488, "loss_scale": 4.0, "retries": 0, "seconds": 615.2, "max_memory_gb": 26.28, "pair": 0.3437313959002495, "sign": 0.3674881015904248, "keep": 0.056155655067414045, "leak": 0.6625224434537813, "replacement": 0.7211838886141777}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 10.706685066223145, "loss_scale": 4.0, "retries": 0, "seconds": 616.4, "max_memory_gb": 26.28, "pair": 0.2636167146265507, "sign": 0.17673315200954676, "keep": 0.1600270905182697, "leak": 0.31761818495579064, "replacement": 0.5114013850688934}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 5.829433917999268, "loss_scale": 4.0, "retries": 0, "seconds": 617.7, "max_memory_gb": 26.28, "pair": 0.35420522280037403, "sign": 0.42666900530457497, "keep": 0.24077366653364152, "leak": 0.21054144576191902, "replacement": 0.740561917424202}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 9.157299995422363, "loss_scale": 4.0, "retries": 0, "seconds": 619.0, "max_memory_gb": 26.28, "pair": 0.2753206081688404, "sign": 0.4018548056483269, "keep": 0.3375107451574877, "leak": 0.2262296830303967, "replacement": 0.7015607953071594}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 7.879817485809326, "loss_scale": 4.0, "retries": 0, "seconds": 620.3, "max_memory_gb": 26.28, "pair": 0.20966607332229614, "sign": 0.1709333099424839, "keep": 0.052205300191417336, "leak": 2.3468476831912994, "replacement": 0.9079472981393337}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 9.692464828491211, "loss_scale": 4.0, "retries": 0, "seconds": 621.6, "max_memory_gb": 26.28, "pair": 0.2970338612794876, "sign": 0.539516968652606, "keep": 0.02237351838266477, "leak": 0.7895460315048695, "replacement": 0.7753656134009361}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 5.7414140701293945, "loss_scale": 4.0, "retries": 0, "seconds": 622.9, "max_memory_gb": 26.28, "pair": 0.21981205977499485, "sign": 0.19207038776949048, "keep": 0.010368625866249204, "leak": 0.36184265837073326, "replacement": 0.4114922396838665}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 10.208468437194824, "loss_scale": 4.0, "retries": 0, "seconds": 624.2, "max_memory_gb": 26.28, "pair": 0.4110234063118696, "sign": 0.4894462749361992, "keep": 0.6185943830059841, "leak": 0.5090815713629127, "replacement": 1.0923141688108444}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 6.199305057525635, "loss_scale": 4.0, "retries": 0, "seconds": 625.5, "max_memory_gb": 26.28, "pair": 0.4407373443245888, "sign": 0.5523569397628307, "keep": 0.3995951348915696, "leak": 0.653507714276202, "replacement": 1.0800903141498566}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 13.168391227722168, "loss_scale": 4.0, "retries": 0, "seconds": 626.8, "max_memory_gb": 26.28, "pair": 0.16986916214227676, "sign": 0.1480295096989721, "keep": 0.0917029504198581, "leak": 0.6290350295603275, "replacement": 0.4469941481947899}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 7.4387383460998535, "loss_scale": 4.0, "retries": 0, "seconds": 628.0, "max_memory_gb": 26.28, "pair": 0.25313510932028294, "sign": 0.4233524762094021, "keep": 0.2769759704824537, "leak": 0.7535944953560829, "replacement": 0.7916979491710663}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 9.599164009094238, "loss_scale": 4.0, "retries": 0, "seconds": 629.3, "max_memory_gb": 26.28, "pair": 0.5237551182508469, "sign": 0.7600788772106171, "keep": 0.4527536053210497, "leak": 0.09172473009675741, "replacement": 1.153102546930313}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 20.193675994873047, "loss_scale": 4.0, "retries": 0, "seconds": 630.6, "max_memory_gb": 26.28, "pair": 0.7388845235109329, "sign": 1.025393282994628, "keep": 1.6358081884682178, "leak": 0.44383440585806966, "replacement": 2.180443748831749}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 11.775117874145508, "loss_scale": 4.0, "retries": 0, "seconds": 631.9, "max_memory_gb": 26.28, "pair": 0.2073325775563717, "sign": 0.46251824125647545, "keep": 0.02910862024873495, "leak": 1.3024844974279404, "replacement": 0.7787671312689781}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 10.534088134765625, "loss_scale": 4.0, "retries": 0, "seconds": 633.2, "max_memory_gb": 26.28, "pair": 0.44277617521584034, "sign": 0.6330259032547474, "keep": 0.7153828428126872, "leak": 0.5356741782452445, "replacement": 1.250899113714695}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 5.59792947769165, "loss_scale": 4.0, "retries": 0, "seconds": 634.4, "max_memory_gb": 26.28, "pair": 0.13895972818136215, "sign": 0.1254048072732985, "keep": 0.07346289046108723, "leak": 0.8647326361387968, "replacement": 0.4545767344534397}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 10.529611587524414, "loss_scale": 4.0, "retries": 0, "seconds": 635.7, "max_memory_gb": 26.28, "pair": 0.295959809795022, "sign": 0.3672817852348089, "keep": 0.16010860726237297, "leak": 1.7434580847620964, "replacement": 0.9955195412039757}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 14.156344413757324, "loss_scale": 4.0, "retries": 0, "seconds": 637.0, "max_memory_gb": 26.28, "pair": 0.15564454346895218, "sign": 0.26837687008082867, "keep": 0.5363723242189735, "leak": 2.0285147302784026, "replacement": 1.0651478171348572}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 8.120471000671387, "loss_scale": 4.0, "retries": 0, "seconds": 638.3, "max_memory_gb": 26.28, "pair": 0.1814711131155491, "sign": 0.2560703959316015, "keep": 0.04452806932386011, "leak": 0.767130209133029, "replacement": 0.5235529020428658}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 12.229398727416992, "loss_scale": 4.0, "retries": 0, "seconds": 639.6, "max_memory_gb": 26.28, "pair": 0.1563173709437251, "sign": 0.31426205299794674, "keep": 0.15018200228223577, "leak": 1.8543138895183802, "replacement": 0.8521178625524044}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 7.778289318084717, "loss_scale": 4.0, "retries": 0, "seconds": 640.9, "max_memory_gb": 26.28, "pair": 0.19685691874474287, "sign": 0.1326964944601059, "keep": 0.02861173392739147, "leak": 1.089259710162878, "replacement": 0.5498259663581848}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 11.874652862548828, "loss_scale": 4.0, "retries": 0, "seconds": 642.2, "max_memory_gb": 26.28, "pair": 0.5417236126959324, "sign": 0.7875974401831627, "keep": 0.31971948861610144, "leak": 1.152879236266017, "replacement": 1.3836019560694695}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 10.086359977722168, "loss_scale": 4.0, "retries": 0, "seconds": 643.5, "max_memory_gb": 26.28, "pair": 0.10176803916692734, "sign": 0.23261522874236107, "keep": 0.05591309163719416, "leak": 0.8865060077514499, "replacement": 0.46765870600938797}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 6.133735656738281, "loss_scale": 4.0, "retries": 0, "seconds": 644.9, "max_memory_gb": 26.28, "pair": 0.17012792080640793, "sign": 0.22368794213980436, "keep": 0.03701193421147764, "leak": 0.7316958718001842, "replacement": 0.4834018275141716}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 5.583495616912842, "loss_scale": 4.0, "retries": 0, "seconds": 646.1, "max_memory_gb": 26.28, "pair": 0.20955539215356112, "sign": 0.06411385908722878, "keep": 0.09511744696646929, "leak": 0.9813427675326238, "replacement": 0.534506730735302}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 22.231321334838867, "loss_scale": 4.0, "retries": 0, "seconds": 647.4, "max_memory_gb": 26.28, "pair": 0.3490856159478426, "sign": 0.5771957412362099, "keep": 0.8871286397334188, "leak": 1.2769568995572627, "replacement": 1.400487031787634}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 7.2265095710754395, "loss_scale": 4.0, "retries": 0, "seconds": 648.7, "max_memory_gb": 26.28, "pair": 0.40371283888816833, "sign": 0.3756448933854699, "keep": 0.24879621271975338, "leak": 1.1635072430203763, "replacement": 1.0068102404475212}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 10.007608413696289, "loss_scale": 4.0, "retries": 0, "seconds": 650.0, "max_memory_gb": 26.28, "pair": 0.39661335572600365, "sign": 0.3935224171727896, "keep": 0.44099891785299405, "leak": 0.869666799902916, "replacement": 1.0312907099723816}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 16.840471267700195, "loss_scale": 4.0, "retries": 0, "seconds": 651.3, "max_memory_gb": 26.28, "pair": 0.35489723086357117, "sign": 0.5814467754680663, "keep": 0.8230163803091273, "leak": 0.7439644697587937, "replacement": 1.2431199625134468}
diff --git a/runs/single_llava/replacement/L08_s13/adapter/README.md b/runs/single_llava/replacement/L08_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L08_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L08_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L08_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..30cb922adbf88a2db0ff6ffd38e299b03ca64277
--- /dev/null
+++ b/runs/single_llava/replacement/L08_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.8.mlp.gate_proj",
+    "model.language_model.layers.8.self_attn.v_proj",
+    "model.language_model.layers.8.self_attn.k_proj",
+    "model.language_model.layers.8.self_attn.o_proj",
+    "model.language_model.layers.8.mlp.up_proj",
+    "model.language_model.layers.8.self_attn.q_proj",
+    "model.language_model.layers.8.mlp.down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L08_s13/config.json b/runs/single_llava/replacement/L08_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..f6509190bc061bbe9bcb52d0c4708eb34d793cce
--- /dev/null
+++ b/runs/single_llava/replacement/L08_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 8,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L08_s13/steps.jsonl b/runs/single_llava/replacement/L08_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ccd8d0a959c077892b2a728d561cf26406d6e6ab
--- /dev/null
+++ b/runs/single_llava/replacement/L08_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.5765212774276733, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 25.23, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.8158561587333679, "loss_scale": 4.0, "retries": 0, "seconds": 2.6, "max_memory_gb": 25.24, "pair": 1.097008615732193, "sign": 0.5359146799892187, "keep": 5.373814415186644, "leak": 2.9637972716045624e-05, "replacement": 4.051880478858948}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.2313497066497803, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 25.34, "pair": 1.0954424142837524, "sign": 0.5414890274405479, "keep": 10.805621147155762, "leak": 6.124038787902464e-05, "replacement": 6.769012808799744}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 1.0584816932678223, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 25.36, "pair": 1.0901202261447906, "sign": 0.47044648230075836, "keep": 7.670173165388405, "leak": 0.00011932874100750723, "replacement": 5.1604598462581635}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.6006823778152466, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 25.36, "pair": 1.097646176815033, "sign": 0.7761100977659225, "keep": 0.15872790291905403, "leak": 1.8130561052487337e-07, "replacement": 1.5650652050971985}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.6321249008178711, "loss_scale": 4.0, "retries": 0, "seconds": 7.6, "max_memory_gb": 25.36, "pair": 1.0920039415359497, "sign": 0.7355471104383469, "keep": 2.9286814481019974, "leak": 0.0002356696768979205, "replacement": 2.9241771399974823}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 1.1966615915298462, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 25.36, "pair": 1.0675483345985413, "sign": 0.5458253119140863, "keep": 6.872829545289278, "leak": 0.0009069191648904962, "replacement": 4.777102440595627}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 1.1015018224716187, "loss_scale": 4.0, "retries": 0, "seconds": 10.1, "max_memory_gb": 25.36, "pair": 1.078220009803772, "sign": 0.5848022699356079, "keep": 7.378617957234383, "leak": 0.00044040033594683337, "replacement": 5.060040295124054}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 1.04661226272583, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 25.36, "pair": 1.0648942291736603, "sign": 0.6177874221466482, "keep": 5.707426756620407, "leak": 0.0016633575550939472, "replacement": 4.2279170751571655}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.8277616500854492, "loss_scale": 4.0, "retries": 0, "seconds": 12.6, "max_memory_gb": 25.36, "pair": 1.0750190615653992, "sign": 0.7787267453968525, "keep": 3.2703857868909836, "leak": 0.0010381975293967116, "replacement": 3.0998348593711853}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 1.3039690256118774, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 25.36, "pair": 1.0424061566591263, "sign": 0.4509822838008404, "keep": 5.249914187937975, "leak": 0.004113549281100859, "replacement": 3.8938828110694885}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 1.168756365776062, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 25.36, "pair": 1.0527074933052063, "sign": 0.6063821204006672, "keep": 3.0404079016298056, "leak": 0.005592452341716125, "replacement": 2.877500742673874}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 1.0085338354110718, "loss_scale": 4.0, "retries": 0, "seconds": 16.2, "max_memory_gb": 25.36, "pair": 1.0581272840499878, "sign": 0.5031227096915245, "keep": 5.336009310558438, "leak": 0.003423095733523951, "replacement": 3.9785490930080414}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.6024436950683594, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 25.36, "pair": 1.011758267879486, "sign": 0.39183283410966396, "keep": 7.762395277619362, "leak": 0.00633418615507253, "replacement": 5.090455949306488}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 1.0075156688690186, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 25.36, "pair": 1.0489063560962677, "sign": 0.6890762895345688, "keep": 5.154959993087687, "leak": 0.004816926593775861, "replacement": 3.9721286594867706}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 1.0042319297790527, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 25.36, "pair": 1.0624982863664627, "sign": 0.7789784073829651, "keep": 2.7474224967882037, "leak": 0.0024923344935814384, "replacement": 2.826321691274643}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.8547900319099426, "loss_scale": 4.0, "retries": 0, "seconds": 21.2, "max_memory_gb": 25.36, "pair": 1.0488169342279434, "sign": 0.7020992189645767, "keep": 3.1664982736110687, "leak": 0.009558082379498956, "replacement": 2.985505312681198}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.4594569206237793, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 25.42, "pair": 1.0304217636585236, "sign": 0.5246347412467003, "keep": 7.130885556340218, "leak": 0.004720128039480187, "replacement": 4.859361886978149}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 2.129804849624634, "loss_scale": 4.0, "retries": 0, "seconds": 23.7, "max_memory_gb": 25.42, "pair": 0.952722042798996, "sign": 0.5169250946491957, "keep": 10.651281595230103, "leak": 0.0229662605561316, "replacement": 6.542567133903503}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.7252713441848755, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 25.42, "pair": 1.0763904452323914, "sign": 0.4274130016565323, "keep": 0.022990503814071417, "leak": 0.000422711582359625, "replacement": 1.301697850227356}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.320664882659912, "loss_scale": 4.0, "retries": 0, "seconds": 26.2, "max_memory_gb": 25.42, "pair": 1.0120168179273605, "sign": 0.3662112392485142, "keep": 5.5173701462335885, "leak": 0.006617521219595801, "replacement": 3.9554618895053864}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.3294286727905273, "loss_scale": 4.0, "retries": 0, "seconds": 27.4, "max_memory_gb": 25.42, "pair": 0.9934450387954712, "sign": 0.563515480607748, "keep": 5.897952012717724, "leak": 0.028410581850039307, "replacement": 4.2312814593315125}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.5012996196746826, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 25.42, "pair": 0.9462591707706451, "sign": 0.23888478241860867, "keep": 5.298567249439657, "leak": 0.02900861136004096, "replacement": 3.722237318754196}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.3882293701171875, "loss_scale": 4.0, "retries": 0, "seconds": 29.9, "max_memory_gb": 25.42, "pair": 0.8953610956668854, "sign": 0.23661034367978573, "keep": 9.97956895828247, "leak": 0.02650535455904901, "replacement": 6.010077238082886}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.869806170463562, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 25.42, "pair": 0.9149631261825562, "sign": 0.3768521212041378, "keep": 8.401681706309319, "leak": 0.03702803451233194, "replacement": 5.313487023115158}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.3667752742767334, "loss_scale": 4.0, "retries": 0, "seconds": 32.5, "max_memory_gb": 25.42, "pair": 0.871290534734726, "sign": 0.6525606382638216, "keep": 10.42143726348877, "leak": 0.06763037899509072, "replacement": 6.425197124481201}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.9061627388000488, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 25.42, "pair": 0.9161076843738556, "sign": 0.5706954412162304, "keep": 8.34340389072895, "leak": 0.041804634674917907, "replacement": 5.38360857963562}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.4966455698013306, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 25.42, "pair": 0.9920758157968521, "sign": 0.7608871757984161, "keep": 5.579606398940086, "leak": 0.026462475609150715, "replacement": 4.168938159942627}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.3080294132232666, "loss_scale": 4.0, "retries": 0, "seconds": 36.2, "max_memory_gb": 25.42, "pair": 1.0062486380338669, "sign": 0.8414292335510254, "keep": 5.18822814989835, "leak": 0.020075941584764223, "replacement": 4.0260962545871735}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.5730091333389282, "loss_scale": 4.0, "retries": 0, "seconds": 37.5, "max_memory_gb": 25.42, "pair": 0.8820859491825104, "sign": 0.3382950136438012, "keep": 5.466029614210129, "leak": 0.07789806160872104, "replacement": 3.803722709417343}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.4752647876739502, "loss_scale": 4.0, "retries": 0, "seconds": 38.7, "max_memory_gb": 25.42, "pair": 0.9396856427192688, "sign": 0.7992085050791502, "keep": 5.632654070854187, "leak": 0.08177839564450551, "replacement": 4.176061570644379}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.5887157917022705, "loss_scale": 4.0, "retries": 0, "seconds": 40.0, "max_memory_gb": 25.42, "pair": 0.860206201672554, "sign": 0.3611222170293331, "keep": 10.634692192077637, "leak": 0.07854381436482072, "replacement": 6.37774932384491}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.8219945430755615, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 25.42, "pair": 0.937521293759346, "sign": 0.6063241455703974, "keep": 4.845971021801233, "leak": 0.11553295638987038, "replacement": 3.6925520598888397}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.638061761856079, "loss_scale": 4.0, "retries": 0, "seconds": 42.5, "max_memory_gb": 25.42, "pair": 0.9319678843021393, "sign": 0.5036569647490978, "keep": 5.225884044542909, "leak": 0.02761590707814321, "replacement": 3.803642362356186}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.893284559249878, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 25.42, "pair": 0.8377871066331863, "sign": 0.4165854100137949, "keep": 5.403741452842951, "leak": 0.09691661311080679, "replacement": 3.772179812192917}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.6474262475967407, "loss_scale": 4.0, "retries": 0, "seconds": 45.0, "max_memory_gb": 25.42, "pair": 0.8490214794874191, "sign": 0.4183246698230505, "keep": 5.1058241948485374, "leak": 0.08300982297805604, "replacement": 3.631848454475403}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.8502724170684814, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 25.42, "pair": 0.8556177169084549, "sign": 0.5695147402584553, "keep": 4.840626500546932, "leak": 0.13285852621777394, "replacement": 3.5939030051231384}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.006086587905884, "loss_scale": 4.0, "retries": 0, "seconds": 47.6, "max_memory_gb": 25.42, "pair": 0.7875731885433197, "sign": 0.7309417948126793, "keep": 9.323848366737366, "leak": 0.09415045217610896, "replacement": 5.838506102561951}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.5512720346450806, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 25.42, "pair": 0.9546074867248535, "sign": 0.5490302816033363, "keep": 2.786074438598007, "leak": 0.10546612931102572, "replacement": 2.64852637052536}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.074007749557495, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 25.42, "pair": 0.8595430552959442, "sign": 0.4705136679112911, "keep": 5.453724004328251, "leak": 0.06909739493858069, "replacement": 3.8389362394809723}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.9851019978523254, "loss_scale": 4.0, "retries": 0, "seconds": 51.3, "max_memory_gb": 25.42, "pair": 1.0233894884586334, "sign": 1.0102459192276, "keep": 2.9343428015708923, "leak": 0.03247382376139285, "replacement": 3.0038022696971893}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.200721025466919, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 25.42, "pair": 0.9233376383781433, "sign": 0.3845445401966572, "keep": 4.524344288744032, "leak": 0.02351010707207024, "replacement": 3.38365975022316}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.2938902378082275, "loss_scale": 4.0, "retries": 0, "seconds": 53.7, "max_memory_gb": 25.42, "pair": 0.8759017586708069, "sign": 0.49171337485313416, "keep": 4.721547789871693, "leak": 0.04785760532831773, "replacement": 3.4944967925548553}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.4675467014312744, "loss_scale": 4.0, "retries": 0, "seconds": 55.0, "max_memory_gb": 25.42, "pair": 0.921991378068924, "sign": 0.5557964369654655, "keep": 2.724086742848158, "leak": 0.055465360812377185, "replacement": 2.5757994055747986}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.253749132156372, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 25.42, "pair": 0.9997886568307877, "sign": 0.8770933896303177, "keep": 2.755152426660061, "leak": 0.028014777276042402, "replacement": 2.8229153752326965}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.2709262371063232, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 25.42, "pair": 0.8969383090734482, "sign": 0.6883236616849899, "keep": 4.463542357087135, "leak": 0.09639881375915138, "replacement": 3.496971070766449}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.319640874862671, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 25.42, "pair": 0.7851119935512543, "sign": 0.6086901668459177, "keep": 3.9685668917372823, "leak": 0.12080789473839104, "replacement": 3.103942573070526}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 4.02712869644165, "loss_scale": 4.0, "retries": 0, "seconds": 60.0, "max_memory_gb": 25.42, "pair": 0.7076383829116821, "sign": 0.18419071240350604, "keep": 7.270479567348957, "leak": 0.1887433456722647, "replacement": 4.48215925693512}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.5317318439483643, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 25.42, "pair": 0.696734644472599, "sign": 0.5844366336241364, "keep": 6.857986065559089, "leak": 0.24955774703994393, "replacement": 4.4803354144096375}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 3.792085886001587, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 25.42, "pair": 0.8440062999725342, "sign": 0.4112534746527672, "keep": 6.839911375194788, "leak": 0.09613290341076208, "replacement": 4.493621915578842}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.5705476999282837, "loss_scale": 4.0, "retries": 0, "seconds": 63.7, "max_memory_gb": 25.42, "pair": 0.9069118648767471, "sign": 0.5629236288368702, "keep": 2.606173443607986, "leak": 0.09931209070055047, "replacement": 2.5162883698940277}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 3.0446081161499023, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 25.42, "pair": 0.78099425137043, "sign": 0.657068032771349, "keep": 4.425862118601799, "leak": 0.21082855737768114, "replacement": 3.375166356563568}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 3.774160385131836, "loss_scale": 4.0, "retries": 0, "seconds": 66.2, "max_memory_gb": 25.42, "pair": 0.8426375985145569, "sign": 0.7027880623936653, "keep": 7.06965209543705, "leak": 0.17326583793692407, "replacement": 4.772174119949341}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 4.8129072189331055, "loss_scale": 4.0, "retries": 0, "seconds": 67.4, "max_memory_gb": 25.42, "pair": 0.7160635516047478, "sign": 0.5752414688467979, "keep": 7.355488285422325, "leak": 0.22060849773697555, "replacement": 4.736580550670624}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 3.6554484367370605, "loss_scale": 4.0, "retries": 0, "seconds": 68.7, "max_memory_gb": 25.42, "pair": 0.7525698244571686, "sign": 0.3777269124984741, "keep": 4.404232404194772, "leak": 0.1629051441559568, "replacement": 3.1842756867408752}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 7.529952526092529, "loss_scale": 4.0, "retries": 0, "seconds": 69.9, "max_memory_gb": 25.42, "pair": 0.48275724053382874, "sign": 0.13486567698419094, "keep": 8.309267520904541, "leak": 0.345624215900898, "replacement": 4.791229844093323}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 2.0411629676818848, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 25.42, "pair": 0.9048236012458801, "sign": 0.7825366631150246, "keep": 2.232836472801864, "leak": 0.036747480160556734, "replacement": 2.421696960926056}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 4.3505754470825195, "loss_scale": 4.0, "retries": 0, "seconds": 72.3, "max_memory_gb": 25.42, "pair": 0.8136658519506454, "sign": 0.2564026489853859, "keep": 4.224733951734379, "leak": 0.2608320275321603, "replacement": 3.1194422841072083}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.1937875747680664, "loss_scale": 4.0, "retries": 0, "seconds": 73.6, "max_memory_gb": 25.42, "pair": 0.9231660515069962, "sign": 0.6386078149080276, "keep": 2.6704170405864716, "leak": 0.011470995690615382, "replacement": 2.5805461406707764}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 4.914062976837158, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 25.42, "pair": 0.7222043499350548, "sign": 0.6016691029071808, "keep": 4.10953863710165, "leak": 0.14751523034647107, "replacement": 3.114687144756317}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 7.90870475769043, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 25.42, "pair": 0.7323026061058044, "sign": 0.6554791871458292, "keep": 5.694271624088287, "leak": 0.33467969184857793, "replacement": 3.990847885608673}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 3.0537261962890625, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 25.42, "pair": 0.88705874979496, "sign": 0.6823335979133844, "keep": 2.3881483916193247, "leak": 0.04239751482964493, "replacement": 2.432899236679077}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 9.179864883422852, "loss_scale": 4.0, "retries": 0, "seconds": 78.5, "max_memory_gb": 25.42, "pair": 0.745907187461853, "sign": 0.5373192578554153, "keep": 6.123265757225454, "leak": 0.1288689022185281, "replacement": 4.108416944742203}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 10.66357135772705, "loss_scale": 4.0, "retries": 0, "seconds": 79.8, "max_memory_gb": 25.46, "pair": 0.6265549138188362, "sign": 0.3997749164700508, "keep": 5.433262558653951, "leak": 0.3862622210290283, "replacement": 3.6396393477916718}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 9.923813819885254, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 25.46, "pair": 0.7805452346801758, "sign": 0.5218541249632835, "keep": 5.564305454492569, "leak": 0.1842867630184628, "replacement": 3.8696967363357544}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 2.7966277599334717, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 25.46, "pair": 1.0250246971845627, "sign": 0.6113120466470718, "keep": 1.5792881434317678, "leak": 0.0471209765819367, "replacement": 2.1321049630641937}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 12.027156829833984, "loss_scale": 4.0, "retries": 0, "seconds": 83.6, "max_memory_gb": 25.46, "pair": 0.5908755883574486, "sign": 0.3640171820297837, "keep": 4.371624112129211, "leak": 0.7789161924738437, "replacement": 3.1534253358840942}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 4.187835216522217, "loss_scale": 4.0, "retries": 0, "seconds": 84.8, "max_memory_gb": 25.46, "pair": 0.9167548418045044, "sign": 0.4918398968875408, "keep": 1.8641203816514462, "leak": 0.11272883875062689, "replacement": 2.1229170858860016}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 8.873616218566895, "loss_scale": 4.0, "retries": 0, "seconds": 86.0, "max_memory_gb": 25.46, "pair": 0.7614781484007835, "sign": 0.6668198481202126, "keep": 2.836698288563639, "leak": 0.11699560178385582, "replacement": 2.542486160993576}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 6.062408447265625, "loss_scale": 4.0, "retries": 0, "seconds": 87.3, "max_memory_gb": 25.46, "pair": 0.9052592813968658, "sign": 0.5467319339513779, "keep": 2.496545407921076, "leak": 0.4593985246610828, "replacement": 2.541747659444809}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 11.138996124267578, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 25.46, "pair": 0.6792138814926147, "sign": 0.1490212120115757, "keep": 3.651618020609021, "leak": 0.3551695051137358, "replacement": 2.668325901031494}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 12.340920448303223, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 25.46, "pair": 0.5485713630914688, "sign": 0.21771417185664177, "keep": 4.104336440563202, "leak": 0.40898160729557276, "replacement": 2.811842054128647}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.585174560546875, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 25.46, "pair": 0.9680910557508469, "sign": 0.7793017029762268, "keep": 0.4134921059012413, "leak": 0.017045909517037217, "replacement": 1.5687495470046997}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 6.195314407348633, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 25.46, "pair": 0.93171276897192, "sign": 0.49208085983991623, "keep": 1.854853346478194, "leak": 0.12361750967102125, "replacement": 2.136084258556366}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.700856924057007, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 25.46, "pair": 0.8738308548927307, "sign": 0.8830391522496939, "keep": 2.249092325568199, "leak": 0.18058367949561216, "replacement": 2.4850425124168396}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 7.170273303985596, "loss_scale": 4.0, "retries": 0, "seconds": 94.8, "max_memory_gb": 25.46, "pair": 0.7091128677129745, "sign": 0.4169656131416559, "keep": 1.4885892439633608, "leak": 0.2386665504600387, "replacement": 1.7215570211410522}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 10.337739944458008, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 25.46, "pair": 0.8777855187654495, "sign": 0.4638506844639778, "keep": 2.70418119430542, "leak": 0.5952098986599594, "replacement": 2.6106038689613342}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 5.3635029792785645, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 25.46, "pair": 0.8985994458198547, "sign": 0.7140629850327969, "keep": 1.6092321500182152, "leak": 0.5244899559766054, "replacement": 2.1913694739341736}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 3.7580394744873047, "loss_scale": 4.0, "retries": 0, "seconds": 98.5, "max_memory_gb": 25.46, "pair": 0.870919942855835, "sign": 0.5563829988241196, "keep": 0.5898372093215585, "leak": 0.036489418786004535, "replacement": 1.453152358531952}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 6.364425182342529, "loss_scale": 4.0, "retries": 0, "seconds": 99.7, "max_memory_gb": 25.46, "pair": 0.9713619947433472, "sign": 0.39681561291217804, "keep": 1.0302995352540165, "leak": 0.24033088629948907, "replacement": 1.7450023293495178}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 8.966937065124512, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 25.46, "pair": 0.6912001371383667, "sign": 0.442515192553401, "keep": 1.6750352308154106, "leak": 0.7599499735515565, "replacement": 1.939962923526764}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 8.49794864654541, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 25.46, "pair": 0.9616348147392273, "sign": 0.4229799062013626, "keep": 1.2862998992204666, "leak": 0.13152374762285035, "replacement": 1.8491557240486145}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 7.534999847412109, "loss_scale": 4.0, "retries": 0, "seconds": 103.5, "max_memory_gb": 25.46, "pair": 0.7794287353754044, "sign": 0.3771379888057709, "keep": 1.3500701421871781, "leak": 0.763153257779777, "replacement": 1.8338211476802826}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 3.882585048675537, "loss_scale": 4.0, "retries": 0, "seconds": 104.7, "max_memory_gb": 25.46, "pair": 0.9692078232765198, "sign": 0.7100840583443642, "keep": 0.7183718294836581, "leak": 0.5907812586665386, "replacement": 1.8311311602592468}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 2.9872031211853027, "loss_scale": 4.0, "retries": 0, "seconds": 106.0, "max_memory_gb": 25.46, "pair": 1.0153859108686447, "sign": 0.5702120810747147, "keep": 0.12761597707867622, "leak": 0.022591959801502526, "replacement": 1.3699479401111603}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 7.2624640464782715, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 25.46, "pair": 0.9334437549114227, "sign": 0.6660198047757149, "keep": 0.9992249980568886, "leak": 0.06545527020352893, "replacement": 1.7824299931526184}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 5.337795257568359, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 25.46, "pair": 0.6210580393671989, "sign": 0.34829691611230373, "keep": 0.6517519652843475, "leak": 1.198959049768746, "replacement": 1.420822262763977}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 4.596351623535156, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 25.46, "pair": 0.9506511986255646, "sign": 0.6343177724629641, "keep": 0.6147174048237503, "leak": 0.12057693031965755, "replacement": 1.605313003063202}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 3.12906813621521, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 25.46, "pair": 1.0583987832069397, "sign": 0.7012807279825211, "keep": 0.2386076597031206, "leak": 0.0058045508922077715, "replacement": 1.5297942161560059}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 4.147176265716553, "loss_scale": 4.0, "retries": 0, "seconds": 112.2, "max_memory_gb": 25.46, "pair": 0.721603125333786, "sign": 0.5197168365120888, "keep": 0.5272631957195699, "leak": 0.47139439126476645, "replacement": 1.362941712141037}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.5832985639572144, "loss_scale": 4.0, "retries": 0, "seconds": 113.4, "max_memory_gb": 25.46, "pair": 0.9435684531927109, "sign": 0.9610684514045715, "keep": 0.5878043957054615, "leak": 0.030582530782339745, "replacement": 1.7256504595279694}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.0890371799468994, "loss_scale": 4.0, "retries": 0, "seconds": 114.7, "max_memory_gb": 25.46, "pair": 0.8931360095739365, "sign": 0.5121897682547569, "keep": 0.26711086416617036, "leak": 1.3483710977016017, "replacement": 1.6198790669441223}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 5.3210930824279785, "loss_scale": 4.0, "retries": 0, "seconds": 115.9, "max_memory_gb": 25.46, "pair": 0.7703152149915695, "sign": 0.28902885131537914, "keep": 0.47455573990009725, "leak": 0.03930326167034082, "replacement": 1.161933273077011}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 3.3758127689361572, "loss_scale": 4.0, "retries": 0, "seconds": 117.2, "max_memory_gb": 25.46, "pair": 0.5720995739102364, "sign": 0.35007875598967075, "keep": 0.46665582060813904, "leak": 0.6205017781048809, "replacement": 1.135592296719551}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 2.405839443206787, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 25.46, "pair": 0.5777993723750114, "sign": 0.4306214302778244, "keep": 0.6976310908794403, "leak": 0.06324904994107783, "replacement": 1.157737836241722}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 5.86315393447876, "loss_scale": 4.0, "retries": 0, "seconds": 119.7, "max_memory_gb": 25.46, "pair": 0.6808981001377106, "sign": 0.40849672816693783, "keep": 0.487028107047081, "leak": 2.6724364364199573, "replacement": 1.7967697083950043}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.661472797393799, "loss_scale": 4.0, "retries": 0, "seconds": 120.9, "max_memory_gb": 25.46, "pair": 0.740084633231163, "sign": 0.46072274073958397, "keep": 0.4081840179860592, "leak": 0.4358061645179987, "replacement": 1.2834895551204681}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 2.3898255825042725, "loss_scale": 4.0, "retries": 0, "seconds": 122.1, "max_memory_gb": 25.46, "pair": 0.7171211168169975, "sign": 0.4978572092950344, "keep": 0.1880811471492052, "leak": 0.7115734489634633, "replacement": 1.2379836440086365}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 1.9981448650360107, "loss_scale": 4.0, "retries": 0, "seconds": 123.4, "max_memory_gb": 25.46, "pair": 0.9258630126714706, "sign": 0.864818386733532, "keep": 0.3445812235586345, "leak": 0.06246367860967439, "replacement": 1.5461787432432175}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 2.597362756729126, "loss_scale": 4.0, "retries": 0, "seconds": 124.7, "max_memory_gb": 25.46, "pair": 0.6575074382126331, "sign": 0.608895068988204, "keep": 0.503706805408001, "leak": 1.358502991264686, "replacement": 1.5534341037273407}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 2.9697957038879395, "loss_scale": 4.0, "retries": 0, "seconds": 125.9, "max_memory_gb": 25.46, "pair": 0.43242479488253593, "sign": 0.14736171066761017, "keep": 0.3117159456014633, "leak": 0.38034188747406006, "replacement": 0.7570491135120392}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 3.3059306144714355, "loss_scale": 4.0, "retries": 0, "seconds": 127.2, "max_memory_gb": 25.46, "pair": 0.5234435424208641, "sign": 0.30661657452583313, "keep": 0.9319830760359764, "leak": 1.0349412197247148, "replacement": 1.4014786779880524}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 2.198183059692383, "loss_scale": 4.0, "retries": 0, "seconds": 128.4, "max_memory_gb": 25.46, "pair": 0.5834579542279243, "sign": 0.3253142684698105, "keep": 0.15550783136859536, "leak": 1.7413853332400322, "replacement": 1.2592153549194336}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 3.891340494155884, "loss_scale": 4.0, "retries": 0, "seconds": 129.6, "max_memory_gb": 25.46, "pair": 0.7546005100011826, "sign": 0.8323622271418571, "keep": 0.11354915029369295, "leak": 0.19694721105042845, "replacement": 1.2767930030822754}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 2.313025951385498, "loss_scale": 4.0, "retries": 0, "seconds": 130.9, "max_memory_gb": 25.46, "pair": 0.4944312311708927, "sign": 0.35157996974885464, "keep": 0.2829493209719658, "leak": 2.340913087595254, "replacement": 1.3969241380691528}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 1.8645535707473755, "loss_scale": 4.0, "retries": 0, "seconds": 132.1, "max_memory_gb": 25.46, "pair": 0.4143267199397087, "sign": 0.39492356590926647, "keep": 0.26818957179784775, "leak": 2.813131819719274, "replacement": 1.4491662383079529}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.6683807373046875, "loss_scale": 4.0, "retries": 0, "seconds": 133.4, "max_memory_gb": 25.46, "pair": 0.7865010499954224, "sign": 0.6894723922014236, "keep": 0.27133815828710794, "leak": 0.10849593005605129, "replacement": 1.2940303087234497}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.2255144119262695, "loss_scale": 4.0, "retries": 0, "seconds": 134.7, "max_memory_gb": 25.46, "pair": 0.7494060359895229, "sign": 0.68155487626791, "keep": 0.20487324567511678, "leak": 2.186521671304945, "replacement": 1.739250510931015}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 3.151155710220337, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 25.46, "pair": 0.3912345673888922, "sign": 0.4025651100091636, "keep": 0.41207975149154663, "leak": 2.519260580302216, "replacement": 1.4283721819519997}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 2.7232325077056885, "loss_scale": 4.0, "retries": 0, "seconds": 137.3, "max_memory_gb": 25.46, "pair": 0.5666920319199562, "sign": 0.4878858821466565, "keep": 0.3638361841440201, "leak": 1.5489750523120165, "replacement": 1.3797968029975891}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 1.8966833353042603, "loss_scale": 4.0, "retries": 0, "seconds": 138.6, "max_memory_gb": 25.46, "pair": 0.4312582816928625, "sign": 0.3923139818944037, "keep": 0.11776158306747675, "leak": 0.457110732793808, "replacement": 0.8005737587809563}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 1.5954902172088623, "loss_scale": 4.0, "retries": 0, "seconds": 139.8, "max_memory_gb": 25.46, "pair": 0.5025206804275513, "sign": 0.3805612940341234, "keep": 0.20522443298250437, "leak": 0.7928367927670479, "replacement": 0.9936227351427078}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 1.6672351360321045, "loss_scale": 4.0, "retries": 0, "seconds": 141.0, "max_memory_gb": 25.46, "pair": 0.5541345812380314, "sign": 0.48574738297611475, "keep": 0.22399274073541164, "leak": 0.5826385032851249, "replacement": 1.05466428399086}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 3.291398048400879, "loss_scale": 4.0, "retries": 0, "seconds": 142.3, "max_memory_gb": 25.46, "pair": 0.3884265813976526, "sign": 0.4277426479384303, "keep": 0.5699148029088974, "leak": 1.9157171994447708, "replacement": 1.3661845996975899}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 1.904647707939148, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 25.46, "pair": 0.5857160501182079, "sign": 0.535952515900135, "keep": 0.386710450053215, "leak": 1.311348001821898, "replacement": 1.3748845309019089}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 1.9264552593231201, "loss_scale": 4.0, "retries": 0, "seconds": 144.8, "max_memory_gb": 25.46, "pair": 0.6981057375669479, "sign": 0.6096326187252998, "keep": 0.2017857232131064, "leak": 0.19511410052655265, "replacement": 1.1525934785604477}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 2.961179494857788, "loss_scale": 4.0, "retries": 0, "seconds": 146.1, "max_memory_gb": 25.46, "pair": 0.532254409044981, "sign": 0.38958392571657896, "keep": 0.5264092308934778, "leak": 1.2143412493169308, "replacement": 1.2938362658023834}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 1.3430533409118652, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 25.46, "pair": 0.6208670325577259, "sign": 0.5005530123598874, "keep": 0.2837408341001719, "leak": 0.487854055914795, "replacement": 1.1349775046110153}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.0592005252838135, "loss_scale": 4.0, "retries": 0, "seconds": 148.5, "max_memory_gb": 25.46, "pair": 0.759590171277523, "sign": 0.788739874958992, "keep": 0.1836950364522636, "leak": 0.48265969194471836, "replacement": 1.3664725422859192}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 2.7524609565734863, "loss_scale": 4.0, "retries": 0, "seconds": 149.8, "max_memory_gb": 25.46, "pair": 0.49954162910580635, "sign": 0.4669039167929441, "keep": 0.6559891528449953, "leak": 0.9325656129512936, "replacement": 1.2941295728087425}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.1199519634246826, "loss_scale": 4.0, "retries": 0, "seconds": 151.1, "max_memory_gb": 25.46, "pair": 0.705906830728054, "sign": 0.6522185727953911, "keep": 0.5575399016961455, "leak": 2.193146103993058, "replacement": 1.859072595834732}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 2.392951250076294, "loss_scale": 4.0, "retries": 0, "seconds": 152.4, "max_memory_gb": 25.46, "pair": 0.6019912175834179, "sign": 0.5970671437680721, "keep": 0.21272642677649856, "leak": 0.3237980665922784, "replacement": 1.087837502360344}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.773876667022705, "loss_scale": 4.0, "retries": 0, "seconds": 153.6, "max_memory_gb": 25.46, "pair": 0.670114429667592, "sign": 0.7424410381354392, "keep": 0.8604445867240429, "leak": 0.7310540783219039, "replacement": 1.6543207466602325}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 2.747640609741211, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 25.46, "pair": 0.8686773926019669, "sign": 1.0248692780733109, "keep": 0.504659267142415, "leak": 0.08175396043952787, "replacement": 1.653880164027214}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 1.6736572980880737, "loss_scale": 4.0, "retries": 0, "seconds": 156.0, "max_memory_gb": 25.46, "pair": 0.7165298983454704, "sign": 0.7475696876645088, "keep": 0.2265207739546895, "leak": 0.2537155896425247, "replacement": 1.2670040875673294}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 1.498894453048706, "loss_scale": 4.0, "retries": 0, "seconds": 157.2, "max_memory_gb": 25.46, "pair": 0.6148282419890165, "sign": 0.6018518712371588, "keep": 0.5759917162358761, "leak": 0.5282878573780181, "replacement": 1.335821993649006}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 1.5560859441757202, "loss_scale": 4.0, "retries": 0, "seconds": 158.5, "max_memory_gb": 25.46, "pair": 0.5604732111096382, "sign": 0.32901096250861883, "keep": 0.08320952043868601, "leak": 0.26137345284223557, "replacement": 0.8319268301129341}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 1.960890769958496, "loss_scale": 4.0, "retries": 0, "seconds": 159.7, "max_memory_gb": 25.46, "pair": 0.38109543547034264, "sign": 0.2614666451700032, "keep": 0.21508286707103252, "leak": 1.1777261476963758, "replacement": 0.9138017147779465}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 1.659514307975769, "loss_scale": 4.0, "retries": 0, "seconds": 161.0, "max_memory_gb": 25.46, "pair": 0.28207326121628284, "sign": 0.1650098729878664, "keep": 0.14096220582723618, "leak": 2.0091181667667115, "replacement": 0.9373388513922691}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 3.8006887435913086, "loss_scale": 4.0, "retries": 0, "seconds": 162.2, "max_memory_gb": 25.46, "pair": 0.5655856020748615, "sign": 0.8384891822934151, "keep": 0.5437520891427994, "leak": 1.1327902637422085, "replacement": 1.539903849363327}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 2.209486961364746, "loss_scale": 4.0, "retries": 0, "seconds": 163.5, "max_memory_gb": 25.46, "pair": 0.6608124449849129, "sign": 0.6596218273043633, "keep": 0.5451114177703857, "leak": 0.24132823257241398, "replacement": 1.3235111683607101}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 3.215632200241089, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 25.46, "pair": 0.3468355555087328, "sign": 0.11285527050495148, "keep": 0.7756484625861049, "leak": 1.1030246019363403, "replacement": 1.0668435841798782}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 2.633983850479126, "loss_scale": 4.0, "retries": 0, "seconds": 166.0, "max_memory_gb": 25.46, "pair": 0.4306075517088175, "sign": 0.3905102629214525, "keep": 0.5042911805212498, "leak": 2.3572414023801684, "replacement": 1.4673186391592026}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 2.931087017059326, "loss_scale": 4.0, "retries": 0, "seconds": 167.2, "max_memory_gb": 25.46, "pair": 0.7263197898864746, "sign": 0.6579933315515518, "keep": 0.32817721576429904, "leak": 0.28552595619112253, "replacement": 1.2907865643501282}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 3.468492031097412, "loss_scale": 4.0, "retries": 0, "seconds": 168.5, "max_memory_gb": 25.46, "pair": 0.3739943001419306, "sign": 0.3806302179582417, "keep": 0.21112408488988876, "leak": 3.5545534657430835, "replacement": 1.5585098266601562}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 3.4349217414855957, "loss_scale": 4.0, "retries": 0, "seconds": 169.7, "max_memory_gb": 25.46, "pair": 0.5049661882221699, "sign": 0.4411971541121602, "keep": 0.5102320853620768, "leak": 0.31721698865294456, "replacement": 1.0599850341677666}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 4.3043975830078125, "loss_scale": 4.0, "retries": 0, "seconds": 171.0, "max_memory_gb": 25.46, "pair": 0.5642004050314426, "sign": 0.4394514746963978, "keep": 0.2413558978587389, "leak": 2.5448969434946775, "replacement": 1.5408283025026321}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 1.6547106504440308, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 25.46, "pair": 0.36789266765117645, "sign": 0.23853418696671724, "keep": 0.16775397397577763, "leak": 0.2790686001535505, "replacement": 0.6408039107918739}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 4.388149738311768, "loss_scale": 4.0, "retries": 0, "seconds": 173.5, "max_memory_gb": 25.46, "pair": 0.46942292526364326, "sign": 0.4903000555932522, "keep": 0.3413828834891319, "leak": 1.1521280948072672, "replacement": 1.1732963994145393}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 3.985069513320923, "loss_scale": 4.0, "retries": 0, "seconds": 174.8, "max_memory_gb": 25.46, "pair": 0.38184161856770515, "sign": 0.43128827027976513, "keep": 0.12483493145555258, "leak": 1.0992795377969742, "replacement": 0.9347230941057205}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 2.8613638877868652, "loss_scale": 4.0, "retries": 0, "seconds": 176.1, "max_memory_gb": 25.46, "pair": 0.4740369878709316, "sign": 0.461089882068336, "keep": 0.34911734890192747, "leak": 1.4806244596838951, "replacement": 1.249296709895134}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 1.5579191446304321, "loss_scale": 4.0, "retries": 0, "seconds": 177.3, "max_memory_gb": 25.46, "pair": 0.4486776925623417, "sign": 0.3732850197702646, "keep": 0.21889716410078108, "leak": 0.48155779218359385, "replacement": 0.8651582449674606}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 2.0173823833465576, "loss_scale": 4.0, "retries": 0, "seconds": 178.6, "max_memory_gb": 25.46, "pair": 0.8205609768629074, "sign": 1.0052346363663673, "keep": 0.22511075425427407, "leak": 0.29424277879297733, "replacement": 1.5092943608760834}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 2.5234270095825195, "loss_scale": 4.0, "retries": 0, "seconds": 179.8, "max_memory_gb": 25.46, "pair": 0.4168412387371063, "sign": 0.3283370304852724, "keep": 0.22046596836298704, "leak": 1.8431814946234226, "replacement": 1.1520381569862366}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 2.3479905128479004, "loss_scale": 4.0, "retries": 0, "seconds": 181.1, "max_memory_gb": 25.46, "pair": 0.5509093180298805, "sign": 0.308135979808867, "keep": 0.05391780589707196, "leak": 1.2963395342230797, "replacement": 1.0560210943222046}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.7615699768066406, "loss_scale": 4.0, "retries": 0, "seconds": 182.3, "max_memory_gb": 25.46, "pair": 0.5097184386104345, "sign": 0.533819055184722, "keep": 0.19207261968404055, "leak": 1.2194828724022955, "replacement": 1.1775349825620651}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.201246738433838, "loss_scale": 4.0, "retries": 0, "seconds": 183.5, "max_memory_gb": 25.46, "pair": 0.5128363780677319, "sign": 0.4910299889743328, "keep": 0.4344970490783453, "leak": 1.0968560529872775, "replacement": 1.2498138546943665}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.9755903482437134, "loss_scale": 4.0, "retries": 0, "seconds": 184.8, "max_memory_gb": 25.46, "pair": 0.4563001021742821, "sign": 0.25378502160310745, "keep": 0.03966152551583946, "leak": 0.3883260702714324, "replacement": 0.7001048773527145}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 2.1312286853790283, "loss_scale": 4.0, "retries": 0, "seconds": 186.0, "max_memory_gb": 25.46, "pair": 0.4796527735888958, "sign": 0.4156348742544651, "keep": 0.14905225485563278, "leak": 0.25783914187923074, "replacement": 0.8264561221003532}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 2.9642317295074463, "loss_scale": 4.0, "retries": 0, "seconds": 187.3, "max_memory_gb": 25.46, "pair": 0.38885143399238586, "sign": 0.3210312481969595, "keep": 0.4355638027191162, "leak": 2.4601367274299264, "replacement": 1.382183164358139}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 2.222531318664551, "loss_scale": 4.0, "retries": 0, "seconds": 188.6, "max_memory_gb": 25.46, "pair": 0.35055746138095856, "sign": 0.3349364213645458, "keep": 0.28228708100505173, "leak": 0.4912984073162079, "replacement": 0.7819938063621521}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 4.177172660827637, "loss_scale": 4.0, "retries": 0, "seconds": 189.8, "max_memory_gb": 25.46, "pair": 0.24829662963747978, "sign": 0.1746234307065606, "keep": 0.16940925363451242, "leak": 2.5501937568187714, "replacement": 1.057861402630806}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 2.4844706058502197, "loss_scale": 4.0, "retries": 0, "seconds": 191.1, "max_memory_gb": 25.46, "pair": 0.31118167005479336, "sign": 0.2217374425381422, "keep": 0.07084681233391166, "leak": 1.7716152593493462, "replacement": 0.9003776162862778}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.7736825942993164, "loss_scale": 4.0, "retries": 0, "seconds": 192.4, "max_memory_gb": 25.46, "pair": 0.24706029146909714, "sign": 0.16507769748568535, "keep": 0.7344416379928589, "leak": 1.0249793529510498, "replacement": 0.953064814209938}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.5654921531677246, "loss_scale": 4.0, "retries": 0, "seconds": 193.6, "max_memory_gb": 25.46, "pair": 0.41394011676311493, "sign": 0.38850899413228035, "keep": 0.9478186359629035, "leak": 2.6310522258281708, "replacement": 1.739867016673088}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.6969847679138184, "loss_scale": 4.0, "retries": 0, "seconds": 194.9, "max_memory_gb": 25.46, "pair": 0.5369509980082512, "sign": 0.49136912263929844, "keep": 0.2311325641348958, "leak": 0.4984515003161505, "replacement": 1.022814691066742}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 2.6161391735076904, "loss_scale": 4.0, "retries": 0, "seconds": 196.1, "max_memory_gb": 25.46, "pair": 0.7325547859072685, "sign": 1.1581815108656883, "keep": 0.2516860947944224, "leak": 0.2826455545146018, "replacement": 1.5081500113010406}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 2.8419082164764404, "loss_scale": 4.0, "retries": 0, "seconds": 197.3, "max_memory_gb": 25.46, "pair": 0.5368076004087925, "sign": 0.6181028001010418, "keep": 0.47278746124356985, "leak": 0.45179577078670263, "replacement": 1.1952016651630402}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 2.7953875064849854, "loss_scale": 4.0, "retries": 0, "seconds": 198.6, "max_memory_gb": 25.46, "pair": 0.26895715296268463, "sign": 0.15528463013470173, "keep": 0.08328555570915341, "leak": 2.312748968601227, "replacement": 0.9664295017719269}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 5.746050834655762, "loss_scale": 4.0, "retries": 0, "seconds": 199.9, "max_memory_gb": 25.46, "pair": 0.5893913693726063, "sign": 0.3461974896490574, "keep": 0.24690166488289833, "leak": 1.4574695713232018, "replacement": 1.2503083497285843}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 2.289844036102295, "loss_scale": 4.0, "retries": 0, "seconds": 201.1, "max_memory_gb": 25.46, "pair": 0.5983962677419186, "sign": 0.5710284672677517, "keep": 0.44362768973223865, "leak": 0.6915214546024799, "replacement": 1.2786047458648682}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 3.7500038146972656, "loss_scale": 4.0, "retries": 0, "seconds": 202.3, "max_memory_gb": 25.46, "pair": 0.5609228974208236, "sign": 0.5725439479574561, "keep": 0.15098555851727724, "leak": 0.3828357843449339, "replacement": 1.0183965563774109}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 1.7499603033065796, "loss_scale": 4.0, "retries": 0, "seconds": 203.5, "max_memory_gb": 25.46, "pair": 0.4184737950563431, "sign": 0.2883016923442483, "keep": 0.08751443587243557, "leak": 0.7864803075790405, "replacement": 0.8030019551515579}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 2.99115252494812, "loss_scale": 4.0, "retries": 0, "seconds": 204.8, "max_memory_gb": 25.46, "pair": 0.607595793902874, "sign": 0.8016421748325229, "keep": 0.42830174416303635, "leak": 0.7850175115504499, "replacement": 1.418822169303894}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 2.8278400897979736, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 25.46, "pair": 0.42553336173295975, "sign": 0.5044813193380833, "keep": 0.15898949932307005, "leak": 1.477659584954381, "replacement": 1.1266836822032928}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 3.1911051273345947, "loss_scale": 4.0, "retries": 0, "seconds": 207.3, "max_memory_gb": 25.46, "pair": 0.42680639028549194, "sign": 0.4466741681098938, "keep": 0.21984338015317917, "leak": 1.7488555266572803, "replacement": 1.1972790360450745}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 3.2994112968444824, "loss_scale": 4.0, "retries": 0, "seconds": 208.6, "max_memory_gb": 25.46, "pair": 0.34403506200760603, "sign": 0.3797034826129675, "keep": 0.5602949783205986, "leak": 1.4827295988798141, "replacement": 1.1847166828811169}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 3.265748977661133, "loss_scale": 4.0, "retries": 0, "seconds": 209.8, "max_memory_gb": 25.46, "pair": 0.4036215730011463, "sign": 0.36919867992401123, "keep": 0.14714948926120996, "leak": 1.443340688943863, "replacement": 1.022630825638771}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 4.858243465423584, "loss_scale": 4.0, "retries": 0, "seconds": 211.1, "max_memory_gb": 25.46, "pair": 0.33295080438256264, "sign": 0.3834956372156739, "keep": 0.7420753035694361, "leak": 2.1736421901732683, "replacement": 1.4391468912363052}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 3.1989831924438477, "loss_scale": 4.0, "retries": 0, "seconds": 212.3, "max_memory_gb": 25.46, "pair": 0.27849696204066277, "sign": 0.08082049060612917, "keep": 0.08956725290045142, "leak": 1.128506850451231, "replacement": 0.6458175331354141}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 6.047194957733154, "loss_scale": 4.0, "retries": 0, "seconds": 213.6, "max_memory_gb": 25.46, "pair": 0.46280201338231564, "sign": 0.3792658010497689, "keep": 0.11593468533828855, "leak": 1.4974404694512486, "replacement": 1.0847623720765114}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 4.067115306854248, "loss_scale": 4.0, "retries": 0, "seconds": 214.9, "max_memory_gb": 25.46, "pair": 0.22961031086742878, "sign": 0.07339124567806721, "keep": 0.18476027809083462, "leak": 4.698603123426437, "replacement": 1.5333368182182312}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 4.406335830688477, "loss_scale": 4.0, "retries": 0, "seconds": 216.1, "max_memory_gb": 25.46, "pair": 0.24515990540385246, "sign": 0.3646055683493614, "keep": 0.3665420413017273, "leak": 2.2665172815322876, "replacement": 1.1773630380630493}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 3.197601079940796, "loss_scale": 4.0, "retries": 0, "seconds": 217.3, "max_memory_gb": 25.46, "pair": 0.40308911353349686, "sign": 0.25224436074495316, "keep": 0.20566792460158467, "leak": 1.2210478521883488, "replacement": 0.9373071789741516}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.4480245113372803, "loss_scale": 4.0, "retries": 0, "seconds": 218.6, "max_memory_gb": 25.46, "pair": 0.4349781796336174, "sign": 0.48971639946103096, "keep": 0.23346683010458946, "leak": 1.8054782776162028, "replacement": 1.2479393482208252}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 2.8550403118133545, "loss_scale": 4.0, "retries": 0, "seconds": 219.9, "max_memory_gb": 25.46, "pair": 0.34161168709397316, "sign": 0.1330677168443799, "keep": 0.17887504398822784, "leak": 2.1163187995553017, "replacement": 1.0266627594828606}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 5.2962727546691895, "loss_scale": 4.0, "retries": 0, "seconds": 221.1, "max_memory_gb": 25.46, "pair": 0.4670259766280651, "sign": 0.34899470768868923, "keep": 0.40179675654508173, "leak": 0.7412310816143872, "replacement": 1.027729481458664}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.5846216678619385, "loss_scale": 4.0, "retries": 0, "seconds": 222.3, "max_memory_gb": 25.46, "pair": 0.5378578901290894, "sign": 0.5158015210181475, "keep": 0.21481616771779954, "leak": 1.6244707619771361, "replacement": 1.309284433722496}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 1.7752580642700195, "loss_scale": 4.0, "retries": 0, "seconds": 223.6, "max_memory_gb": 25.46, "pair": 0.7076135724782944, "sign": 0.7625620849430561, "keep": 0.5833791820332408, "leak": 0.7334587769000791, "replacement": 1.5639488846063614}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 9.139033317565918, "loss_scale": 4.0, "retries": 0, "seconds": 224.8, "max_memory_gb": 25.46, "pair": 0.7697168216109276, "sign": 0.6513388268649578, "keep": 0.9597962535917759, "leak": 0.5737330745905638, "replacement": 1.7187176197767258}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 3.3677210807800293, "loss_scale": 4.0, "retries": 0, "seconds": 226.1, "max_memory_gb": 25.46, "pair": 0.32317934557795525, "sign": 0.13065186375752091, "keep": 0.17524266056716442, "leak": 2.573838584125042, "replacement": 1.1195861920714378}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 4.35480260848999, "loss_scale": 4.0, "retries": 0, "seconds": 227.3, "max_memory_gb": 25.46, "pair": 0.5771195814013481, "sign": 0.595903730019927, "keep": 0.5793726111296564, "leak": 1.0153169557452202, "replacement": 1.4185869842767715}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.486374855041504, "loss_scale": 4.0, "retries": 0, "seconds": 228.6, "max_memory_gb": 25.46, "pair": 0.2246411330997944, "sign": 0.1201649010181427, "keep": 0.04718940728344023, "leak": 1.4044121205806732, "replacement": 0.6594213172793388}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 5.9837164878845215, "loss_scale": 4.0, "retries": 0, "seconds": 229.8, "max_memory_gb": 25.46, "pair": 0.4279569052159786, "sign": 0.3219478251412511, "keep": 0.16699762642383575, "leak": 0.8771565286442637, "replacement": 0.891718789935112}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 4.9045610427856445, "loss_scale": 4.0, "retries": 0, "seconds": 231.1, "max_memory_gb": 25.46, "pair": 0.40783168002963066, "sign": 0.24930999474599957, "keep": 0.08794566616415977, "leak": 2.937878548298613, "replacement": 1.310929149389267}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 2.890777111053467, "loss_scale": 4.0, "retries": 0, "seconds": 232.4, "max_memory_gb": 25.46, "pair": 0.34711456298828125, "sign": 0.2349533447995782, "keep": 0.18233217054512352, "leak": 0.7898427955879015, "replacement": 0.753218024969101}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 4.480012893676758, "loss_scale": 4.0, "retries": 0, "seconds": 233.7, "max_memory_gb": 25.46, "pair": 0.3938850536942482, "sign": 0.07969141937792301, "keep": 0.14182001445442438, "leak": 1.5148379541933537, "replacement": 0.8833502233028412}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 2.511979341506958, "loss_scale": 4.0, "retries": 0, "seconds": 234.9, "max_memory_gb": 25.46, "pair": 0.494708389043808, "sign": 0.5234521813690662, "keep": 0.23729801713488996, "leak": 0.5033313997555524, "replacement": 1.0009163320064545}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 3.026123046875, "loss_scale": 4.0, "retries": 0, "seconds": 236.1, "max_memory_gb": 25.46, "pair": 0.29194705188274384, "sign": 0.19461644999682903, "keep": 0.11038627428933978, "leak": 1.0983963900944218, "replacement": 0.7190474942326546}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 2.428586959838867, "loss_scale": 4.0, "retries": 0, "seconds": 237.4, "max_memory_gb": 25.46, "pair": 0.2529454827308655, "sign": 0.1895979642868042, "keep": 0.07129285950213671, "leak": 1.503622144460678, "replacement": 0.7592964470386505}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 4.292140960693359, "loss_scale": 4.0, "retries": 0, "seconds": 238.5, "max_memory_gb": 25.46, "pair": 0.5811455994844437, "sign": 0.6061936356127262, "keep": 0.719660556409508, "leak": 0.6377651728689671, "replacement": 1.403513953089714}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 1.464437484741211, "loss_scale": 4.0, "retries": 0, "seconds": 239.8, "max_memory_gb": 25.46, "pair": 0.5548438653349876, "sign": 0.46600330248475075, "keep": 0.3488933062180877, "leak": 0.777896239640711, "replacement": 1.1567662358283997}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 3.5529463291168213, "loss_scale": 4.0, "retries": 0, "seconds": 241.0, "max_memory_gb": 25.46, "pair": 0.3395977020263672, "sign": 0.16621003486216068, "keep": 0.15530488383956254, "leak": 0.6144484728574753, "replacement": 0.6539672911167145}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 6.976126670837402, "loss_scale": 4.0, "retries": 0, "seconds": 242.2, "max_memory_gb": 25.46, "pair": 0.66429303586483, "sign": 0.8430791608989239, "keep": 1.1328237345442176, "leak": 1.2314701303839684, "replacement": 1.960112065076828}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 4.376607894897461, "loss_scale": 4.0, "retries": 0, "seconds": 243.5, "max_memory_gb": 25.46, "pair": 0.5094339996576309, "sign": 0.3614430259913206, "keep": 0.3663168828934431, "leak": 1.6361504904925823, "replacement": 1.2823516130447388}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 1.7838901281356812, "loss_scale": 4.0, "retries": 0, "seconds": 244.8, "max_memory_gb": 25.46, "pair": 0.21669001877307892, "sign": 0.09820597106590867, "keep": 0.05359869636595249, "leak": 0.9047048091888428, "replacement": 0.5187685638666153}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 3.6181328296661377, "loss_scale": 4.0, "retries": 0, "seconds": 246.1, "max_memory_gb": 25.46, "pair": 0.5212863236665726, "sign": 0.5208661546930671, "keep": 0.167018735781312, "leak": 0.5747438815305941, "replacement": 1.008914738893509}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 3.269237995147705, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 25.46, "pair": 0.4718015193939209, "sign": 0.5385235100984573, "keep": 0.24149926099926233, "leak": 0.9230093043299148, "replacement": 1.0925652235746384}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.602322578430176, "loss_scale": 4.0, "retries": 0, "seconds": 248.5, "max_memory_gb": 25.46, "pair": 0.24319535866379738, "sign": 0.11604934185743332, "keep": 0.2808944280259311, "leak": 2.1683104634284973, "replacement": 0.9837448596954346}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 2.131117105484009, "loss_scale": 4.0, "retries": 0, "seconds": 249.8, "max_memory_gb": 25.46, "pair": 0.4885899983346462, "sign": 0.5914807450026274, "keep": 0.4589870891068131, "leak": 0.8027569595724344, "replacement": 1.2145131677389145}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 2.9627301692962646, "loss_scale": 4.0, "retries": 0, "seconds": 251.0, "max_memory_gb": 25.46, "pair": 0.461422435939312, "sign": 0.5882171117700636, "keep": 0.22841255087405443, "leak": 0.6803900226950645, "replacement": 1.0398347824811935}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 2.9712913036346436, "loss_scale": 4.0, "retries": 0, "seconds": 252.3, "max_memory_gb": 25.46, "pair": 0.3942568711936474, "sign": 0.31685449555516243, "keep": 0.010045496514067054, "leak": 0.7566477581858635, "replacement": 0.746868833899498}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 5.665511131286621, "loss_scale": 4.0, "retries": 0, "seconds": 253.5, "max_memory_gb": 25.46, "pair": 0.5398470908403397, "sign": 0.4942805012688041, "keep": 0.31347978115081787, "leak": 1.3382736005587503, "replacement": 1.2782956212759018}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 4.354238033294678, "loss_scale": 4.0, "retries": 0, "seconds": 254.8, "max_memory_gb": 25.46, "pair": 0.5877532586455345, "sign": 0.4957917174324393, "keep": 0.2524020420387387, "leak": 1.443275124183856, "replacement": 1.3226689249277115}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 3.3896684646606445, "loss_scale": 4.0, "retries": 0, "seconds": 256.0, "max_memory_gb": 25.46, "pair": 0.24268881604075432, "sign": 0.29186402074992657, "keep": 0.13848316960502416, "leak": 1.6790673360228539, "replacement": 0.8776292502880096}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 3.2893807888031006, "loss_scale": 4.0, "retries": 0, "seconds": 257.3, "max_memory_gb": 25.46, "pair": 0.4310937859117985, "sign": 0.38653961196541786, "keep": 0.1445906877052039, "leak": 2.2139586261473596, "replacement": 1.250148594379425}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 8.163008689880371, "loss_scale": 4.0, "retries": 0, "seconds": 258.6, "max_memory_gb": 25.46, "pair": 0.916423287242651, "sign": 1.0242241099476814, "keep": 1.5259388585109264, "leak": 0.4285639689769596, "replacement": 2.2986457347869873}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 3.006709575653076, "loss_scale": 4.0, "retries": 0, "seconds": 259.8, "max_memory_gb": 25.46, "pair": 0.30658602342009544, "sign": 0.08180185593664646, "keep": 0.05384383932687342, "leak": 1.3662035167217255, "replacement": 0.7159597426652908}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 2.3921263217926025, "loss_scale": 4.0, "retries": 0, "seconds": 261.1, "max_memory_gb": 25.46, "pair": 0.5254057012498379, "sign": 0.5717350244522095, "keep": 0.32579625584185123, "leak": 0.9928662093589082, "replacement": 1.2223879173398018}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 2.9042062759399414, "loss_scale": 4.0, "retries": 0, "seconds": 262.4, "max_memory_gb": 25.46, "pair": 0.4056093357503414, "sign": 0.4698976967483759, "keep": 0.5617730026133358, "leak": 2.184362666681409, "replacement": 1.4675354063510895}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 3.846318483352661, "loss_scale": 4.0, "retries": 0, "seconds": 263.6, "max_memory_gb": 25.46, "pair": 0.5796868009492755, "sign": 0.575993500649929, "keep": 0.3475453592836857, "leak": 0.7363440985791385, "replacement": 1.2255422472953796}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 5.2923903465271, "loss_scale": 4.0, "retries": 0, "seconds": 264.9, "max_memory_gb": 25.46, "pair": 0.47510088235139847, "sign": 0.6186340544372797, "keep": 0.5624515898525715, "leak": 0.7542677465826273, "replacement": 1.2542106285691261}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 3.368168830871582, "loss_scale": 4.0, "retries": 0, "seconds": 266.1, "max_memory_gb": 25.46, "pair": 0.3959057405591011, "sign": 0.45866669341921806, "keep": 0.8619070504792035, "leak": 1.7317625656432938, "replacement": 1.489133283495903}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 2.1488733291625977, "loss_scale": 4.0, "retries": 0, "seconds": 267.4, "max_memory_gb": 25.46, "pair": 0.5831096023321152, "sign": 0.5880156904459, "keep": 0.49655329261440784, "leak": 0.31615803969907574, "replacement": 1.2044335901737213}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 5.691702842712402, "loss_scale": 4.0, "retries": 0, "seconds": 268.7, "max_memory_gb": 25.46, "pair": 0.797217957675457, "sign": 0.9799352064728737, "keep": 1.5628064904594794, "leak": 0.5009547564695822, "replacement": 2.1938274651765823}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 2.019329071044922, "loss_scale": 4.0, "retries": 0, "seconds": 269.9, "max_memory_gb": 25.46, "pair": 0.2880808524787426, "sign": 0.27621906250715256, "keep": 0.2891467453446239, "leak": 0.7916981466114521, "replacement": 0.768688291311264}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 4.191067695617676, "loss_scale": 4.0, "retries": 0, "seconds": 271.2, "max_memory_gb": 25.46, "pair": 0.22175890859216452, "sign": 0.19519029837101698, "keep": 0.05210514087229967, "leak": 1.7109414637088776, "replacement": 0.7731419950723648}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 3.910053014755249, "loss_scale": 4.0, "retries": 0, "seconds": 272.4, "max_memory_gb": 25.46, "pair": 0.819578543305397, "sign": 1.0249881595373154, "keep": 0.4146951816510409, "leak": 0.2093090182170272, "replacement": 1.5917474180459976}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 3.599221706390381, "loss_scale": 4.0, "retries": 0, "seconds": 273.6, "max_memory_gb": 25.46, "pair": 0.43029608577489853, "sign": 0.5020546615123749, "keep": 0.215671754674986, "leak": 0.34997617607950815, "replacement": 0.876653328537941}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 2.908499240875244, "loss_scale": 4.0, "retries": 0, "seconds": 274.9, "max_memory_gb": 25.46, "pair": 0.3762265332043171, "sign": 0.4644891023635864, "keep": 0.2556091103469953, "leak": 0.6739362329244614, "replacement": 0.9047597050666809}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 2.0259828567504883, "loss_scale": 4.0, "retries": 0, "seconds": 276.1, "max_memory_gb": 25.46, "pair": 0.3161818999797106, "sign": 0.26827770844101906, "keep": 0.15471909567713737, "leak": 0.9662827159045264, "replacement": 0.7692509815096855}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 5.1031718254089355, "loss_scale": 4.0, "retries": 0, "seconds": 277.4, "max_memory_gb": 25.46, "pair": 0.4185567796230316, "sign": 0.2333855489268899, "keep": 0.4588704891502857, "leak": 1.3874019496142864, "replacement": 1.1115352511405945}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 5.512309551239014, "loss_scale": 4.0, "retries": 0, "seconds": 278.6, "max_memory_gb": 25.46, "pair": 0.5990847945213318, "sign": 0.6775719150900841, "keep": 0.35307861265027896, "leak": 0.1386493967147544, "replacement": 1.1490723937749863}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 2.6065993309020996, "loss_scale": 4.0, "retries": 0, "seconds": 279.9, "max_memory_gb": 25.46, "pair": 0.3451095260679722, "sign": 0.3672016803175211, "keep": 0.11177377594867721, "leak": 1.2445698622614145, "replacement": 0.8957397639751434}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 6.24539852142334, "loss_scale": 4.0, "retries": 0, "seconds": 281.1, "max_memory_gb": 25.46, "pair": 0.5253870189189911, "sign": 0.5790087776258588, "keep": 0.12594028748571873, "leak": 1.0027212611821597, "replacement": 1.1285418719053268}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 3.592371702194214, "loss_scale": 4.0, "retries": 0, "seconds": 282.4, "max_memory_gb": 25.46, "pair": 0.291248120367527, "sign": 0.2712911982089281, "keep": 0.23545183520764112, "leak": 0.741830050945282, "replacement": 0.7300771772861481}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.6948530673980713, "loss_scale": 4.0, "retries": 0, "seconds": 283.6, "max_memory_gb": 25.46, "pair": 0.32489901781082153, "sign": 0.1832682890817523, "keep": 0.031348977936431766, "leak": 1.1992178782820702, "replacement": 0.7320121228694916}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 2.8495707511901855, "loss_scale": 4.0, "retries": 0, "seconds": 284.8, "max_memory_gb": 25.46, "pair": 0.23539266735315323, "sign": 0.3150556869804859, "keep": 0.0341907748952508, "leak": 0.9110996127128601, "replacement": 0.6377907991409302}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 3.533703088760376, "loss_scale": 4.0, "retries": 0, "seconds": 286.1, "max_memory_gb": 25.46, "pair": 0.5170104950666428, "sign": 0.6505056009627879, "keep": 0.5254473662935197, "leak": 0.6843016818165779, "replacement": 1.2760624289512634}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 4.474168300628662, "loss_scale": 4.0, "retries": 0, "seconds": 287.3, "max_memory_gb": 25.46, "pair": 0.6662367433309555, "sign": 0.7391752190887928, "keep": 0.6109314583591186, "leak": 0.5581058524549007, "replacement": 1.4808165729045868}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 3.248711585998535, "loss_scale": 4.0, "retries": 0, "seconds": 288.5, "max_memory_gb": 25.46, "pair": 0.6525793224573135, "sign": 0.7595920749008656, "keep": 0.5211411821655929, "leak": 0.26395178213715553, "replacement": 1.358933910727501}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 3.999918222427368, "loss_scale": 4.0, "retries": 0, "seconds": 289.8, "max_memory_gb": 25.46, "pair": 0.25495360791683197, "sign": 0.37712755613029003, "keep": 0.05529100447893143, "leak": 1.2246997775509953, "replacement": 0.777337808161974}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 4.461903095245361, "loss_scale": 4.0, "retries": 0, "seconds": 291.0, "max_memory_gb": 25.46, "pair": 0.3127659484744072, "sign": 0.18260851688683033, "keep": 0.37313452549278736, "leak": 0.9635220115897027, "replacement": 0.8315179646015167}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 3.5588700771331787, "loss_scale": 4.0, "retries": 0, "seconds": 292.2, "max_memory_gb": 25.46, "pair": 0.3481983672827482, "sign": 0.4046263061463833, "keep": 0.28330216091126204, "leak": 1.7064304761588573, "replacement": 1.1187702119350433}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 4.0737457275390625, "loss_scale": 4.0, "retries": 0, "seconds": 293.4, "max_memory_gb": 25.46, "pair": 0.6163907200098038, "sign": 0.6367105841636658, "keep": 0.7472355104982853, "leak": 0.6244151368737221, "replacement": 1.4644675850868225}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 3.471719980239868, "loss_scale": 4.0, "retries": 0, "seconds": 294.7, "max_memory_gb": 25.46, "pair": 0.5516354739665985, "sign": 0.4808650575578213, "keep": 0.22875273053068668, "leak": 0.29782157577574253, "replacement": 0.980899766087532}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.4880144596099854, "loss_scale": 4.0, "retries": 0, "seconds": 295.9, "max_memory_gb": 25.46, "pair": 0.49946022033691406, "sign": 0.42823489010334015, "keep": 0.40236456104321405, "leak": 0.42543667438440025, "replacement": 1.0211191028356552}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 3.3988797664642334, "loss_scale": 4.0, "retries": 0, "seconds": 297.1, "max_memory_gb": 25.46, "pair": 0.2337002493441105, "sign": 0.14293285319581628, "keep": 0.08649132354184985, "leak": 1.3277395814657211, "replacement": 0.6803472116589546}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 4.999448776245117, "loss_scale": 4.0, "retries": 0, "seconds": 298.4, "max_memory_gb": 25.46, "pair": 0.4702496323734522, "sign": 0.5593414753675461, "keep": 0.3281364035792649, "leak": 3.138783954549581, "replacement": 1.698684573173523}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.168520450592041, "loss_scale": 4.0, "retries": 0, "seconds": 299.6, "max_memory_gb": 25.46, "pair": 0.33673450350761414, "sign": 0.27884782291948795, "keep": 0.08963434206089005, "leak": 0.5305946189910173, "replacement": 0.6536242514848709}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 3.5074269771575928, "loss_scale": 4.0, "retries": 0, "seconds": 300.9, "max_memory_gb": 25.46, "pair": 0.4591188281774521, "sign": 0.5850820168852806, "keep": 0.17396725434809923, "leak": 0.3788254917599261, "replacement": 0.9333498477935791}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 3.0175726413726807, "loss_scale": 4.0, "retries": 0, "seconds": 302.1, "max_memory_gb": 25.46, "pair": 0.47848493978381157, "sign": 0.5431498307734728, "keep": 0.20613323780708015, "leak": 0.5833130981773138, "replacement": 0.9989547654986382}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 3.397008180618286, "loss_scale": 4.0, "retries": 0, "seconds": 303.4, "max_memory_gb": 25.46, "pair": 0.29430417343974113, "sign": 0.21743174083530903, "keep": 0.05286048143170774, "leak": 2.7752778017893434, "replacement": 1.123269721865654}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 8.64281177520752, "loss_scale": 4.0, "retries": 0, "seconds": 304.7, "max_memory_gb": 25.46, "pair": 0.750302642583847, "sign": 0.9336595796048641, "keep": 1.0598839758895338, "leak": 0.6449631992727518, "replacement": 1.9083151742815971}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 2.7989847660064697, "loss_scale": 4.0, "retries": 0, "seconds": 305.9, "max_memory_gb": 25.46, "pair": 0.843747541308403, "sign": 0.982910368591547, "keep": 0.6971332840621471, "leak": 0.27442159596830606, "replacement": 1.752374842762947}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 3.0595507621765137, "loss_scale": 4.0, "retries": 0, "seconds": 307.2, "max_memory_gb": 25.46, "pair": 0.21308243460953236, "sign": 0.19550148909911513, "keep": 0.06594179396051913, "leak": 1.3926422595977783, "replacement": 0.6919646263122559}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 3.9678280353546143, "loss_scale": 4.0, "retries": 0, "seconds": 308.4, "max_memory_gb": 25.46, "pair": 0.3593708798289299, "sign": 0.37541887909173965, "keep": 0.13068519285297953, "leak": 0.5833463668823242, "replacement": 0.758259505033493}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.9576008319854736, "loss_scale": 4.0, "retries": 0, "seconds": 309.6, "max_memory_gb": 25.46, "pair": 0.2347452249377966, "sign": 0.12367797503247857, "keep": 0.15539633063599467, "leak": 0.6197067876928486, "replacement": 0.5292090699076653}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 5.810108661651611, "loss_scale": 4.0, "retries": 0, "seconds": 310.9, "max_memory_gb": 25.46, "pair": 0.514445599168539, "sign": 0.7363831400871277, "keep": 0.10054096020758152, "leak": 0.8453920630854554, "replacement": 1.144255667924881}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 4.1612982749938965, "loss_scale": 4.0, "retries": 0, "seconds": 312.1, "max_memory_gb": 25.46, "pair": 0.19701837562024593, "sign": 0.20885483548045158, "keep": 0.05160984513349831, "leak": 2.080652803182602, "replacement": 0.8474138975143433}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 4.091100692749023, "loss_scale": 4.0, "retries": 0, "seconds": 313.4, "max_memory_gb": 25.46, "pair": 0.394250713288784, "sign": 0.42352492548525333, "keep": 0.1741736235562712, "leak": 1.8389811927918345, "replacement": 1.1528452932834625}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 3.0013387203216553, "loss_scale": 4.0, "retries": 0, "seconds": 314.7, "max_memory_gb": 25.46, "pair": 0.2328979466110468, "sign": 0.1438652933575213, "keep": 0.027736546471714973, "leak": 1.19861201941967, "replacement": 0.6183518916368484}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 5.875640869140625, "loss_scale": 4.0, "retries": 0, "seconds": 315.9, "max_memory_gb": 25.46, "pair": 0.2706473069265485, "sign": 0.3127960809506476, "keep": 0.22420837508980185, "leak": 2.0217456333339214, "replacement": 1.044585943222046}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 3.9207334518432617, "loss_scale": 4.0, "retries": 0, "seconds": 317.2, "max_memory_gb": 25.46, "pair": 0.23765485733747482, "sign": 0.2815672564320266, "keep": 0.3661274593323469, "leak": 1.4598536118865013, "replacement": 0.9264655858278275}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 3.5937180519104004, "loss_scale": 4.0, "retries": 0, "seconds": 318.4, "max_memory_gb": 25.46, "pair": 0.5113693885505199, "sign": 0.6774391829967499, "keep": 0.1508275584783405, "leak": 0.5322482399642467, "replacement": 1.0585647970438004}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 3.4739127159118652, "loss_scale": 4.0, "retries": 0, "seconds": 319.6, "max_memory_gb": 25.46, "pair": 0.4718665685504675, "sign": 0.5531508848071098, "keep": 0.11783921159803867, "leak": 0.9785216990858316, "replacement": 1.051992043852806}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 6.3869147300720215, "loss_scale": 4.0, "retries": 0, "seconds": 320.9, "max_memory_gb": 25.46, "pair": 0.4614454321563244, "sign": 0.48896409571170807, "keep": 0.19187401130329818, "leak": 0.8057721722871065, "replacement": 1.0033074915409088}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 2.526806354522705, "loss_scale": 4.0, "retries": 0, "seconds": 322.1, "max_memory_gb": 25.46, "pair": 0.19462691992521286, "sign": 0.13699612393975258, "keep": 0.22539120400324464, "leak": 1.326066717505455, "replacement": 0.7073372602462769}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.5022032260894775, "loss_scale": 4.0, "retries": 0, "seconds": 323.4, "max_memory_gb": 25.46, "pair": 0.20207133516669273, "sign": 0.1982279196381569, "keep": 0.00702454021666199, "leak": 1.2285992950201035, "replacement": 0.611847385764122}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 4.4141154289245605, "loss_scale": 4.0, "retries": 0, "seconds": 324.6, "max_memory_gb": 25.46, "pair": 0.2427549483254552, "sign": 0.31728839315474033, "keep": 0.18138410663232207, "leak": 1.1503678411245346, "replacement": 0.7796831727027893}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.7246758937835693, "loss_scale": 4.0, "retries": 0, "seconds": 325.9, "max_memory_gb": 25.46, "pair": 0.36357446759939194, "sign": 0.33197919093072414, "keep": 0.24983021710067987, "leak": 1.0666312073008157, "replacement": 0.9211369827389717}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 6.46436882019043, "loss_scale": 4.0, "retries": 0, "seconds": 327.1, "max_memory_gb": 25.46, "pair": 0.5189238265156746, "sign": 0.5630042105913162, "keep": 0.5738088940270245, "leak": 0.332249965518713, "replacement": 1.1703928858041763}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 4.179322719573975, "loss_scale": 4.0, "retries": 0, "seconds": 328.4, "max_memory_gb": 25.46, "pair": 0.2875550538301468, "sign": 0.32601368241012096, "keep": 0.3412892948836088, "leak": 1.9287355579435825, "replacement": 1.1033904701471329}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 6.219083786010742, "loss_scale": 4.0, "retries": 0, "seconds": 329.6, "max_memory_gb": 25.46, "pair": 0.22732432559132576, "sign": 0.12345933960750699, "keep": 0.49692903959657997, "leak": 2.874553419649601, "replacement": 1.2561568915843964}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 4.03037166595459, "loss_scale": 4.0, "retries": 0, "seconds": 330.9, "max_memory_gb": 25.46, "pair": 0.2977343462407589, "sign": 0.303116662427783, "keep": 0.09938588715158403, "leak": 1.7623401042073965, "replacement": 0.9395706430077553}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 7.072060585021973, "loss_scale": 4.0, "retries": 0, "seconds": 332.1, "max_memory_gb": 25.46, "pair": 0.5382793992757797, "sign": 0.5773448143154383, "keep": 0.9253366943448782, "leak": 0.6986032128334045, "replacement": 1.4642709828913212}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.1200759410858154, "loss_scale": 4.0, "retries": 0, "seconds": 333.4, "max_memory_gb": 25.46, "pair": 0.42272601649165154, "sign": 0.442196317948401, "keep": 0.27763570472598076, "leak": 0.6345375152304769, "replacement": 0.9412763640284538}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 1.8256710767745972, "loss_scale": 4.0, "retries": 0, "seconds": 334.6, "max_memory_gb": 25.46, "pair": 0.5848440453410149, "sign": 0.7155640497803688, "keep": 0.3624502863967791, "leak": 0.41155910259112716, "replacement": 1.22674098610878}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.680541515350342, "loss_scale": 4.0, "retries": 0, "seconds": 335.9, "max_memory_gb": 25.46, "pair": 0.3155960328876972, "sign": 0.3981463648378849, "keep": 1.2743194608483464, "leak": 0.7338789771310985, "replacement": 1.3352986723184586}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 6.1489715576171875, "loss_scale": 4.0, "retries": 0, "seconds": 337.1, "max_memory_gb": 25.46, "pair": 0.31318848114460707, "sign": 0.4359766086563468, "keep": 0.5582205569371581, "leak": 1.0288330614566803, "replacement": 1.067495383322239}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 4.313681125640869, "loss_scale": 4.0, "retries": 0, "seconds": 338.4, "max_memory_gb": 25.46, "pair": 0.21989681199193, "sign": 0.22942777164280415, "keep": 0.08152451273053885, "leak": 2.3914298340678215, "replacement": 0.9732304215431213}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 2.6416845321655273, "loss_scale": 4.0, "retries": 0, "seconds": 339.6, "max_memory_gb": 25.46, "pair": 0.15786450542509556, "sign": 0.04865576047450304, "keep": 0.06358460430055857, "leak": 0.41685082390904427, "replacement": 0.3181973919272423}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 3.0744643211364746, "loss_scale": 4.0, "retries": 0, "seconds": 340.9, "max_memory_gb": 25.46, "pair": 0.3056006394326687, "sign": 0.1622292399406433, "keep": 0.04649623861769214, "leak": 0.27617837488651276, "replacement": 0.4790079742670059}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 6.765883445739746, "loss_scale": 4.0, "retries": 0, "seconds": 342.1, "max_memory_gb": 25.46, "pair": 0.5341056287288666, "sign": 0.7623143196105957, "keep": 0.189284211024642, "leak": 0.26884939859155566, "replacement": 1.0771172642707825}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 2.687265157699585, "loss_scale": 4.0, "retries": 0, "seconds": 343.4, "max_memory_gb": 25.46, "pair": 0.3883245512843132, "sign": 0.4675012044608593, "keep": 1.4371191374957561, "leak": 0.43367107688391116, "replacement": 1.4490525498986244}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 2.4066359996795654, "loss_scale": 4.0, "retries": 0, "seconds": 344.6, "max_memory_gb": 25.46, "pair": 0.13454893603920937, "sign": 0.07995020411908627, "keep": 0.07716438116040081, "leak": 0.724965634290129, "replacement": 0.3943476378917694}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 1.9352612495422363, "loss_scale": 4.0, "retries": 0, "seconds": 345.9, "max_memory_gb": 25.46, "pair": 0.30379467085003853, "sign": 0.26244378089904785, "keep": 0.02343240613117814, "leak": 0.4686628356575966, "replacement": 0.5638984814286232}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 7.780647277832031, "loss_scale": 4.0, "retries": 0, "seconds": 347.1, "max_memory_gb": 25.46, "pair": 0.410529550164938, "sign": 0.5805697664618492, "keep": 0.633355428930372, "leak": 0.6883506095036864, "replacement": 1.1895797848701477}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 4.136693954467773, "loss_scale": 4.0, "retries": 0, "seconds": 348.4, "max_memory_gb": 25.46, "pair": 0.3517339862883091, "sign": 0.2422076053917408, "keep": 0.1174830140080303, "leak": 1.483843332156539, "replacement": 0.9025401249527931}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 2.39827561378479, "loss_scale": 4.0, "retries": 0, "seconds": 349.6, "max_memory_gb": 25.46, "pair": 0.21800151094794273, "sign": 0.14326344802975655, "keep": 0.0442913721781224, "leak": 0.7445503812050447, "replacement": 0.4979165196418762}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 3.047612428665161, "loss_scale": 4.0, "retries": 0, "seconds": 350.8, "max_memory_gb": 25.46, "pair": 0.2846823204308748, "sign": 0.27354144398123026, "keep": 0.3542433388065547, "leak": 0.6755981042515486, "replacement": 0.7674742266535759}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 2.680694818496704, "loss_scale": 4.0, "retries": 0, "seconds": 352.1, "max_memory_gb": 25.46, "pair": 0.28172849398106337, "sign": 0.295524584595114, "keep": 0.1293752535711974, "leak": 0.3526809914037585, "replacement": 0.5823486521840096}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 3.758967399597168, "loss_scale": 4.0, "retries": 0, "seconds": 353.4, "max_memory_gb": 25.46, "pair": 0.2175303716212511, "sign": 0.09671263955533504, "keep": 0.1529707545414567, "leak": 2.634520798921585, "replacement": 1.0010022521018982}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.456888437271118, "loss_scale": 4.0, "retries": 0, "seconds": 354.6, "max_memory_gb": 25.46, "pair": 0.20022407546639442, "sign": 0.17120062187314034, "keep": 0.12920844019390643, "leak": 1.0259609560016543, "replacement": 0.6069188639521599}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 5.780380725860596, "loss_scale": 4.0, "retries": 0, "seconds": 355.9, "max_memory_gb": 25.46, "pair": 0.3419722430408001, "sign": 0.48746513575315475, "keep": 0.1615694637876004, "leak": 1.9913992844813038, "replacement": 1.1643393486738205}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 6.113957405090332, "loss_scale": 4.0, "retries": 0, "seconds": 357.2, "max_memory_gb": 25.46, "pair": 0.3765180353075266, "sign": 0.5489557273685932, "keep": 0.041399153880774975, "leak": 2.3218950405716896, "replacement": 1.2521692365407944}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 9.271957397460938, "loss_scale": 4.0, "retries": 0, "seconds": 358.5, "max_memory_gb": 25.46, "pair": 0.35326750576496124, "sign": 0.4090894181281328, "keep": 0.45252970419824123, "leak": 0.2598952248808928, "replacement": 0.8490508422255516}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 6.07159948348999, "loss_scale": 4.0, "retries": 0, "seconds": 359.9, "max_memory_gb": 25.46, "pair": 0.3309821905568242, "sign": 0.3519181115552783, "keep": 0.05995134310796857, "leak": 0.5959367509931326, "replacement": 0.6859010979533195}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 5.127730369567871, "loss_scale": 4.0, "retries": 0, "seconds": 361.2, "max_memory_gb": 25.46, "pair": 0.2638659030199051, "sign": 0.1784888245165348, "keep": 0.09083419851958752, "leak": 1.817038707435131, "replacement": 0.8527870923280716}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 5.624468803405762, "loss_scale": 4.0, "retries": 0, "seconds": 362.4, "max_memory_gb": 25.46, "pair": 0.33771458081901073, "sign": 0.42522307951003313, "keep": 0.3374989323783666, "leak": 0.9763893336057663, "replacement": 0.9631729051470757}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 4.967114448547363, "loss_scale": 4.0, "retries": 0, "seconds": 363.7, "max_memory_gb": 25.46, "pair": 0.12001818418502808, "sign": 0.05506711918860674, "keep": 0.03045304905390367, "leak": 2.200463280081749, "replacement": 0.712894082069397}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 3.468278408050537, "loss_scale": 4.0, "retries": 0, "seconds": 364.9, "max_memory_gb": 25.46, "pair": 0.18818650022149086, "sign": 0.15152418985962868, "keep": 0.0364894843660295, "leak": 0.9155889302492142, "replacement": 0.511090561747551}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 3.0828254222869873, "loss_scale": 4.0, "retries": 0, "seconds": 366.1, "max_memory_gb": 25.46, "pair": 0.29082145169377327, "sign": 0.28448787052184343, "keep": 0.07060470897704363, "leak": 0.6542071402072906, "replacement": 0.6319195032119751}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 5.172684192657471, "loss_scale": 4.0, "retries": 0, "seconds": 367.4, "max_memory_gb": 25.46, "pair": 0.2319682240486145, "sign": 0.14966874476522207, "keep": 0.2600122503936291, "leak": 1.2622515298426151, "replacement": 0.752371609210968}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 9.812440872192383, "loss_scale": 4.0, "retries": 0, "seconds": 368.6, "max_memory_gb": 25.46, "pair": 0.2154323272407055, "sign": 0.34491660445928574, "keep": 0.3014172570547089, "leak": 2.0687230452895164, "replacement": 1.0557800233364105}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 8.413714408874512, "loss_scale": 4.0, "retries": 0, "seconds": 369.9, "max_memory_gb": 25.46, "pair": 0.44732766784727573, "sign": 0.6987957060337067, "keep": 0.6475583421997726, "leak": 1.3103963239118457, "replacement": 1.448103815317154}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 5.727752208709717, "loss_scale": 4.0, "retries": 0, "seconds": 371.1, "max_memory_gb": 25.46, "pair": 0.5135489869862795, "sign": 0.7708241753280163, "keep": 0.5636661584721878, "leak": 0.6572168953716755, "replacement": 1.3450983613729477}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 4.334502220153809, "loss_scale": 4.0, "retries": 0, "seconds": 372.3, "max_memory_gb": 25.46, "pair": 0.49040840566158295, "sign": 0.6629445441067219, "keep": 0.6841763681732118, "leak": 0.3256449643522501, "replacement": 1.245380088686943}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 4.647542953491211, "loss_scale": 4.0, "retries": 0, "seconds": 373.6, "max_memory_gb": 25.46, "pair": 0.370386715978384, "sign": 0.12666145525872707, "keep": 0.06338469823822379, "leak": 1.3880707379430532, "replacement": 0.8124275058507919}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 7.497790813446045, "loss_scale": 4.0, "retries": 0, "seconds": 374.8, "max_memory_gb": 25.46, "pair": 0.6142344512045383, "sign": 0.8447438217699528, "keep": 0.5652418006211519, "leak": 1.8035313518485054, "replacement": 1.7701101303100586}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 4.208549499511719, "loss_scale": 4.0, "retries": 0, "seconds": 376.1, "max_memory_gb": 25.46, "pair": 0.2696394696831703, "sign": 0.3475143173709512, "keep": 0.10329324332997203, "leak": 0.8283052272163332, "replacement": 0.7021195441484451}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 2.7712807655334473, "loss_scale": 4.0, "retries": 0, "seconds": 377.3, "max_memory_gb": 25.46, "pair": 0.18428411707282066, "sign": 0.053982126992195845, "keep": 0.09412647469434887, "leak": 0.44911667061114713, "replacement": 0.37061757594347}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 5.413727760314941, "loss_scale": 4.0, "retries": 0, "seconds": 378.6, "max_memory_gb": 25.46, "pair": 0.3229979556053877, "sign": 0.2827771808952093, "keep": 0.08479439141228795, "leak": 1.7426309511065483, "replacement": 0.9424414858222008}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 5.778692245483398, "loss_scale": 4.0, "retries": 0, "seconds": 379.8, "max_memory_gb": 25.46, "pair": 0.3850948675535619, "sign": 0.4667505966499448, "keep": 0.05325123481452465, "leak": 1.5267453073392971, "replacement": 1.0267821326851845}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 9.571980476379395, "loss_scale": 4.0, "retries": 0, "seconds": 381.1, "max_memory_gb": 25.46, "pair": 0.3756231665611267, "sign": 0.1945854090154171, "keep": 0.10787925164913759, "leak": 1.804367296397686, "replacement": 0.9779473394155502}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 3.394632339477539, "loss_scale": 4.0, "retries": 0, "seconds": 382.3, "max_memory_gb": 25.46, "pair": 0.16527186892926693, "sign": 0.03971291380003095, "keep": 0.0569938188418746, "leak": 1.0539003536105156, "replacement": 0.47710032761096954}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 11.625691413879395, "loss_scale": 4.0, "retries": 0, "seconds": 383.6, "max_memory_gb": 25.46, "pair": 0.4350121170282364, "sign": 0.4401124962605536, "keep": 0.4727174984291196, "leak": 0.8429024834185839, "replacement": 1.1021527200937271}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 5.434326171875, "loss_scale": 4.0, "retries": 0, "seconds": 384.9, "max_memory_gb": 25.46, "pair": 0.18794650118798018, "sign": 0.38443307019770145, "keep": 0.08487357365083881, "leak": 1.7871797531843185, "replacement": 0.8693947456777096}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 16.158252716064453, "loss_scale": 4.0, "retries": 0, "seconds": 386.1, "max_memory_gb": 25.46, "pair": 0.6474251858890057, "sign": 0.7596321413293481, "keep": 0.6315075699239969, "leak": 0.8498118054121733, "replacement": 1.5554480403661728}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 2.172910690307617, "loss_scale": 4.0, "retries": 0, "seconds": 387.3, "max_memory_gb": 25.46, "pair": 0.28472306206822395, "sign": 0.10633381083607674, "keep": 0.06134141294751316, "leak": 1.1946248896420002, "replacement": 0.6672168895602226}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 4.1509013175964355, "loss_scale": 4.0, "retries": 0, "seconds": 388.5, "max_memory_gb": 25.46, "pair": 0.3199978172779083, "sign": 0.40965742617845535, "keep": 0.30693183635594323, "leak": 0.5616673976182938, "replacement": 0.8187093138694763}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.668918609619141, "loss_scale": 4.0, "retries": 0, "seconds": 389.8, "max_memory_gb": 25.46, "pair": 0.25396882742643356, "sign": 0.17846847884356976, "keep": 0.05414150597061962, "leak": 0.256366060115397, "replacement": 0.43436533957719803}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 10.82689094543457, "loss_scale": 4.0, "retries": 0, "seconds": 391.0, "max_memory_gb": 25.46, "pair": 0.4762314725667238, "sign": 0.6214602808468044, "keep": 0.7539303787052631, "leak": 0.8679171726107597, "replacement": 1.380906119942665}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 9.139827728271484, "loss_scale": 4.0, "retries": 0, "seconds": 392.3, "max_memory_gb": 25.46, "pair": 0.34056070633232594, "sign": 0.40971092507243156, "keep": 0.5258806669153273, "leak": 1.3220230787992477, "replacement": 1.1388622522354126}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 4.347232818603516, "loss_scale": 4.0, "retries": 0, "seconds": 393.6, "max_memory_gb": 25.46, "pair": 0.22511987760663033, "sign": 0.20430232817307115, "keep": 0.08525823522359133, "leak": 0.828400962986052, "replacement": 0.5770004093647003}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 3.1761257648468018, "loss_scale": 4.0, "retries": 0, "seconds": 394.8, "max_memory_gb": 25.46, "pair": 0.61465884745121, "sign": 0.7739147432148457, "keep": 0.4112493172287941, "leak": 0.838352934923023, "replacement": 1.4168291538953781}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 4.794064044952393, "loss_scale": 4.0, "retries": 0, "seconds": 396.0, "max_memory_gb": 25.46, "pair": 0.32064469903707504, "sign": 0.36033246852457523, "keep": 0.31012640707194805, "leak": 1.2480522617697716, "replacement": 0.967887207865715}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 3.998859405517578, "loss_scale": 4.0, "retries": 0, "seconds": 397.3, "max_memory_gb": 25.46, "pair": 0.2741983821615577, "sign": 0.22326945699751377, "keep": 0.10903369204606861, "leak": 0.8140373192727566, "replacement": 0.6438592821359634}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 5.864408493041992, "loss_scale": 4.0, "retries": 0, "seconds": 398.5, "max_memory_gb": 25.46, "pair": 0.39753849245607853, "sign": 0.33900793828070164, "keep": 0.1690744962543249, "leak": 0.4106034431606531, "replacement": 0.7542305588722229}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 5.616261005401611, "loss_scale": 4.0, "retries": 0, "seconds": 399.7, "max_memory_gb": 25.46, "pair": 0.3574652671813965, "sign": 0.3976777596399188, "keep": 0.36120859475340694, "leak": 0.6146154180169106, "replacement": 0.8905622661113739}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 4.730462074279785, "loss_scale": 4.0, "retries": 0, "seconds": 401.0, "max_memory_gb": 25.46, "pair": 0.2831912189722061, "sign": 0.20072962529957294, "keep": 0.02980174054391682, "leak": 0.802383181406185, "replacement": 0.5990526974201202}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 4.362586498260498, "loss_scale": 4.0, "retries": 0, "seconds": 402.2, "max_memory_gb": 25.46, "pair": 0.3417247235774994, "sign": 0.4350440874695778, "keep": 0.046882330905646086, "leak": 0.2034632358700037, "replacement": 0.6335537359118462}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 9.526244163513184, "loss_scale": 4.0, "retries": 0, "seconds": 403.5, "max_memory_gb": 25.46, "pair": 0.09604664705693722, "sign": 0.09029950760304928, "keep": 0.08741739019751549, "leak": 3.5120998919010162, "replacement": 1.0629300698637962}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 5.608996868133545, "loss_scale": 4.0, "retries": 0, "seconds": 404.7, "max_memory_gb": 25.46, "pair": 0.154694858007133, "sign": 0.08060769038274884, "keep": 0.0963865693192929, "leak": 1.3191060253884643, "replacement": 0.572968490421772}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 10.101175308227539, "loss_scale": 4.0, "retries": 0, "seconds": 406.0, "max_memory_gb": 25.46, "pair": 0.781269371509552, "sign": 0.6872762474231422, "keep": 0.6185344490222633, "leak": 0.11217605699494015, "replacement": 1.4622186794877052}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 10.33308219909668, "loss_scale": 4.0, "retries": 0, "seconds": 407.2, "max_memory_gb": 25.46, "pair": 0.5292898043990135, "sign": 0.6001302525401115, "keep": 0.4814270284259692, "leak": 0.8306389050558209, "replacement": 1.2777281627058983}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 5.413313388824463, "loss_scale": 4.0, "retries": 0, "seconds": 408.5, "max_memory_gb": 25.46, "pair": 0.18993252515792847, "sign": 0.10143343079835176, "keep": 0.39891401771456003, "leak": 0.5333294523879886, "replacement": 0.5734385997056961}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 9.69287395477295, "loss_scale": 4.0, "retries": 0, "seconds": 409.7, "max_memory_gb": 25.46, "pair": 0.4615560248494148, "sign": 0.6532511487603188, "keep": 0.2667792895808816, "leak": 0.2854193579405546, "replacement": 0.9929260760545731}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 7.690202236175537, "loss_scale": 4.0, "retries": 0, "seconds": 411.0, "max_memory_gb": 25.46, "pair": 0.28522706124931574, "sign": 0.39687298610806465, "keep": 0.33728080592118204, "leak": 1.1071947067975998, "replacement": 0.9291026443243027}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 5.241476058959961, "loss_scale": 4.0, "retries": 0, "seconds": 412.2, "max_memory_gb": 25.46, "pair": 0.48789358139038086, "sign": 0.5162013098597527, "keep": 0.3597056125290692, "leak": 0.40622451185481623, "replacement": 1.027403138577938}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 7.152268409729004, "loss_scale": 4.0, "retries": 0, "seconds": 413.5, "max_memory_gb": 25.46, "pair": 0.6685036197304726, "sign": 0.6458547376096249, "keep": 0.7239657928876113, "leak": 0.8520702021196485, "replacement": 1.566431537270546}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 4.935558319091797, "loss_scale": 4.0, "retries": 0, "seconds": 414.8, "max_memory_gb": 25.46, "pair": 0.318821806460619, "sign": 0.3006742522120476, "keep": 0.37116084166336805, "leak": 0.7650568950921297, "replacement": 0.8460035920143127}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 6.306670665740967, "loss_scale": 4.0, "retries": 0, "seconds": 416.0, "max_memory_gb": 25.46, "pair": 0.6275100912898779, "sign": 0.7606086581945419, "keep": 0.5950231388560496, "leak": 0.3382202279390185, "replacement": 1.3898809999227524}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 5.916029453277588, "loss_scale": 4.0, "retries": 0, "seconds": 417.2, "max_memory_gb": 25.46, "pair": 0.5231538861989975, "sign": 0.5403045900166035, "keep": 0.5545162614434958, "leak": 1.3135804124758579, "replacement": 1.398959442973137}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 6.840750217437744, "loss_scale": 4.0, "retries": 0, "seconds": 418.5, "max_memory_gb": 25.46, "pair": 0.528459295630455, "sign": 0.6101627871394157, "keep": 0.9199302061460912, "leak": 0.2831006897613406, "replacement": 1.3642809838056564}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 7.773780345916748, "loss_scale": 4.0, "retries": 0, "seconds": 419.8, "max_memory_gb": 25.46, "pair": 0.23950173519551754, "sign": 0.12912065163254738, "keep": 0.04500868753530085, "leak": 1.229059753473848, "replacement": 0.6338313296437263}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 20.656967163085938, "loss_scale": 4.0, "retries": 0, "seconds": 421.1, "max_memory_gb": 25.46, "pair": 0.6573759019374847, "sign": 0.7875584363937378, "keep": 0.3806120231747627, "leak": 1.3468285230337642, "replacement": 1.5781682878732681}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 6.927961826324463, "loss_scale": 4.0, "retries": 0, "seconds": 422.3, "max_memory_gb": 25.46, "pair": 0.3954401593655348, "sign": 0.5172505704686046, "keep": 0.4372628848068416, "leak": 0.9787280429154634, "replacement": 1.1173788607120514}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 7.857402324676514, "loss_scale": 4.0, "retries": 0, "seconds": 423.6, "max_memory_gb": 25.46, "pair": 0.4223513398319483, "sign": 0.7411779230460525, "keep": 0.36826234869658947, "leak": 2.0795074068009853, "replacement": 1.4969483017921448}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 6.449702739715576, "loss_scale": 4.0, "retries": 0, "seconds": 424.8, "max_memory_gb": 25.46, "pair": 0.3643897566944361, "sign": 0.3795662419870496, "keep": 0.8180543390335515, "leak": 0.679987576790154, "replacement": 1.1331969890743494}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 6.28177547454834, "loss_scale": 4.0, "retries": 0, "seconds": 426.1, "max_memory_gb": 25.46, "pair": 0.22497951798141003, "sign": 0.2802457856014371, "keep": 0.25840812595561147, "leak": 1.5814454071223736, "replacement": 0.8896678388118744}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 9.613076210021973, "loss_scale": 4.0, "retries": 0, "seconds": 427.3, "max_memory_gb": 25.46, "pair": 0.555723013356328, "sign": 0.6621558554470539, "keep": 0.6254597301594913, "leak": 1.159068836364895, "replacement": 1.4892979860305786}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 4.209733009338379, "loss_scale": 4.0, "retries": 0, "seconds": 428.6, "max_memory_gb": 25.46, "pair": 0.3897683285176754, "sign": 0.5380415972322226, "keep": 0.02132997033186257, "leak": 0.3177038077265024, "replacement": 0.7488800548017025}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 12.755878448486328, "loss_scale": 4.0, "retries": 0, "seconds": 429.8, "max_memory_gb": 25.46, "pair": 0.2232019156217575, "sign": 0.22787481732666492, "keep": 0.11270252618123777, "leak": 1.5399964451789856, "replacement": 0.7784896790981293}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 6.111617565155029, "loss_scale": 4.0, "retries": 0, "seconds": 431.0, "max_memory_gb": 25.46, "pair": 0.26056910306215286, "sign": 0.37189921783283353, "keep": 0.2530938198324293, "leak": 0.6829853462259052, "replacement": 0.7438119500875473}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 8.875360488891602, "loss_scale": 4.0, "retries": 0, "seconds": 432.3, "max_memory_gb": 25.46, "pair": 0.49873396568000317, "sign": 0.5651055797934532, "keep": 0.5909108942723833, "leak": 0.9121044054627419, "replacement": 1.3047683238983154}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 4.77016019821167, "loss_scale": 4.0, "retries": 0, "seconds": 433.5, "max_memory_gb": 25.46, "pair": 0.3856996167451143, "sign": 0.3870041435584426, "keep": 0.3184978417120874, "leak": 0.5329037972592232, "replacement": 0.8716765865683556}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 4.660843849182129, "loss_scale": 4.0, "retries": 0, "seconds": 434.8, "max_memory_gb": 25.46, "pair": 0.2981246178969741, "sign": 0.1476105316542089, "keep": 0.018060040078125894, "leak": 1.9813755368813872, "replacement": 0.8763037994503975}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 7.221625804901123, "loss_scale": 4.0, "retries": 0, "seconds": 436.1, "max_memory_gb": 25.46, "pair": 0.3280230909585953, "sign": 0.35747560765594244, "keep": 0.14532712084474042, "leak": 1.6631290479563177, "replacement": 0.9952067360281944}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 6.309030055999756, "loss_scale": 4.0, "retries": 0, "seconds": 437.3, "max_memory_gb": 25.46, "pair": 0.39997976645827293, "sign": 0.6833075061440468, "keep": 0.25644833594560623, "leak": 0.60661441530101, "replacement": 1.021511286497116}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 3.7776317596435547, "loss_scale": 4.0, "retries": 0, "seconds": 438.6, "max_memory_gb": 25.46, "pair": 0.18469005543738604, "sign": 0.09260321222245693, "keep": 0.0643197181634605, "leak": 0.5491461688652635, "replacement": 0.40043807215988636}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 11.519439697265625, "loss_scale": 4.0, "retries": 0, "seconds": 439.8, "max_memory_gb": 25.46, "pair": 0.37610155530273914, "sign": 0.5909452112391591, "keep": 0.5739177791401744, "leak": 1.1754883229732513, "replacement": 1.2524051517248154}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 8.773253440856934, "loss_scale": 4.0, "retries": 0, "seconds": 441.1, "max_memory_gb": 25.46, "pair": 0.41850778087973595, "sign": 0.4922151230275631, "keep": 0.5698694102466106, "leak": 0.6124163679778576, "replacement": 1.102654106914997}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 1.9638042449951172, "loss_scale": 4.0, "retries": 0, "seconds": 442.3, "max_memory_gb": 25.46, "pair": 0.419558584690094, "sign": 0.4354106690734625, "keep": 0.32572155428351834, "leak": 0.2638836630212609, "replacement": 0.8660955876111984}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 6.697314262390137, "loss_scale": 4.0, "retries": 0, "seconds": 443.6, "max_memory_gb": 25.46, "pair": 0.40003200247883797, "sign": 0.41012209258042276, "keep": 0.20103542134165764, "leak": 1.9854339921148494, "replacement": 1.2019692435860634}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 7.133281707763672, "loss_scale": 4.0, "retries": 0, "seconds": 444.8, "max_memory_gb": 25.46, "pair": 0.15524994023144245, "sign": 0.15316697023808956, "keep": 0.042750393971800804, "leak": 2.456178978085518, "replacement": 0.8672533482313156}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 9.288538932800293, "loss_scale": 4.0, "retries": 0, "seconds": 446.1, "max_memory_gb": 25.46, "pair": 0.26812940277159214, "sign": 0.09134497796185315, "keep": 0.17962696589529514, "leak": 1.6830714866518974, "replacement": 0.8243832588195801}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 5.548547744750977, "loss_scale": 4.0, "retries": 0, "seconds": 447.3, "max_memory_gb": 25.46, "pair": 0.5598958134651184, "sign": 0.6801715474575758, "keep": 0.4866702890722081, "leak": 0.19213900723843835, "replacement": 1.1913514956831932}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 2.7761006355285645, "loss_scale": 4.0, "retries": 0, "seconds": 448.6, "max_memory_gb": 25.46, "pair": 0.3633723855018616, "sign": 0.48694006353616714, "keep": 0.09108217956963927, "leak": 0.7777996417135, "replacement": 0.8468334153294563}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 6.752566814422607, "loss_scale": 4.0, "retries": 0, "seconds": 449.9, "max_memory_gb": 25.46, "pair": 0.2302368599921465, "sign": 0.19302861392498016, "keep": 0.041699636640259996, "leak": 1.7652257601730525, "replacement": 0.7889073938131332}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 8.439687728881836, "loss_scale": 4.0, "retries": 0, "seconds": 451.1, "max_memory_gb": 25.46, "pair": 0.2983120419085026, "sign": 0.466291431337595, "keep": 0.0367574154806789, "leak": 1.313756424933672, "replacement": 0.8782756179571152}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 6.9331583976745605, "loss_scale": 4.0, "retries": 0, "seconds": 452.3, "max_memory_gb": 25.46, "pair": 0.425291633233428, "sign": 0.44453428126871586, "keep": 0.7708656887989491, "leak": 0.549590287682804, "replacement": 1.1703891307115555}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 8.461650848388672, "loss_scale": 4.0, "retries": 0, "seconds": 453.6, "max_memory_gb": 25.46, "pair": 0.4648529216647148, "sign": 0.788044486194849, "keep": 0.058219908736646175, "leak": 0.3485032421340293, "replacement": 0.9751109108328819}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 4.875121116638184, "loss_scale": 4.0, "retries": 0, "seconds": 454.9, "max_memory_gb": 25.46, "pair": 0.29436594247817993, "sign": 0.06193281104788184, "keep": 0.05112184816971421, "leak": 0.7445801142603159, "replacement": 0.5370382890105247}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 4.705489158630371, "loss_scale": 4.0, "retries": 0, "seconds": 456.1, "max_memory_gb": 25.46, "pair": 0.3435169607400894, "sign": 0.3747842349112034, "keep": 0.22501365304924548, "leak": 0.18861868791282177, "replacement": 0.6905705481767654}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 7.581827640533447, "loss_scale": 4.0, "retries": 0, "seconds": 457.4, "max_memory_gb": 25.46, "pair": 0.2887905668467283, "sign": 0.29915418289601803, "keep": 0.07773314788937569, "leak": 2.3383397683501244, "replacement": 1.061819151043892}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 4.796443939208984, "loss_scale": 4.0, "retries": 0, "seconds": 458.6, "max_memory_gb": 25.46, "pair": 0.4054899252951145, "sign": 0.4265010226517916, "keep": 0.4319485833402723, "leak": 0.6667597852647305, "replacement": 1.0014047175645828}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 5.52311372756958, "loss_scale": 4.0, "retries": 0, "seconds": 459.8, "max_memory_gb": 25.46, "pair": 0.24245556816458702, "sign": 0.12472930364310741, "keep": 0.04191065288614482, "leak": 0.47337090970540885, "replacement": 0.44411826878786087}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 6.566603183746338, "loss_scale": 4.0, "retries": 0, "seconds": 461.1, "max_memory_gb": 25.46, "pair": 0.43298454210162163, "sign": 0.402136102784425, "keep": 0.07716275891289115, "leak": 0.9473906999483006, "replacement": 0.9094816148281097}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 10.05368423461914, "loss_scale": 4.0, "retries": 0, "seconds": 462.3, "max_memory_gb": 25.46, "pair": 0.22943726554512978, "sign": 0.2883790656924248, "keep": 1.0052675269544125, "leak": 0.4343692627735436, "replacement": 0.9848528653383255}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.744368076324463, "loss_scale": 4.0, "retries": 0, "seconds": 463.6, "max_memory_gb": 25.46, "pair": 0.16806875728070736, "sign": 0.06655835220590234, "keep": 0.31767371692694724, "leak": 0.8156669549643993, "replacement": 0.5641015134751797}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 9.251585960388184, "loss_scale": 4.0, "retries": 0, "seconds": 464.8, "max_memory_gb": 25.46, "pair": 0.29661508835852146, "sign": 0.28176192101091146, "keep": 0.1130759374354966, "leak": 1.0115710478276014, "replacement": 0.7469267882406712}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 4.02011251449585, "loss_scale": 4.0, "retries": 0, "seconds": 466.0, "max_memory_gb": 25.46, "pair": 0.5050577968358994, "sign": 0.6985405758023262, "keep": 0.05596692470135167, "leak": 0.3693378893658519, "replacement": 0.9746460095047951}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 7.210978984832764, "loss_scale": 4.0, "retries": 0, "seconds": 467.3, "max_memory_gb": 25.46, "pair": 0.3085107896476984, "sign": 0.31100855581462383, "keep": 0.48808784992434084, "leak": 0.7533202958293259, "replacement": 0.8963890969753265}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 4.819110870361328, "loss_scale": 4.0, "retries": 0, "seconds": 468.6, "max_memory_gb": 25.46, "pair": 0.45847950130701065, "sign": 0.49577294662594795, "keep": 0.45005319948541, "leak": 0.37723685521632433, "replacement": 1.0257017612457275}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.1934306621551514, "loss_scale": 4.0, "retries": 0, "seconds": 469.8, "max_memory_gb": 25.46, "pair": 0.18943284545093775, "sign": 0.1843999559059739, "keep": 0.09175987855996937, "leak": 0.3204127661883831, "replacement": 0.4076159596443176}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 4.593886375427246, "loss_scale": 4.0, "retries": 0, "seconds": 471.0, "max_memory_gb": 25.46, "pair": 0.27100830525159836, "sign": 0.3078255979344249, "keep": 0.05625391309149563, "leak": 0.7021725005470216, "replacement": 0.6285911686718464}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 4.80058479309082, "loss_scale": 4.0, "retries": 0, "seconds": 472.3, "max_memory_gb": 25.46, "pair": 0.19279731810092926, "sign": 0.15816893428564072, "keep": 0.0747601306065917, "leak": 0.6924212370067835, "replacement": 0.4823671542108059}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 13.411952018737793, "loss_scale": 4.0, "retries": 0, "seconds": 473.5, "max_memory_gb": 25.46, "pair": 0.2732607237994671, "sign": 0.37446818221360445, "keep": 0.378264625556767, "leak": 0.2955457968637347, "replacement": 0.7235135808587074}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 11.337331771850586, "loss_scale": 4.0, "retries": 0, "seconds": 474.8, "max_memory_gb": 25.46, "pair": 0.21401014039292932, "sign": 0.30399176152423024, "keep": 0.06993807386606932, "leak": 3.0024424977600574, "replacement": 1.1515856832265854}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 3.931109666824341, "loss_scale": 4.0, "retries": 0, "seconds": 476.0, "max_memory_gb": 25.46, "pair": 0.5707836709916592, "sign": 0.7754157558083534, "keep": 0.3807599386200309, "leak": 0.281360583845526, "replacement": 1.219211608171463}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 7.301445007324219, "loss_scale": 4.0, "retries": 0, "seconds": 477.3, "max_memory_gb": 25.46, "pair": 0.38328222930431366, "sign": 0.5364417596720159, "keep": 1.0745819676667452, "leak": 0.6408395208418369, "replacement": 1.3490039519965649}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 3.892526865005493, "loss_scale": 4.0, "retries": 0, "seconds": 478.5, "max_memory_gb": 25.46, "pair": 0.24703477323055267, "sign": 0.26481733098626137, "keep": 0.14739118237048388, "leak": 0.86754534766078, "replacement": 0.6700253784656525}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 5.435022830963135, "loss_scale": 4.0, "retries": 0, "seconds": 479.7, "max_memory_gb": 25.46, "pair": 0.08625867636874318, "sign": 0.05493883602321148, "keep": 0.03488699812442064, "leak": 0.9433127716183662, "replacement": 0.3669997826218605}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 4.197782039642334, "loss_scale": 4.0, "retries": 0, "seconds": 481.0, "max_memory_gb": 25.46, "pair": 0.38686276227235794, "sign": 0.47824327275156975, "keep": 0.30173361918423325, "leak": 0.6150330619420856, "replacement": 0.9306094571948051}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 6.255158424377441, "loss_scale": 4.0, "retries": 0, "seconds": 482.3, "max_memory_gb": 25.46, "pair": 0.2428848510608077, "sign": 0.15298225497826934, "keep": 0.039993690443225205, "leak": 0.3498798299115151, "replacement": 0.4268427789211273}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 6.744281768798828, "loss_scale": 4.0, "retries": 0, "seconds": 483.6, "max_memory_gb": 25.46, "pair": 0.18932647537440062, "sign": 0.39031951129436493, "keep": 0.08837422705255449, "leak": 1.0968704130500555, "replacement": 0.7028909474611282}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 6.561184883117676, "loss_scale": 4.0, "retries": 0, "seconds": 484.9, "max_memory_gb": 25.46, "pair": 0.17864062264561653, "sign": 0.22610437497496605, "keep": 0.09044475381961092, "leak": 0.6634561084210873, "replacement": 0.5027792230248451}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 4.925157070159912, "loss_scale": 4.0, "retries": 0, "seconds": 486.2, "max_memory_gb": 25.46, "pair": 0.2703659888356924, "sign": 0.11852789763361216, "keep": 0.14524522796273232, "leak": 1.8540868610143661, "replacement": 0.8657742738723755}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 16.441423416137695, "loss_scale": 4.0, "retries": 0, "seconds": 487.4, "max_memory_gb": 25.46, "pair": 0.32678121514618397, "sign": 0.5353038869798183, "keep": 0.4919407411944121, "leak": 1.3997268229722977, "replacement": 1.1903352290391922}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 6.359706878662109, "loss_scale": 4.0, "retries": 0, "seconds": 488.7, "max_memory_gb": 25.46, "pair": 0.5146736800670624, "sign": 0.7536637969315052, "keep": 0.5695765633136034, "leak": 0.7713489253073931, "replacement": 1.369131125509739}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 10.445344924926758, "loss_scale": 4.0, "retries": 0, "seconds": 489.9, "max_memory_gb": 25.46, "pair": 0.13054238446056843, "sign": 0.09962789434939623, "keep": 0.07528963539516553, "leak": 0.47452718764543533, "replacement": 0.33663294464349747}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 7.727506160736084, "loss_scale": 4.0, "retries": 0, "seconds": 491.2, "max_memory_gb": 25.46, "pair": 0.11772955860942602, "sign": 0.17401933390647173, "keep": 0.06520652677863836, "leak": 1.1546119004487991, "replacement": 0.5259954743087292}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 9.740571022033691, "loss_scale": 4.0, "retries": 0, "seconds": 492.5, "max_memory_gb": 25.46, "pair": 0.359606608748436, "sign": 0.4242287864908576, "keep": 0.15111401723697782, "leak": 0.7788430470245657, "replacement": 0.841988779604435}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 10.16200065612793, "loss_scale": 4.0, "retries": 0, "seconds": 493.7, "max_memory_gb": 25.46, "pair": 0.10268967412412167, "sign": 0.12176623847335577, "keep": 0.031142514024395496, "leak": 2.591898888349533, "replacement": 0.8271187767386436}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 4.5075812339782715, "loss_scale": 4.0, "retries": 0, "seconds": 494.9, "max_memory_gb": 25.46, "pair": 0.1536857057362795, "sign": 0.06424755696207285, "keep": 0.1453394964337349, "leak": 0.44313280656933784, "replacement": 0.3692624308168888}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 6.791235446929932, "loss_scale": 4.0, "retries": 0, "seconds": 496.2, "max_memory_gb": 25.46, "pair": 0.2854363080114126, "sign": 0.4543821942061186, "keep": 0.23194173350930214, "leak": 0.7049617278389633, "replacement": 0.8048387132585049}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 7.442294120788574, "loss_scale": 4.0, "retries": 0, "seconds": 497.4, "max_memory_gb": 25.46, "pair": 0.35241672210395336, "sign": 0.3675050400197506, "keep": 0.22330234886612743, "leak": 0.6061199894174933, "replacement": 0.7993504405021667}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 12.462393760681152, "loss_scale": 4.0, "retries": 0, "seconds": 498.7, "max_memory_gb": 25.46, "pair": 0.4051373340189457, "sign": 0.6719154776073992, "keep": 0.8237601639702916, "leak": 0.530196662992239, "replacement": 1.2855243161320686}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 10.094801902770996, "loss_scale": 4.0, "retries": 0, "seconds": 500.0, "max_memory_gb": 25.46, "pair": 0.4096067100763321, "sign": 0.2643470950424671, "keep": 0.21582834422588348, "leak": 1.1555146984755993, "replacement": 0.9385731220245361}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 29.882598876953125, "loss_scale": 4.0, "retries": 0, "seconds": 501.3, "max_memory_gb": 25.46, "pair": 0.4040592946112156, "sign": 0.463833000510931, "keep": 0.43894744373392314, "leak": 1.3491040356457233, "replacement": 1.1927254870533943}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 9.321873664855957, "loss_scale": 4.0, "retries": 0, "seconds": 502.5, "max_memory_gb": 25.46, "pair": 0.6237927116453648, "sign": 0.7491380022838712, "keep": 0.27803870290517807, "leak": 0.5538450560852652, "replacement": 1.275842323899269}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 4.987494468688965, "loss_scale": 4.0, "retries": 0, "seconds": 503.8, "max_memory_gb": 25.46, "pair": 0.20326543413102627, "sign": 0.11164867877960205, "keep": 0.0505508144851774, "leak": 1.5042339526116848, "replacement": 0.660423681139946}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 15.669327735900879, "loss_scale": 4.0, "retries": 0, "seconds": 505.0, "max_memory_gb": 25.46, "pair": 0.3025658940896392, "sign": 0.4621686968021095, "keep": 0.451944200322032, "leak": 0.8240637332201004, "replacement": 0.9656382463872433}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 6.001237869262695, "loss_scale": 4.0, "retries": 0, "seconds": 506.3, "max_memory_gb": 25.46, "pair": 0.2041943296790123, "sign": 0.17484288476407528, "keep": 0.050579712144099176, "leak": 0.5915587581694126, "replacement": 0.4647953175008297}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 9.383561134338379, "loss_scale": 4.0, "retries": 0, "seconds": 507.6, "max_memory_gb": 25.46, "pair": 0.715680219233036, "sign": 0.782391294836998, "keep": 0.31010501831769943, "leak": 0.45263553224503994, "replacement": 1.3750872611999512}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 15.543074607849121, "loss_scale": 4.0, "retries": 0, "seconds": 508.8, "max_memory_gb": 25.46, "pair": 0.7398139536380768, "sign": 0.989165079779923, "keep": 2.275540016591549, "leak": 0.8361561830388382, "replacement": 2.5812056213617325}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 6.024654388427734, "loss_scale": 4.0, "retries": 0, "seconds": 510.1, "max_memory_gb": 25.46, "pair": 0.31208540173247457, "sign": 0.4347946997731924, "keep": 0.4505763644701801, "leak": 0.7701965330634266, "replacement": 0.9473200142383575}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 6.261988639831543, "loss_scale": 4.0, "retries": 0, "seconds": 511.3, "max_memory_gb": 25.46, "pair": 0.4469820447266102, "sign": 0.434458464384079, "keep": 0.4014986274414696, "leak": 0.8084792296285741, "replacement": 1.067080408334732}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 8.25140380859375, "loss_scale": 4.0, "retries": 0, "seconds": 512.6, "max_memory_gb": 25.46, "pair": 0.14114294946193695, "sign": 0.10078463470563293, "keep": 0.36012447108805645, "leak": 1.7226706147193909, "replacement": 0.8022651672363281}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 11.163050651550293, "loss_scale": 4.0, "retries": 0, "seconds": 513.9, "max_memory_gb": 25.46, "pair": 0.35424969531595707, "sign": 0.5290687084197998, "keep": 0.8016843227669597, "leak": 1.2199465094599873, "replacement": 1.3246128298342228}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 15.284759521484375, "loss_scale": 4.0, "retries": 0, "seconds": 515.1, "max_memory_gb": 25.46, "pair": 0.4423240236938, "sign": 0.6473347283899784, "keep": 0.33076218888163567, "leak": 0.243120557628572, "replacement": 0.9921526461839676}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 8.677412986755371, "loss_scale": 4.0, "retries": 0, "seconds": 516.3, "max_memory_gb": 25.46, "pair": 0.10951072163879871, "sign": 0.09571225009858608, "keep": 0.021576860977802426, "leak": 2.0869951844215393, "replacement": 0.6899040639400482}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 17.578060150146484, "loss_scale": 4.0, "retries": 0, "seconds": 517.6, "max_memory_gb": 25.46, "pair": 0.4924707394093275, "sign": 0.452102230861783, "keep": 1.0461009186692536, "leak": 0.39773232862353325, "replacement": 1.3410053178668022}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 8.509010314941406, "loss_scale": 4.0, "retries": 0, "seconds": 518.8, "max_memory_gb": 25.46, "pair": 0.3886520741507411, "sign": 0.3800894967280328, "keep": 1.5316273784264922, "leak": 0.6210865404436845, "replacement": 1.4997821971774101}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 4.242802619934082, "loss_scale": 4.0, "retries": 0, "seconds": 520.0, "max_memory_gb": 25.46, "pair": 0.18465572409331799, "sign": 0.12386510614305735, "keep": 0.14143929514102638, "leak": 0.43264035135507584, "replacement": 0.42546800151467323}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 21.381738662719727, "loss_scale": 4.0, "retries": 0, "seconds": 521.2, "max_memory_gb": 25.46, "pair": 0.8323977254331112, "sign": 1.2480455711483955, "keep": 1.6618438365403563, "leak": 0.7878679670393467, "replacement": 2.484309360384941}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 4.843331813812256, "loss_scale": 4.0, "retries": 0, "seconds": 522.4, "max_memory_gb": 25.46, "pair": 0.13011762965470552, "sign": 0.18890128005295992, "keep": 0.035731382260564715, "leak": 0.34993919415865093, "replacement": 0.3299187570810318}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 11.630579948425293, "loss_scale": 4.0, "retries": 0, "seconds": 523.6, "max_memory_gb": 25.46, "pair": 0.3711885139346123, "sign": 0.6126252338290215, "keep": 0.07896457010065205, "leak": 1.8454628679901361, "replacement": 1.1783490926027298}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 15.951029777526855, "loss_scale": 4.0, "retries": 0, "seconds": 524.9, "max_memory_gb": 25.46, "pair": 0.18716744985431433, "sign": 0.38885101303458214, "keep": 0.6182429655455053, "leak": 2.2703523077070713, "replacement": 1.2583025470376015}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 10.180791854858398, "loss_scale": 4.0, "retries": 0, "seconds": 526.1, "max_memory_gb": 25.46, "pair": 0.18324541952461004, "sign": 0.252568231895566, "keep": 0.07806286634877324, "leak": 1.846273812057916, "replacement": 0.8101294189691544}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 8.99378776550293, "loss_scale": 4.0, "retries": 0, "seconds": 527.3, "max_memory_gb": 25.46, "pair": 0.5902224443852901, "sign": 0.5698165465146303, "keep": 0.5241968473419547, "leak": 0.12996907242268207, "replacement": 1.1697214618325233}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 9.643141746520996, "loss_scale": 4.0, "retries": 0, "seconds": 528.6, "max_memory_gb": 25.46, "pair": 0.5843901820480824, "sign": 0.7916518524289131, "keep": 0.8947248719632626, "leak": 0.5077276974916458, "replacement": 1.5545104518532753}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 4.066357135772705, "loss_scale": 4.0, "retries": 0, "seconds": 529.8, "max_memory_gb": 25.46, "pair": 0.12071173451840878, "sign": 0.06650479137897491, "keep": 0.0878333318978548, "leak": 0.8503191508352757, "replacement": 0.41046058014035225}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 10.044111251831055, "loss_scale": 4.0, "retries": 0, "seconds": 531.1, "max_memory_gb": 25.46, "pair": 0.2862518262118101, "sign": 0.19114286825060844, "keep": 0.05578858591616154, "leak": 0.6154048973694444, "replacement": 0.5635688006877899}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 7.605839252471924, "loss_scale": 4.0, "retries": 0, "seconds": 532.3, "max_memory_gb": 25.46, "pair": 0.5639841370284557, "sign": 0.6356123276054859, "keep": 0.3794150557368994, "leak": 0.37037335336208344, "replacement": 1.16409120708704}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 7.752927303314209, "loss_scale": 4.0, "retries": 0, "seconds": 533.6, "max_memory_gb": 25.46, "pair": 0.290283540263772, "sign": 0.2413501963019371, "keep": 0.16368858609348536, "leak": 2.1966274264268577, "replacement": 1.0419598072767258}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 5.857110500335693, "loss_scale": 4.0, "retries": 0, "seconds": 534.9, "max_memory_gb": 25.46, "pair": 0.21920248121023178, "sign": 0.12235993333160877, "keep": 0.017015010700561106, "leak": 0.218377290351782, "replacement": 0.3434842824935913}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 5.285023212432861, "loss_scale": 4.0, "retries": 0, "seconds": 536.1, "max_memory_gb": 25.46, "pair": 0.1902116946876049, "sign": 0.0633827056735754, "keep": 0.03747523861238733, "leak": 0.5032729748636484, "replacement": 0.3664589002728462}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 4.939691066741943, "loss_scale": 4.0, "retries": 0, "seconds": 537.3, "max_memory_gb": 25.46, "pair": 0.24159464798867702, "sign": 0.2492311056703329, "keep": 0.21923444932326674, "leak": 0.5799146573990583, "replacement": 0.6208060830831528}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 7.336266994476318, "loss_scale": 4.0, "retries": 0, "seconds": 538.6, "max_memory_gb": 25.46, "pair": 0.10789913032203913, "sign": 0.05772109376266599, "keep": 0.16998506305390038, "leak": 1.5003013228997588, "replacement": 0.5968275517225266}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 4.946091651916504, "loss_scale": 4.0, "retries": 0, "seconds": 539.9, "max_memory_gb": 25.46, "pair": 0.14540748298168182, "sign": 0.17598143965005875, "keep": 0.12626104382798076, "leak": 0.31923653604462743, "replacement": 0.37633785977959633}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 6.932140350341797, "loss_scale": 4.0, "retries": 0, "seconds": 541.2, "max_memory_gb": 25.46, "pair": 0.15398742444813251, "sign": 0.22188489884138107, "keep": 0.05737925763241947, "leak": 1.3266968727111816, "replacement": 0.6252937167882919}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 4.164759159088135, "loss_scale": 4.0, "retries": 0, "seconds": 542.5, "max_memory_gb": 25.46, "pair": 0.15972433798015118, "sign": 0.12739324569702148, "keep": 0.0204228128422983, "leak": 0.20469038002192974, "replacement": 0.28480496630072594}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 14.196515083312988, "loss_scale": 4.0, "retries": 0, "seconds": 543.9, "max_memory_gb": 25.46, "pair": 0.4490805473178625, "sign": 0.45481074042618275, "keep": 0.571601735195145, "leak": 0.6676262483815663, "replacement": 1.1291933227330446}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 8.103026390075684, "loss_scale": 4.0, "retries": 0, "seconds": 545.2, "max_memory_gb": 25.46, "pair": 0.27310641668736935, "sign": 0.27642732486128807, "keep": 0.17013105656951666, "leak": 1.061071440577507, "replacement": 0.7616534531116486}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 15.528787612915039, "loss_scale": 4.0, "retries": 0, "seconds": 546.5, "max_memory_gb": 25.46, "pair": 0.3241694774478674, "sign": 0.5508868684992194, "keep": 0.7748735975474119, "leak": 1.393187753856182, "replacement": 1.3353466838598251}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 3.3469319343566895, "loss_scale": 4.0, "retries": 0, "seconds": 547.8, "max_memory_gb": 25.46, "pair": 0.11276690009981394, "sign": 0.07709538890048862, "keep": 0.06776753772282973, "leak": 0.5019436860457063, "replacement": 0.31068428605794907}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 5.428112983703613, "loss_scale": 4.0, "retries": 0, "seconds": 549.0, "max_memory_gb": 25.46, "pair": 0.40518438071012497, "sign": 0.42443254776299, "keep": 0.41459768638014793, "leak": 0.30682914005592465, "replacement": 0.9014067873358727}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 9.241846084594727, "loss_scale": 4.0, "retries": 0, "seconds": 550.3, "max_memory_gb": 25.46, "pair": 0.26842164993286133, "sign": 0.37724684923887253, "keep": 0.02316460246220231, "leak": 0.5820622788742185, "replacement": 0.6141429394483566}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 22.767990112304688, "loss_scale": 4.0, "retries": 0, "seconds": 551.5, "max_memory_gb": 25.46, "pair": 0.7326114922761917, "sign": 0.8687730878591537, "keep": 1.1877267025556648, "leak": 1.4545400142669678, "replacement": 2.1244963631033897}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 7.704097270965576, "loss_scale": 4.0, "retries": 0, "seconds": 552.7, "max_memory_gb": 25.46, "pair": 0.40391951985657215, "sign": 0.4280605036765337, "keep": 0.10651617776602507, "leak": 2.032799605280161, "replacement": 1.1794077455997467}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 10.553803443908691, "loss_scale": 4.0, "retries": 0, "seconds": 554.0, "max_memory_gb": 25.46, "pair": 0.23089931532740593, "sign": 0.3017386496067047, "keep": 0.06927347474265844, "leak": 0.8969916421920061, "replacement": 0.6406532973051071}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 6.658617973327637, "loss_scale": 4.0, "retries": 0, "seconds": 555.3, "max_memory_gb": 25.46, "pair": 0.25071093440055847, "sign": 0.3910521576181054, "keep": 0.2344354975502938, "leak": 0.35692630987614393, "replacement": 0.6526863351464272}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 6.468037128448486, "loss_scale": 4.0, "retries": 0, "seconds": 556.5, "max_memory_gb": 25.46, "pair": 0.6394182704389095, "sign": 0.7980564441531897, "keep": 0.6936569414101541, "leak": 0.7856373633258045, "replacement": 1.5816843584179878}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 6.4308624267578125, "loss_scale": 4.0, "retries": 0, "seconds": 557.7, "max_memory_gb": 25.46, "pair": 0.15820514969527721, "sign": 0.20170604903250933, "keep": 0.22192406130488962, "leak": 0.8352261901600286, "replacement": 0.5788267478346825}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 8.517395973205566, "loss_scale": 4.0, "retries": 0, "seconds": 559.0, "max_memory_gb": 25.46, "pair": 0.4107420900836587, "sign": 0.5489633027464151, "keep": 0.4393623590003699, "leak": 2.124314741056878, "replacement": 1.4359835237264633}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 6.794530391693115, "loss_scale": 4.0, "retries": 0, "seconds": 560.2, "max_memory_gb": 25.46, "pair": 0.2051963247358799, "sign": 0.24026467092335224, "keep": 0.26992585766129196, "leak": 0.46413245087023824, "replacement": 0.5763247087597847}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 5.373289108276367, "loss_scale": 4.0, "retries": 0, "seconds": 561.5, "max_memory_gb": 25.46, "pair": 0.11467359960079193, "sign": 0.145871726796031, "keep": 0.04039878991898149, "leak": 1.113339900970459, "replacement": 0.486143846064806}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 8.766640663146973, "loss_scale": 4.0, "retries": 0, "seconds": 562.7, "max_memory_gb": 25.46, "pair": 0.23096866719424725, "sign": 0.13002624083310366, "keep": 0.3965237506781705, "leak": 1.520128783187829, "replacement": 0.8742758333683014}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 5.287772178649902, "loss_scale": 4.0, "retries": 0, "seconds": 564.0, "max_memory_gb": 25.46, "pair": 0.15711969323456287, "sign": 0.1366134062409401, "keep": 0.04553876619320363, "leak": 0.8564180135726929, "replacement": 0.46230029687285423}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 6.065706729888916, "loss_scale": 4.0, "retries": 0, "seconds": 565.2, "max_memory_gb": 25.46, "pair": 0.35828207805752754, "sign": 0.3194994144141674, "keep": 0.426569826900959, "leak": 1.350315686315298, "replacement": 1.0688956379890442}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 17.675622940063477, "loss_scale": 4.0, "retries": 0, "seconds": 566.5, "max_memory_gb": 25.46, "pair": 0.3393172174692154, "sign": 0.3072113674134016, "keep": 0.060044095385819674, "leak": 0.9165442325174809, "replacement": 0.7520810142159462}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 11.163031578063965, "loss_scale": 4.0, "retries": 0, "seconds": 567.7, "max_memory_gb": 25.46, "pair": 0.3610015846788883, "sign": 0.26169524528086185, "keep": 0.4829341843724251, "leak": 1.198641799390316, "replacement": 1.0329767540097237}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 6.316989898681641, "loss_scale": 4.0, "retries": 0, "seconds": 569.0, "max_memory_gb": 25.46, "pair": 0.19369961135089397, "sign": 0.21648425050079823, "keep": 0.2732120882719755, "leak": 1.7661636471748352, "replacement": 0.8800886571407318}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 8.820100784301758, "loss_scale": 4.0, "retries": 0, "seconds": 570.2, "max_memory_gb": 25.46, "pair": 0.3618841767311096, "sign": 0.7051505115814507, "keep": 0.06634767726063728, "leak": 0.9488866712199524, "replacement": 0.984854981303215}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 5.59903621673584, "loss_scale": 4.0, "retries": 0, "seconds": 571.5, "max_memory_gb": 25.46, "pair": 0.15351484715938568, "sign": 0.2328601386398077, "keep": 0.014239949276088737, "leak": 0.4611574038863182, "replacement": 0.39235424622893333}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 16.170658111572266, "loss_scale": 4.0, "retries": 0, "seconds": 572.7, "max_memory_gb": 25.46, "pair": 0.3211185000836849, "sign": 0.44710971415042877, "keep": 0.3443647753447294, "leak": 1.1345145006380335, "replacement": 1.000484362244606}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 5.053550720214844, "loss_scale": 4.0, "retries": 0, "seconds": 574.0, "max_memory_gb": 25.46, "pair": 0.23987418226897717, "sign": 0.2640765327960253, "keep": 0.053906526998616755, "leak": 0.31263499055057764, "replacement": 0.47702446579933167}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 4.774438858032227, "loss_scale": 4.0, "retries": 0, "seconds": 575.3, "max_memory_gb": 25.46, "pair": 0.24188606441020966, "sign": 0.2706908229738474, "keep": 0.041289061307907104, "leak": 0.25217268615961075, "replacement": 0.46091917157173157}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 14.390824317932129, "loss_scale": 4.0, "retries": 0, "seconds": 576.5, "max_memory_gb": 25.46, "pair": 0.2555265882983804, "sign": 0.6414171420037746, "keep": 0.8000909690745175, "leak": 1.4483282417058945, "replacement": 1.3383627235889435}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 4.02102518081665, "loss_scale": 4.0, "retries": 0, "seconds": 577.8, "max_memory_gb": 25.46, "pair": 0.1562588643282652, "sign": 0.19131837505847216, "keep": 0.018449447947205044, "leak": 0.571929108351469, "replacement": 0.40412504598498344}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 5.73765754699707, "loss_scale": 4.0, "retries": 0, "seconds": 579.0, "max_memory_gb": 25.46, "pair": 0.1868299813941121, "sign": 0.2003499288111925, "keep": 0.03771351705654524, "leak": 0.11073448532260954, "replacement": 0.3335453271865845}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 13.963183403015137, "loss_scale": 4.0, "retries": 0, "seconds": 580.3, "max_memory_gb": 25.46, "pair": 0.15075436420738697, "sign": 0.1661724904552102, "keep": 0.022394362778868526, "leak": 1.8813251480460167, "replacement": 0.7153690792620182}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 9.495767593383789, "loss_scale": 4.0, "retries": 0, "seconds": 581.6, "max_memory_gb": 25.46, "pair": 0.09681737702339888, "sign": 0.12988494150340557, "keep": 0.1243088599294424, "leak": 1.8063098639249802, "replacement": 0.6754917576909065}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 13.748888969421387, "loss_scale": 4.0, "retries": 0, "seconds": 582.8, "max_memory_gb": 25.46, "pair": 0.15532566653564572, "sign": 0.20583471050485969, "keep": 0.1164812445640564, "leak": 1.332528853788972, "replacement": 0.6496158614754677}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 11.020374298095703, "loss_scale": 4.0, "retries": 0, "seconds": 584.1, "max_memory_gb": 25.46, "pair": 0.12143867090344429, "sign": 0.15716180112212896, "keep": 0.060457466868683696, "leak": 1.309925147332251, "replacement": 0.5577295795083046}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 5.972825050354004, "loss_scale": 4.0, "retries": 0, "seconds": 585.3, "max_memory_gb": 25.46, "pair": 0.136493518948555, "sign": 0.12071604514494538, "keep": 0.01823945698561147, "leak": 0.37565658800303936, "replacement": 0.29988541454076767}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 11.95910358428955, "loss_scale": 4.0, "retries": 0, "seconds": 586.6, "max_memory_gb": 25.46, "pair": 0.35525945387780666, "sign": 0.18607516214251518, "keep": 0.5165921696461737, "leak": 1.1587796164967585, "replacement": 0.9962880238890648}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 8.553791999816895, "loss_scale": 4.0, "retries": 0, "seconds": 587.8, "max_memory_gb": 25.46, "pair": 0.17419766820967197, "sign": 0.08749768603593111, "keep": 0.020652360166423023, "leak": 0.32975469596567564, "replacement": 0.3107113614678383}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 10.584920883178711, "loss_scale": 4.0, "retries": 0, "seconds": 589.1, "max_memory_gb": 25.46, "pair": 0.4119324404746294, "sign": 0.5585018545389175, "keep": 0.5584810553118587, "leak": 0.3730643317103386, "replacement": 1.0636899694800377}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 17.15692138671875, "loss_scale": 4.0, "retries": 0, "seconds": 590.3, "max_memory_gb": 25.46, "pair": 0.30046372301876545, "sign": 0.34285436756908894, "keep": 0.17592229042202234, "leak": 0.8333200663328171, "replacement": 0.768182098865509}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 10.290932655334473, "loss_scale": 4.0, "retries": 0, "seconds": 591.5, "max_memory_gb": 25.46, "pair": 0.18942144326865673, "sign": 0.13808652944862843, "keep": 0.22557849856093526, "leak": 0.27726079345120525, "replacement": 0.4405691400170326}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 9.258404731750488, "loss_scale": 4.0, "retries": 0, "seconds": 592.8, "max_memory_gb": 25.46, "pair": 0.31542997620999813, "sign": 0.42382980696856976, "keep": 0.38549865991808474, "leak": 0.2393593192100525, "replacement": 0.7799340225756168}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 13.504412651062012, "loss_scale": 4.0, "retries": 0, "seconds": 594.1, "max_memory_gb": 25.46, "pair": 0.24813663214445114, "sign": 0.4199214391410351, "keep": 0.46970894269179553, "leak": 0.5942985936999321, "replacement": 0.8415264673531055}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 7.516763210296631, "loss_scale": 4.0, "retries": 0, "seconds": 595.4, "max_memory_gb": 25.46, "pair": 0.23376214876770973, "sign": 0.19085466675460339, "keep": 0.08942804019898176, "leak": 1.699653036892414, "replacement": 0.7988167554140091}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 10.635259628295898, "loss_scale": 4.0, "retries": 0, "seconds": 596.6, "max_memory_gb": 25.46, "pair": 0.24136612564325333, "sign": 0.3746415195055306, "keep": 0.023325488495174795, "leak": 0.7086378559251898, "replacement": 0.6175091005861759}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 6.195342540740967, "loss_scale": 4.0, "retries": 0, "seconds": 597.8, "max_memory_gb": 25.46, "pair": 0.25095562264323235, "sign": 0.22454250836744905, "keep": 0.046161160222254694, "leak": 0.2717962050810456, "replacement": 0.4542565196752548}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 15.663986206054688, "loss_scale": 4.0, "retries": 0, "seconds": 599.1, "max_memory_gb": 25.46, "pair": 0.5851017171517015, "sign": 0.5279973484575748, "keep": 0.6167211333522573, "leak": 0.6966093615628779, "replacement": 1.3316133320331573}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 6.206048011779785, "loss_scale": 4.0, "retries": 0, "seconds": 600.4, "max_memory_gb": 25.46, "pair": 0.4151443541049957, "sign": 0.5201440434902906, "keep": 0.43479033559560776, "leak": 0.5229940968565643, "replacement": 1.0233601033687592}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 7.995371341705322, "loss_scale": 4.0, "retries": 0, "seconds": 601.6, "max_memory_gb": 25.46, "pair": 0.16018550004810095, "sign": 0.12152927881106734, "keep": 0.07284936250653118, "leak": 0.45851996168494225, "replacement": 0.37200480699539185}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 7.247772216796875, "loss_scale": 4.0, "retries": 0, "seconds": 602.9, "max_memory_gb": 25.46, "pair": 0.2506899759173393, "sign": 0.41929043270647526, "keep": 0.26088398415595293, "leak": 0.6282739476300776, "replacement": 0.7478456944227219}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 14.704791069030762, "loss_scale": 4.0, "retries": 0, "seconds": 604.1, "max_memory_gb": 25.46, "pair": 0.39952193945646286, "sign": 0.7090069577097893, "keep": 0.47246615774929523, "leak": 0.20942072081379592, "replacement": 1.0426136404275894}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 17.498947143554688, "loss_scale": 4.0, "retries": 0, "seconds": 605.4, "max_memory_gb": 25.46, "pair": 0.6504897698760033, "sign": 0.9618581645190716, "keep": 1.5149776823818684, "leak": 0.3174576684832573, "replacement": 1.9682721346616745}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 13.684887886047363, "loss_scale": 4.0, "retries": 0, "seconds": 606.6, "max_memory_gb": 25.46, "pair": 0.32467127637937665, "sign": 0.5585938766598701, "keep": 0.06788693577982485, "leak": 1.0537679120898247, "replacement": 0.9013536274433136}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 12.141219139099121, "loss_scale": 4.0, "retries": 0, "seconds": 607.8, "max_memory_gb": 25.46, "pair": 0.3200058192014694, "sign": 0.5838113054633141, "keep": 0.7372047400567681, "leak": 0.9210200668312609, "replacement": 1.2107688821852207}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 7.758179187774658, "loss_scale": 4.0, "retries": 0, "seconds": 609.0, "max_memory_gb": 25.46, "pair": 0.1455530822277069, "sign": 0.13356966525316238, "keep": 0.06428863760083914, "leak": 0.8918110388949572, "replacement": 0.46743498742580414}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 10.861918449401855, "loss_scale": 4.0, "retries": 0, "seconds": 610.3, "max_memory_gb": 25.46, "pair": 0.21067721582949162, "sign": 0.3304506530985236, "keep": 0.1824327171780169, "leak": 0.9019109606742859, "replacement": 0.6925966441631317}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 15.975656509399414, "loss_scale": 4.0, "retries": 0, "seconds": 611.5, "max_memory_gb": 25.46, "pair": 0.10440724343061447, "sign": 0.24972478672862053, "keep": 0.7343497591209598, "leak": 2.147748850286007, "replacement": 1.1333817318081856}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 12.187748908996582, "loss_scale": 4.0, "retries": 0, "seconds": 612.7, "max_memory_gb": 25.46, "pair": 0.15448191948235035, "sign": 0.19413496414199471, "keep": 0.02565929346019402, "leak": 1.046916700899601, "replacement": 0.5261082351207733}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 26.911832809448242, "loss_scale": 4.0, "retries": 0, "seconds": 614.0, "max_memory_gb": 25.46, "pair": 0.1263565761037171, "sign": 0.3215382620692253, "keep": 0.16742849012371153, "leak": 2.8514191322028637, "replacement": 1.0836947336792946}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 12.956449508666992, "loss_scale": 4.0, "retries": 0, "seconds": 615.3, "max_memory_gb": 25.46, "pair": 0.15222434513270855, "sign": 0.15059299115091562, "keep": 0.03566943798796274, "leak": 1.3421085104346275, "replacement": 0.5808826684951782}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 15.843085289001465, "loss_scale": 4.0, "retries": 0, "seconds": 616.5, "max_memory_gb": 25.46, "pair": 0.5077431201934814, "sign": 0.8547204211354256, "keep": 0.4914836304669734, "leak": 0.8697161749005318, "replacement": 1.3982741758227348}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 9.514325141906738, "loss_scale": 4.0, "retries": 0, "seconds": 617.8, "max_memory_gb": 25.46, "pair": 0.06231064070016146, "sign": 0.1859345061238855, "keep": 0.028611502377316356, "leak": 0.7995182612794451, "replacement": 0.3694631978869438}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 11.352274894714355, "loss_scale": 4.0, "retries": 0, "seconds": 619.0, "max_memory_gb": 25.46, "pair": 0.16306837182492018, "sign": 0.22643251297995448, "keep": 0.059956374345347285, "leak": 1.2635032534599304, "replacement": 0.6221386268734932}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 6.803197860717773, "loss_scale": 4.0, "retries": 0, "seconds": 620.3, "max_memory_gb": 25.46, "pair": 0.13028803281486034, "sign": 0.06406894978135824, "keep": 0.07946950918994844, "leak": 1.0702613797038794, "replacement": 0.4696226119995117}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 40.54663848876953, "loss_scale": 4.0, "retries": 0, "seconds": 621.5, "max_memory_gb": 25.46, "pair": 0.37424415722489357, "sign": 0.7209047600626945, "keep": 1.261728672485333, "leak": 1.70408128015697, "replacement": 1.7915811836719513}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 9.69116497039795, "loss_scale": 4.0, "retries": 0, "seconds": 622.7, "max_memory_gb": 25.46, "pair": 0.36458129808306694, "sign": 0.3666997104883194, "keep": 0.2120187897235155, "leak": 1.1793709489284083, "replacement": 0.9487832710146904}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 14.015608787536621, "loss_scale": 4.0, "retries": 0, "seconds": 623.9, "max_memory_gb": 25.46, "pair": 0.39026420284062624, "sign": 0.38612684700638056, "keep": 0.47523044294212013, "leak": 0.515100316493772, "replacement": 0.949717927724123}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 20.182357788085938, "loss_scale": 4.0, "retries": 0, "seconds": 625.2, "max_memory_gb": 25.46, "pair": 0.46892669051885605, "sign": 0.6122697591781616, "keep": 0.8953223808202893, "leak": 0.5957256923575187, "replacement": 1.3716542199254036}
diff --git a/runs/single_llava/replacement/L10_s13/adapter/README.md b/runs/single_llava/replacement/L10_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L10_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L10_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L10_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..660b51e08f84d629a9e3b45cf008027d199f3303
--- /dev/null
+++ b/runs/single_llava/replacement/L10_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.10.self_attn.o_proj",
+    "model.language_model.layers.10.mlp.gate_proj",
+    "model.language_model.layers.10.self_attn.k_proj",
+    "model.language_model.layers.10.mlp.down_proj",
+    "model.language_model.layers.10.self_attn.v_proj",
+    "model.language_model.layers.10.mlp.up_proj",
+    "model.language_model.layers.10.self_attn.q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L10_s13/config.json b/runs/single_llava/replacement/L10_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..62a16b367a24d7968da18de4d780c78ac042e7e0
--- /dev/null
+++ b/runs/single_llava/replacement/L10_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 10,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L10_s13/steps.jsonl b/runs/single_llava/replacement/L10_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a33a0d31ab3c1a7b87071b3ada11537db69a3aa4
--- /dev/null
+++ b/runs/single_llava/replacement/L10_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.6421976089477539, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 24.43, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.8989347815513611, "loss_scale": 4.0, "retries": 0, "seconds": 2.6, "max_memory_gb": 24.43, "pair": 1.097475916147232, "sign": 0.5365806855261326, "keep": 5.376199863385409, "leak": 1.1840443111310606e-05, "replacement": 4.053869187831879}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.3758400678634644, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 24.52, "pair": 1.0925927460193634, "sign": 0.5399863515049219, "keep": 10.801673650741577, "leak": 6.97355440024694e-05, "replacement": 6.763440370559692}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 1.1909126043319702, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 24.54, "pair": 1.0909123420715332, "sign": 0.47037837468087673, "keep": 7.668957281857729, "leak": 0.00017053311336212573, "replacement": 5.160622805356979}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.6874906420707703, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 24.54, "pair": 1.0951424837112427, "sign": 0.7751915603876114, "keep": 0.1586596965789795, "leak": 9.591236633355038e-06, "replacement": 1.5620705485343933}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.7116218209266663, "loss_scale": 4.0, "retries": 0, "seconds": 7.6, "max_memory_gb": 24.54, "pair": 1.0904470682144165, "sign": 0.73518256098032, "keep": 2.9299242421984673, "leak": 7.214821146650507e-05, "replacement": 2.923018604516983}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 1.3535466194152832, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 24.54, "pair": 1.0684316754341125, "sign": 0.5457584010437131, "keep": 6.869886350817978, "leak": 0.0009726448170113144, "replacement": 4.7764972448349}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 1.240414023399353, "loss_scale": 4.0, "retries": 0, "seconds": 10.1, "max_memory_gb": 24.54, "pair": 1.075718343257904, "sign": 0.5841367691755295, "keep": 7.375524342060089, "leak": 0.0005166335600961247, "replacement": 5.055678188800812}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 1.2119890451431274, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 24.54, "pair": 1.0619597733020782, "sign": 0.6171679301187396, "keep": 5.701266288757324, "leak": 0.0019205842833811548, "replacement": 4.2216571271419525}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.9984360337257385, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 24.54, "pair": 1.0717637538909912, "sign": 0.777775876224041, "keep": 3.270669564604759, "leak": 0.000991064229310723, "replacement": 3.0962342023849487}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 1.527964472770691, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 24.54, "pair": 1.0379591286182404, "sign": 0.4498559758067131, "keep": 5.250334408134222, "leak": 0.004393219827761641, "replacement": 3.889152765274048}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 1.3263143301010132, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 24.54, "pair": 1.0492450445890427, "sign": 0.6054352186620235, "keep": 3.0385120483115315, "leak": 0.006236448036361253, "replacement": 2.8727777898311615}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 1.2247421741485596, "loss_scale": 4.0, "retries": 0, "seconds": 16.2, "max_memory_gb": 24.54, "pair": 1.053802102804184, "sign": 0.5015069991350174, "keep": 5.332975733093917, "leak": 0.00368140797228822, "replacement": 3.9719639122486115}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.8493744134902954, "loss_scale": 4.0, "retries": 0, "seconds": 17.4, "max_memory_gb": 24.54, "pair": 1.0115599036216736, "sign": 0.391393706202507, "keep": 7.75747412443161, "leak": 0.005989511457300978, "replacement": 5.087491184473038}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 1.1973263025283813, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 24.54, "pair": 1.043030470609665, "sign": 0.6883252412080765, "keep": 5.151364832418039, "leak": 0.004496322573686484, "replacement": 3.9639995992183685}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 1.1855002641677856, "loss_scale": 4.0, "retries": 0, "seconds": 19.8, "max_memory_gb": 24.54, "pair": 1.0567747056484222, "sign": 0.778691291809082, "keep": 2.7447689480613917, "leak": 0.003105263807810843, "replacement": 2.8192811608314514}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 1.0067344903945923, "loss_scale": 4.0, "retries": 0, "seconds": 21.0, "max_memory_gb": 24.54, "pair": 1.042281061410904, "sign": 0.7006261944770813, "keep": 3.161676754243672, "leak": 0.010431813469040208, "replacement": 2.9760403633117676}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.782070517539978, "loss_scale": 4.0, "retries": 0, "seconds": 22.3, "max_memory_gb": 24.6, "pair": 1.0203298926353455, "sign": 0.5220812186598778, "keep": 7.117829725146294, "leak": 0.005561667057918385, "replacement": 4.841675728559494}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 2.5760464668273926, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 24.6, "pair": 0.9436760544776917, "sign": 0.5132379159331322, "keep": 10.631195068359375, "leak": 0.026807887479662895, "replacement": 6.522594571113586}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.8871542811393738, "loss_scale": 4.0, "retries": 0, "seconds": 24.7, "max_memory_gb": 24.6, "pair": 1.0689516365528107, "sign": 0.42561232298612595, "keep": 0.02229927876032889, "leak": 0.000887745525687933, "replacement": 1.2931293845176697}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.630619764328003, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 24.6, "pair": 1.0002976208925247, "sign": 0.36343479715287685, "keep": 5.501697952859104, "leak": 0.0067152845149394125, "replacement": 3.934542715549469}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.6127679347991943, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 24.6, "pair": 0.9805646538734436, "sign": 0.5604230910539627, "keep": 5.8760304525494576, "leak": 0.03196353037492372, "replacement": 4.20678237080574}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.8265204429626465, "loss_scale": 4.0, "retries": 0, "seconds": 28.4, "max_memory_gb": 24.6, "pair": 0.9268487840890884, "sign": 0.23603160213679075, "keep": 5.282213832717389, "leak": 0.03313153712952044, "replacement": 3.6942544281482697}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 3.0496058464050293, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 24.6, "pair": 0.8780990540981293, "sign": 0.23267133347690105, "keep": 9.945882558822632, "leak": 0.030997057794593275, "replacement": 5.975125193595886}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 2.28716778755188, "loss_scale": 4.0, "retries": 0, "seconds": 30.9, "max_memory_gb": 24.6, "pair": 0.8959072530269623, "sign": 0.37330025248229504, "keep": 8.373201929032803, "leak": 0.04560961732204305, "replacement": 5.28056076169014}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.977276563644409, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 24.6, "pair": 0.8404591530561447, "sign": 0.6450533773750067, "keep": 10.364725112915039, "leak": 0.07813446782529354, "replacement": 6.364881873130798}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 2.364788293838501, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 24.6, "pair": 0.8868348896503448, "sign": 0.5641890317201614, "keep": 8.30210231244564, "leak": 0.053037396690342575, "replacement": 5.333240032196045}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.8094367980957031, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 24.6, "pair": 0.9756792932748795, "sign": 0.7579032704234123, "keep": 5.545539289712906, "leak": 0.03571522911715874, "replacement": 4.136329412460327}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.559699296951294, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 24.6, "pair": 0.9895625412464142, "sign": 0.8382462412118912, "keep": 5.16819217056036, "leak": 0.024593457046648837, "replacement": 3.9989301562309265}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.9163103103637695, "loss_scale": 4.0, "retries": 0, "seconds": 37.0, "max_memory_gb": 24.6, "pair": 0.8455082774162292, "sign": 0.3331757131963968, "keep": 5.430893238633871, "leak": 0.10525299495930085, "replacement": 3.7538560330867767}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.897806167602539, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 24.6, "pair": 0.9237860441207886, "sign": 0.795178109779954, "keep": 5.580069646239281, "leak": 0.08675227107687533, "replacement": 4.133098125457764}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 3.26993465423584, "loss_scale": 4.0, "retries": 0, "seconds": 39.5, "max_memory_gb": 24.6, "pair": 0.827374130487442, "sign": 0.3524563033133745, "keep": 10.53848385810852, "leak": 0.1031896835193038, "replacement": 6.298641562461853}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 2.1592066287994385, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 24.6, "pair": 0.8929412662982941, "sign": 0.5945987645536661, "keep": 4.7911385940387845, "leak": 0.14268990616801602, "replacement": 3.6214822828769684}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 2.0679118633270264, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 24.6, "pair": 0.9010877162218094, "sign": 0.4966788850724697, "keep": 5.171163000166416, "leak": 0.03719046915648505, "replacement": 3.7443061769008636}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 2.310605525970459, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 24.6, "pair": 0.7957908362150192, "sign": 0.4103595046326518, "keep": 5.338777953758836, "leak": 0.12976697320118546, "replacement": 3.702801376581192}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 2.053135633468628, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 24.6, "pair": 0.815217986702919, "sign": 0.41071413177996874, "keep": 5.053053747862577, "leak": 0.10597086092457175, "replacement": 3.573594808578491}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 2.3606557846069336, "loss_scale": 4.0, "retries": 0, "seconds": 45.7, "max_memory_gb": 24.6, "pair": 0.8204848915338516, "sign": 0.5607962943613529, "keep": 4.755456380546093, "leak": 0.16294164964347146, "replacement": 3.519346684217453}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.6562676429748535, "loss_scale": 4.0, "retries": 0, "seconds": 47.0, "max_memory_gb": 24.6, "pair": 0.7461231052875519, "sign": 0.7064758576452732, "keep": 9.197211384773254, "leak": 0.1183572297450155, "replacement": 5.727556109428406}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.9507498741149902, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 24.6, "pair": 0.9249256551265717, "sign": 0.541202874854207, "keep": 2.721752239856869, "leak": 0.1327212644428073, "replacement": 2.589583456516266}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.5618481636047363, "loss_scale": 4.0, "retries": 0, "seconds": 49.5, "max_memory_gb": 24.6, "pair": 0.8195400834083557, "sign": 0.45966317877173424, "keep": 5.403389483690262, "leak": 0.0884721155744046, "replacement": 3.7731845378875732}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.24345064163208, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 24.6, "pair": 1.0009499192237854, "sign": 1.0023772567510605, "keep": 2.89172849804163, "leak": 0.04817930243734736, "replacement": 2.960047632455826}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.873577833175659, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 24.6, "pair": 0.8647009134292603, "sign": 0.37274814397096634, "keep": 4.377170855179429, "leak": 0.03826055442914367, "replacement": 3.249225616455078}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.8773698806762695, "loss_scale": 4.0, "retries": 0, "seconds": 53.2, "max_memory_gb": 24.6, "pair": 0.8402177691459656, "sign": 0.4828266389667988, "keep": 4.587333859875798, "leak": 0.06633886322379112, "replacement": 3.391882687807083}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.7347151041030884, "loss_scale": 4.0, "retries": 0, "seconds": 54.4, "max_memory_gb": 24.6, "pair": 0.8734854608774185, "sign": 0.5423956550657749, "keep": 2.6565486658364534, "leak": 0.09676231653429568, "replacement": 2.4971481263637543}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.6298233270645142, "loss_scale": 4.0, "retries": 0, "seconds": 55.6, "max_memory_gb": 24.6, "pair": 0.9598273485898972, "sign": 0.8642075061798096, "keep": 2.6840424686670303, "leak": 0.05028023666045556, "replacement": 2.746522396802902}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.975637435913086, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 24.6, "pair": 0.8534768372774124, "sign": 0.670278038829565, "keep": 4.307578220963478, "leak": 0.10231560206739232, "replacement": 3.367983788251877}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 3.1189630031585693, "loss_scale": 4.0, "retries": 0, "seconds": 58.1, "max_memory_gb": 24.6, "pair": 0.7355737537145615, "sign": 0.5853907987475395, "keep": 3.788450830616057, "leak": 0.13596206990769133, "replacement": 2.956485092639923}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 4.951417446136475, "loss_scale": 4.0, "retries": 0, "seconds": 59.3, "max_memory_gb": 24.6, "pair": 0.6509206593036652, "sign": 0.17736892867833376, "keep": 6.974865645170212, "leak": 0.24593130638822913, "replacement": 4.288520723581314}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 4.457693099975586, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 24.6, "pair": 0.6614544466137886, "sign": 0.5973236570134759, "keep": 6.545231085270643, "leak": 0.3337720683775842, "replacement": 4.31617471575737}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 5.039570331573486, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 24.6, "pair": 0.7857729345560074, "sign": 0.39411547034978867, "keep": 6.493909923359752, "leak": 0.12767989607527852, "replacement": 4.261705756187439}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 2.0287179946899414, "loss_scale": 4.0, "retries": 0, "seconds": 63.0, "max_memory_gb": 24.6, "pair": 0.8495683073997498, "sign": 0.5476835463196039, "keep": 2.4732469273731112, "leak": 0.13790157187031582, "replacement": 2.3945089280605316}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 3.7771503925323486, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 24.6, "pair": 0.7314356118440628, "sign": 0.6450127689167857, "keep": 4.196417607367039, "leak": 0.27851075027137995, "replacement": 3.221778452396393}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 5.041511535644531, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 24.6, "pair": 0.7729276716709137, "sign": 0.672648087143898, "keep": 6.6971278339624405, "leak": 0.2986254126881249, "replacement": 4.532472014427185}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 6.042774677276611, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 24.6, "pair": 0.6606491431593895, "sign": 0.5531095042824745, "keep": 6.880231082439423, "leak": 0.33355508209206164, "replacement": 4.460708260536194}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 4.370914459228516, "loss_scale": 4.0, "retries": 0, "seconds": 67.8, "max_memory_gb": 24.6, "pair": 0.7062018439173698, "sign": 0.35523049533367157, "keep": 4.059852181002498, "leak": 0.2730846470221877, "replacement": 2.9820142090320587}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 9.003718376159668, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 24.6, "pair": 0.41526078432798386, "sign": 0.12226679362356663, "keep": 7.711905002593994, "leak": 0.45951858907938004, "replacement": 4.447226285934448}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 2.4630024433135986, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 24.6, "pair": 0.8501890823245049, "sign": 0.777817290276289, "keep": 2.0708434395492077, "leak": 0.060627290047705173, "replacement": 2.2896762788295746}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 5.070964813232422, "loss_scale": 4.0, "retries": 0, "seconds": 71.6, "max_memory_gb": 24.6, "pair": 0.7438448294997215, "sign": 0.24693654291331768, "keep": 3.821982942055911, "leak": 0.3800389305688441, "replacement": 2.873314321041107}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 3.048515796661377, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 24.6, "pair": 0.8617843836545944, "sign": 0.6171279773116112, "keep": 2.4803447276353836, "leak": 0.0275758304487681, "replacement": 2.4174146354198456}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 5.291022300720215, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 24.6, "pair": 0.6440880745649338, "sign": 0.5897174514830112, "keep": 3.7280424805358052, "leak": 0.2596176927909255, "replacement": 2.8678725361824036}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 7.772779941558838, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 24.6, "pair": 0.6541186049580574, "sign": 0.646246301010251, "keep": 5.1164374351501465, "leak": 0.48896872537443414, "replacement": 3.657702684402466}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 2.944521427154541, "loss_scale": 4.0, "retries": 0, "seconds": 76.5, "max_memory_gb": 24.6, "pair": 0.836679145693779, "sign": 0.6667499486356974, "keep": 2.249608291313052, "leak": 0.07293502797256224, "replacement": 2.313092052936554}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 8.443150520324707, "loss_scale": 4.0, "retries": 0, "seconds": 77.7, "max_memory_gb": 24.6, "pair": 0.6865383982658386, "sign": 0.522952489554882, "keep": 5.701299364678562, "leak": 0.17127480823546648, "replacement": 3.84148308634758}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 9.321531295776367, "loss_scale": 4.0, "retries": 0, "seconds": 79.0, "max_memory_gb": 24.64, "pair": 0.5268938466906548, "sign": 0.35175582859665155, "keep": 4.722165875136852, "leak": 0.8458314528688788, "replacement": 3.2753124833106995}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 8.663352012634277, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 24.64, "pair": 0.6970313191413879, "sign": 0.5006173197180033, "keep": 5.048815339803696, "leak": 0.32201124634593725, "replacement": 3.5522505044937134}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 3.272129535675049, "loss_scale": 4.0, "retries": 0, "seconds": 81.6, "max_memory_gb": 24.64, "pair": 0.9439729005098343, "sign": 0.5755694210529327, "keep": 1.4136592831928283, "leak": 0.08175954894977622, "replacement": 1.9590271413326263}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 8.767812728881836, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 24.64, "pair": 0.5361047573387623, "sign": 0.3606999069452286, "keep": 3.765684127807617, "leak": 0.9652647051261738, "replacement": 2.8406129479408264}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 3.9754836559295654, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 24.64, "pair": 0.869666263461113, "sign": 0.48416705057024956, "keep": 1.7401626182254404, "leak": 0.1345531539991498, "replacement": 2.0154693722724915}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 7.052104949951172, "loss_scale": 4.0, "retries": 0, "seconds": 85.3, "max_memory_gb": 24.64, "pair": 0.7190051823854446, "sign": 0.6931453254073858, "keep": 2.6984073910862207, "leak": 0.11278081813361496, "replacement": 2.442976802587509}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 6.600144386291504, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 24.64, "pair": 0.7429098784923553, "sign": 0.5088870413601398, "keep": 2.2755861673504114, "leak": 0.5934780091047287, "replacement": 2.2835159301757812}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 10.9645414352417, "loss_scale": 4.0, "retries": 0, "seconds": 87.8, "max_memory_gb": 24.64, "pair": 0.4963243380188942, "sign": 0.12836123071610928, "keep": 3.248472456820309, "leak": 0.5054793632298242, "replacement": 2.3111109733581543}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 12.951584815979004, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 24.64, "pair": 0.3689776435494423, "sign": 0.18892737850546837, "keep": 3.557925134897232, "leak": 0.7088339738547802, "replacement": 2.4196122884750366}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.7146985530853271, "loss_scale": 4.0, "retries": 0, "seconds": 90.2, "max_memory_gb": 24.64, "pair": 0.9067521542310715, "sign": 0.7554513365030289, "keep": 0.3819567933678627, "leak": 0.043637244307319634, "replacement": 1.4863655269145966}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 7.045468807220459, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 24.64, "pair": 0.7310297824442387, "sign": 0.4731356902047992, "keep": 1.619953524786979, "leak": 0.34572083010971255, "replacement": 1.864004522562027}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.5504746437072754, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 24.64, "pair": 0.7958417683839798, "sign": 0.8724753009155393, "keep": 2.1834366843104362, "leak": 0.4334975515957922, "replacement": 2.432172119617462}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 6.400834083557129, "loss_scale": 4.0, "retries": 0, "seconds": 93.9, "max_memory_gb": 24.64, "pair": 0.5390903800725937, "sign": 0.38812213484197855, "keep": 1.2954155765473843, "leak": 0.8995697107166052, "replacement": 1.6057516932487488}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 9.738636016845703, "loss_scale": 4.0, "retries": 0, "seconds": 95.1, "max_memory_gb": 24.64, "pair": 0.6086395531892776, "sign": 0.4242438580840826, "keep": 2.6380624771118164, "leak": 1.3408775143325329, "replacement": 2.4750121235847473}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 5.47092342376709, "loss_scale": 4.0, "retries": 0, "seconds": 96.3, "max_memory_gb": 24.64, "pair": 0.7391854971647263, "sign": 0.6981657706201077, "keep": 1.2744252420961857, "leak": 0.7574763000011444, "replacement": 1.9148500859737396}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 3.408005475997925, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 24.64, "pair": 0.7615154050290585, "sign": 0.53883882984519, "keep": 0.45738967321813107, "leak": 0.28684816823806614, "replacement": 1.3313416838645935}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 6.025548458099365, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 24.64, "pair": 0.635550320148468, "sign": 0.3650819528847933, "keep": 0.9099162025377154, "leak": 0.9959999276325107, "replacement": 1.522049367427826}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 8.40108871459961, "loss_scale": 4.0, "retries": 0, "seconds": 99.9, "max_memory_gb": 24.64, "pair": 0.418486513197422, "sign": 0.37954227440059185, "keep": 1.3017429411411285, "leak": 1.469175432343036, "replacement": 1.626422941684723}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 7.015248775482178, "loss_scale": 4.0, "retries": 0, "seconds": 101.1, "max_memory_gb": 24.64, "pair": 0.4793262016028166, "sign": 0.36958035035058856, "keep": 0.8090613931417465, "leak": 0.9929690451972419, "replacement": 1.3168893456459045}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 5.3138861656188965, "loss_scale": 4.0, "retries": 0, "seconds": 102.3, "max_memory_gb": 24.64, "pair": 0.5327020063996315, "sign": 0.3074037432670593, "keep": 0.6935408147983253, "leak": 1.0786230247467756, "replacement": 1.3028300404548645}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 2.9556407928466797, "loss_scale": 4.0, "retries": 0, "seconds": 103.6, "max_memory_gb": 24.64, "pair": 0.787309505045414, "sign": 0.6377087607979774, "keep": 0.45951302582398057, "leak": 0.9010836537927389, "replacement": 1.5611912906169891}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 2.3176729679107666, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 24.64, "pair": 0.9433252662420273, "sign": 0.5586091727018356, "keep": 0.09852783451788127, "leak": 0.055522319627925754, "replacement": 1.2857743501663208}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 6.555110454559326, "loss_scale": 4.0, "retries": 0, "seconds": 106.1, "max_memory_gb": 24.64, "pair": 0.5319293364882469, "sign": 0.636379774659872, "keep": 0.6974851861596107, "leak": 0.5053375049174065, "replacement": 1.3251961916685104}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 3.2381651401519775, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 24.64, "pair": 0.4592975713312626, "sign": 0.29614201933145523, "keep": 0.28579759038984776, "leak": 2.6628175127552822, "replacement": 1.4159717857837677}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.3001320362091064, "loss_scale": 4.0, "retries": 0, "seconds": 108.5, "max_memory_gb": 24.64, "pair": 0.68980473279953, "sign": 0.6149888541549444, "keep": 0.42970402888022363, "leak": 0.6274861092679203, "replacement": 1.369022697210312}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 2.4669852256774902, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 24.64, "pair": 0.9791116863489151, "sign": 0.698834978044033, "keep": 0.22830859059467912, "leak": 0.0314010304864496, "replacement": 1.4505336582660675}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.50168776512146, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 24.64, "pair": 0.5620428249239922, "sign": 0.45086556673049927, "keep": 0.33383319480344653, "leak": 0.6881953834090382, "replacement": 1.1264410763978958}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.9580161571502686, "loss_scale": 4.0, "retries": 0, "seconds": 112.1, "max_memory_gb": 24.64, "pair": 0.8778165429830551, "sign": 0.9313306957483292, "keep": 0.5664404518902302, "leak": 0.08443932863883674, "replacement": 1.6478119194507599}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.9951814413070679, "loss_scale": 4.0, "retries": 0, "seconds": 113.4, "max_memory_gb": 24.64, "pair": 0.8443828374147415, "sign": 0.4972860887646675, "keep": 0.20699381781741977, "leak": 1.5192812457680702, "replacement": 1.5763430893421173}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 2.2106523513793945, "loss_scale": 4.0, "retries": 0, "seconds": 114.6, "max_memory_gb": 24.64, "pair": 0.5040759034454823, "sign": 0.25073669850826263, "keep": 0.21680764365009964, "leak": 0.3942713588476181, "replacement": 0.8364159017801285}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 2.4014761447906494, "loss_scale": 4.0, "retries": 0, "seconds": 115.8, "max_memory_gb": 24.64, "pair": 0.37684375792741776, "sign": 0.3298069331794977, "keep": 0.2772245891392231, "leak": 1.4656586307683028, "replacement": 1.0467741787433624}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 1.6706616878509521, "loss_scale": 4.0, "retries": 0, "seconds": 117.0, "max_memory_gb": 24.64, "pair": 0.47651586309075356, "sign": 0.4118087161332369, "keep": 0.4763742443174124, "leak": 0.5594135440769605, "replacement": 1.0604607313871384}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 2.198917865753174, "loss_scale": 4.0, "retries": 0, "seconds": 118.3, "max_memory_gb": 24.64, "pair": 0.42918383702635765, "sign": 0.3734575370326638, "keep": 0.2469999548047781, "leak": 3.1392750890245225, "replacement": 1.524231344461441}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.234622001647949, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 24.64, "pair": 0.5515839233994484, "sign": 0.4157165065407753, "keep": 0.328883308917284, "leak": 0.9693480283021927, "replacement": 1.1662208586931229}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 2.881514310836792, "loss_scale": 4.0, "retries": 0, "seconds": 120.7, "max_memory_gb": 24.64, "pair": 0.6499184723943472, "sign": 0.5170719809830189, "keep": 0.09376408904790878, "leak": 1.3135363347828388, "replacement": 1.2837205827236176}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 2.16239070892334, "loss_scale": 4.0, "retries": 0, "seconds": 121.9, "max_memory_gb": 24.64, "pair": 0.8654889464378357, "sign": 0.8520420268177986, "keep": 0.30413942714221776, "leak": 0.096785600297153, "replacement": 1.4677760601043701}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 1.4893382787704468, "loss_scale": 4.0, "retries": 0, "seconds": 123.1, "max_memory_gb": 24.64, "pair": 0.5440699979662895, "sign": 0.5809135772287846, "keep": 0.28358528204262257, "leak": 1.8462446828543762, "replacement": 1.437880590558052}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 1.5591365098953247, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 24.64, "pair": 0.30214504338800907, "sign": 0.12853909586556256, "keep": 0.16618762910366058, "leak": 0.8571760654449463, "replacement": 0.6638024374842644}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 2.8966856002807617, "loss_scale": 4.0, "retries": 0, "seconds": 125.6, "max_memory_gb": 24.64, "pair": 0.34848860278725624, "sign": 0.2730226796120405, "keep": 0.7247693687677383, "leak": 1.815628414042294, "replacement": 1.3012917637825012}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 2.2250125408172607, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 24.64, "pair": 0.5028536170721054, "sign": 0.2943124733865261, "keep": 0.09986718092113733, "leak": 1.8099167495965958, "replacement": 1.152422621846199}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 4.672296524047852, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 24.64, "pair": 0.7027349472045898, "sign": 0.8354066759347916, "keep": 0.07649339293129742, "leak": 0.27507615368813276, "replacement": 1.2274539768695831}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 2.4402859210968018, "loss_scale": 4.0, "retries": 0, "seconds": 129.2, "max_memory_gb": 24.64, "pair": 0.41002586111426353, "sign": 0.31187425553798676, "keep": 0.18732929602265358, "leak": 2.4048733562231064, "replacement": 1.2608459293842316}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 2.514288902282715, "loss_scale": 4.0, "retries": 0, "seconds": 130.5, "max_memory_gb": 24.64, "pair": 0.3745056986808777, "sign": 0.3753740657120943, "keep": 0.19291931949555874, "leak": 2.7062238419021014, "replacement": 1.335208386182785}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 3.16184401512146, "loss_scale": 4.0, "retries": 0, "seconds": 131.7, "max_memory_gb": 24.64, "pair": 0.7312777265906334, "sign": 0.6848836913704872, "keep": 0.19721577037125826, "leak": 0.22577423474285752, "replacement": 1.2287710309028625}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.7770235538482666, "loss_scale": 4.0, "retries": 0, "seconds": 132.9, "max_memory_gb": 24.64, "pair": 0.7167354933917522, "sign": 0.6615644302219152, "keep": 0.1862272759899497, "leak": 2.0838230950175785, "replacement": 1.6615870893001556}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 2.9175362586975098, "loss_scale": 4.0, "retries": 0, "seconds": 134.2, "max_memory_gb": 24.64, "pair": 0.3763012960553169, "sign": 0.4025493711233139, "keep": 0.3721181992441416, "leak": 2.486867464147508, "replacement": 1.3853519335389137}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 3.150252342224121, "loss_scale": 4.0, "retries": 0, "seconds": 135.4, "max_memory_gb": 24.64, "pair": 0.49460404831916094, "sign": 0.4553217557258904, "keep": 0.2769532538950443, "leak": 1.6936711817979813, "replacement": 1.2841593325138092}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 2.156512498855591, "loss_scale": 4.0, "retries": 0, "seconds": 136.6, "max_memory_gb": 24.64, "pair": 0.4301963448524475, "sign": 0.4081349018961191, "keep": 0.08897420251742005, "leak": 0.4222830682992935, "replacement": 0.7843216806650162}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 2.2528040409088135, "loss_scale": 4.0, "retries": 0, "seconds": 137.8, "max_memory_gb": 24.64, "pair": 0.4550880528986454, "sign": 0.35227776132524014, "keep": 0.19558168901130557, "leak": 0.7586461342871189, "replacement": 0.9186792969703674}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 2.0898640155792236, "loss_scale": 4.0, "retries": 0, "seconds": 139.1, "max_memory_gb": 24.64, "pair": 0.5058761667460203, "sign": 0.4631048236042261, "keep": 0.20488475635647774, "leak": 0.6858612629584968, "replacement": 1.0113362818956375}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 3.0425312519073486, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 24.64, "pair": 0.36881902627646923, "sign": 0.4091026717796922, "keep": 0.5201213918626308, "leak": 2.1126277055591345, "replacement": 1.3615879714488983}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 2.3272035121917725, "loss_scale": 4.0, "retries": 0, "seconds": 141.5, "max_memory_gb": 24.64, "pair": 0.5108457468450069, "sign": 0.5071965772658587, "keep": 0.37066776119172573, "leak": 1.3882612778106704, "replacement": 1.296843245625496}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 2.457858085632324, "loss_scale": 4.0, "retries": 0, "seconds": 142.8, "max_memory_gb": 24.64, "pair": 0.6477781683206558, "sign": 0.5804212987422943, "keep": 0.18613015464507043, "leak": 0.22657419450115412, "replacement": 1.0876974165439606}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 3.9346439838409424, "loss_scale": 4.0, "retries": 0, "seconds": 144.0, "max_memory_gb": 24.64, "pair": 0.5680029988288879, "sign": 0.3777454597875476, "keep": 0.4505599152762443, "leak": 0.9710857914760709, "replacement": 1.224927231669426}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 1.3574806451797485, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 24.64, "pair": 0.6324603706598282, "sign": 0.49793362617492676, "keep": 0.2653186568059027, "leak": 0.3991677004378289, "replacement": 1.1138784289360046}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.5899295806884766, "loss_scale": 4.0, "retries": 0, "seconds": 146.5, "max_memory_gb": 24.64, "pair": 0.7290598601102829, "sign": 0.7677480429410934, "keep": 0.15597774693742394, "leak": 0.43956908164545894, "replacement": 1.3008150458335876}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 3.658804416656494, "loss_scale": 4.0, "retries": 0, "seconds": 147.7, "max_memory_gb": 24.64, "pair": 0.49497411400079727, "sign": 0.47617420367896557, "keep": 0.653337748721242, "leak": 0.8569105006754398, "replacement": 1.2739576883614063}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.9653825759887695, "loss_scale": 4.0, "retries": 0, "seconds": 149.0, "max_memory_gb": 24.64, "pair": 0.6645359545946121, "sign": 0.6297788098454475, "keep": 0.5156982867047191, "leak": 2.098685674369335, "replacement": 1.7619459182024002}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 3.138784646987915, "loss_scale": 4.0, "retries": 0, "seconds": 150.2, "max_memory_gb": 24.64, "pair": 0.6270346343517303, "sign": 0.6043407786637545, "keep": 0.19270846107974648, "leak": 0.5092504264139279, "replacement": 1.1528718322515488}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.1471521854400635, "loss_scale": 4.0, "retries": 0, "seconds": 151.4, "max_memory_gb": 24.64, "pair": 0.6511423327028751, "sign": 0.7367853540927172, "keep": 1.0137961497530341, "leak": 0.5825371779937996, "replacement": 1.6720673590898514}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 3.176664113998413, "loss_scale": 4.0, "retries": 0, "seconds": 152.6, "max_memory_gb": 24.64, "pair": 0.8578635007143021, "sign": 1.0459925383329391, "keep": 0.5960035645402968, "leak": 0.13899409817531705, "replacement": 1.713610053062439}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.3574254512786865, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 24.64, "pair": 0.6455308869481087, "sign": 0.7051645219326019, "keep": 0.20778636075556278, "leak": 0.21555751701816916, "replacement": 1.1558957397937775}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 1.8984867334365845, "loss_scale": 4.0, "retries": 0, "seconds": 155.0, "max_memory_gb": 24.64, "pair": 0.6416467949748039, "sign": 0.5939771216362715, "keep": 0.48435643687844276, "leak": 0.39089220738969743, "replacement": 1.27853661775589}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 1.7975956201553345, "loss_scale": 4.0, "retries": 0, "seconds": 156.2, "max_memory_gb": 24.64, "pair": 0.5184115096926689, "sign": 0.30416853353381157, "keep": 0.053831682074815035, "leak": 0.3551229014992714, "replacement": 0.7861923575401306}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 2.398238182067871, "loss_scale": 4.0, "retries": 0, "seconds": 157.4, "max_memory_gb": 24.64, "pair": 0.32115988433361053, "sign": 0.2238002228550613, "keep": 0.14798632869496942, "leak": 1.2569437623023987, "replacement": 0.8212890923023224}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 2.3526172637939453, "loss_scale": 4.0, "retries": 0, "seconds": 158.7, "max_memory_gb": 24.64, "pair": 0.3273499384522438, "sign": 0.18643952254205942, "keep": 0.10898048151284456, "leak": 2.007921482436359, "replacement": 0.9770403131842613}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 5.347670078277588, "loss_scale": 4.0, "retries": 0, "seconds": 159.9, "max_memory_gb": 24.64, "pair": 0.575728140771389, "sign": 0.8782982118427753, "keep": 0.5344494776800275, "leak": 1.141810867935419, "replacement": 1.5675546526908875}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 3.7361977100372314, "loss_scale": 4.0, "retries": 0, "seconds": 161.1, "max_memory_gb": 24.64, "pair": 0.5677638985216618, "sign": 0.6336853802204132, "keep": 0.49944257736206055, "leak": 0.4683333458378911, "replacement": 1.2514111995697021}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 2.736006021499634, "loss_scale": 4.0, "retries": 0, "seconds": 162.3, "max_memory_gb": 24.64, "pair": 0.36164760775864124, "sign": 0.11478975228965282, "keep": 0.8475142177194357, "leak": 1.1290495544672012, "replacement": 1.1250619739294052}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 4.190500259399414, "loss_scale": 4.0, "retries": 0, "seconds": 163.5, "max_memory_gb": 24.64, "pair": 0.4149271696805954, "sign": 0.4020801763981581, "keep": 0.487054243683815, "leak": 2.515351891517639, "replacement": 1.488332360982895}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 3.275089740753174, "loss_scale": 4.0, "retries": 0, "seconds": 164.8, "max_memory_gb": 24.64, "pair": 0.6870298981666565, "sign": 0.633674830198288, "keep": 0.3057854752987623, "leak": 0.29287583008408546, "replacement": 1.2299790233373642}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 3.2976202964782715, "loss_scale": 4.0, "retries": 0, "seconds": 166.0, "max_memory_gb": 24.64, "pair": 0.42242057248950005, "sign": 0.4088710444048047, "keep": 0.21124686300754547, "leak": 3.15117944823578, "replacement": 1.520274430513382}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 4.314662933349609, "loss_scale": 4.0, "retries": 0, "seconds": 167.2, "max_memory_gb": 24.64, "pair": 0.5113742128014565, "sign": 0.43706309981644154, "keep": 0.5180154563859105, "leak": 0.41141162998974323, "replacement": 1.0917663723230362}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 4.597841262817383, "loss_scale": 4.0, "retries": 0, "seconds": 168.5, "max_memory_gb": 24.64, "pair": 0.5001371465623379, "sign": 0.4309093081392348, "keep": 0.2782091833651066, "leak": 2.787627376616001, "replacement": 1.5516032576560974}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 1.9053986072540283, "loss_scale": 4.0, "retries": 0, "seconds": 169.7, "max_memory_gb": 24.64, "pair": 0.33849767223000526, "sign": 0.2188434638082981, "keep": 0.16458587534725666, "leak": 0.41431924887001514, "replacement": 0.6337921693921089}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 6.086415767669678, "loss_scale": 4.0, "retries": 0, "seconds": 171.0, "max_memory_gb": 24.64, "pair": 0.4558589309453964, "sign": 0.4687397740781307, "keep": 0.34764861688017845, "leak": 1.11741079762578, "replacement": 1.1434058621525764}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 2.1008782386779785, "loss_scale": 4.0, "retries": 0, "seconds": 172.2, "max_memory_gb": 24.64, "pair": 0.37301265448331833, "sign": 0.40028018318116665, "keep": 0.07740089809522033, "leak": 0.8359152376651764, "replacement": 0.8208320289850235}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 3.757345199584961, "loss_scale": 4.0, "retries": 0, "seconds": 173.5, "max_memory_gb": 24.64, "pair": 0.4429250806570053, "sign": 0.4781787805259228, "keep": 0.38696264009922743, "leak": 1.38179337605834, "replacement": 1.2209440767765045}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.3086559772491455, "loss_scale": 4.0, "retries": 0, "seconds": 174.7, "max_memory_gb": 24.64, "pair": 0.45407770201563835, "sign": 0.360130007378757, "keep": 0.22344663832336664, "leak": 0.5049364755977876, "replacement": 0.8721001297235489}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 2.8232853412628174, "loss_scale": 4.0, "retries": 0, "seconds": 175.9, "max_memory_gb": 24.64, "pair": 0.7684165500104427, "sign": 0.9864592254161835, "keep": 0.25868684344459325, "leak": 0.3946857117116451, "replacement": 1.489661082625389}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 3.045344591140747, "loss_scale": 4.0, "retries": 0, "seconds": 177.2, "max_memory_gb": 24.64, "pair": 0.33761759474873543, "sign": 0.30083909165114164, "keep": 0.2125895619392395, "leak": 1.9009600579738617, "replacement": 1.0695719420909882}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 2.9155993461608887, "loss_scale": 4.0, "retries": 0, "seconds": 178.4, "max_memory_gb": 24.64, "pair": 0.5201906301081181, "sign": 0.28465699637308717, "keep": 0.04799622134305537, "leak": 1.417243055999279, "replacement": 1.0408280342817307}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.3317956924438477, "loss_scale": 4.0, "retries": 0, "seconds": 179.6, "max_memory_gb": 24.64, "pair": 0.4900365862995386, "sign": 0.5234846984967589, "keep": 0.161968725733459, "leak": 1.2948099055793136, "replacement": 1.156465783715248}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.565225124359131, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 24.64, "pair": 0.5005019418895245, "sign": 0.49202557653188705, "keep": 0.4627339905127883, "leak": 1.1028695814311504, "replacement": 1.2535991221666336}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.8101905584335327, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 24.64, "pair": 0.43045520782470703, "sign": 0.2460508719086647, "keep": 0.032156519358977675, "leak": 0.5150454062968493, "replacement": 0.6983202546834946}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 2.2748091220855713, "loss_scale": 4.0, "retries": 0, "seconds": 183.3, "max_memory_gb": 24.64, "pair": 0.43473800271749496, "sign": 0.3901995085179806, "keep": 0.10332660563290119, "leak": 0.36091010086238384, "replacement": 0.7717285677790642}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 2.4524145126342773, "loss_scale": 4.0, "retries": 0, "seconds": 184.5, "max_memory_gb": 24.64, "pair": 0.380788154900074, "sign": 0.3132694251835346, "keep": 0.4330008216202259, "leak": 2.3636520747095346, "replacement": 1.3448363244533539}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 2.754702568054199, "loss_scale": 4.0, "retries": 0, "seconds": 185.8, "max_memory_gb": 24.64, "pair": 0.3171916529536247, "sign": 0.30203332006931305, "keep": 0.32852351083420217, "leak": 0.5157047063112259, "replacement": 0.7613962441682816}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 3.0674941539764404, "loss_scale": 4.0, "retries": 0, "seconds": 187.0, "max_memory_gb": 24.64, "pair": 0.21403472684323788, "sign": 0.15945584094151855, "keep": 0.12259032251313329, "leak": 2.877336546778679, "replacement": 1.0743919610977173}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 2.451131582260132, "loss_scale": 4.0, "retries": 0, "seconds": 188.3, "max_memory_gb": 24.64, "pair": 0.29570598900318146, "sign": 0.20751342363655567, "keep": 0.06229609018191695, "leak": 1.815207302570343, "replacement": 0.8844125717878342}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 3.1594455242156982, "loss_scale": 4.0, "retries": 0, "seconds": 189.5, "max_memory_gb": 24.64, "pair": 0.24413738679140806, "sign": 0.15273395786061883, "keep": 0.7739014364778996, "leak": 1.0658258646726608, "replacement": 0.9739115685224533}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.590388536453247, "loss_scale": 4.0, "retries": 0, "seconds": 190.8, "max_memory_gb": 24.64, "pair": 0.3848648630082607, "sign": 0.39672939106822014, "keep": 0.9960081956814975, "leak": 2.196459323167801, "replacement": 1.630348488688469}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.898511528968811, "loss_scale": 4.0, "retries": 0, "seconds": 192.0, "max_memory_gb": 24.64, "pair": 0.5033610351383686, "sign": 0.45858192536979914, "keep": 0.1717889083083719, "leak": 0.4211342497728765, "replacement": 0.9238300174474716}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 3.5235443115234375, "loss_scale": 4.0, "retries": 0, "seconds": 193.2, "max_memory_gb": 24.64, "pair": 0.6535257771611214, "sign": 1.080588474869728, "keep": 0.20323948515579104, "leak": 0.35933716129511595, "replacement": 1.3852741122245789}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 2.678786277770996, "loss_scale": 4.0, "retries": 0, "seconds": 194.4, "max_memory_gb": 24.64, "pair": 0.4687004406005144, "sign": 0.5645641926676035, "keep": 0.3877509590238333, "leak": 0.5778892352245748, "replacement": 1.0893303453922272}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 2.9412741661071777, "loss_scale": 4.0, "retries": 0, "seconds": 195.7, "max_memory_gb": 24.64, "pair": 0.19618256110697985, "sign": 0.13475423213094473, "keep": 0.05899615027010441, "leak": 2.3871701508760452, "replacement": 0.8898502886295319}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 4.474888801574707, "loss_scale": 4.0, "retries": 0, "seconds": 196.9, "max_memory_gb": 24.64, "pair": 0.43922723829746246, "sign": 0.30653083277866244, "keep": 0.17713671550154686, "leak": 1.7296985329448944, "replacement": 1.1134856566786766}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 3.540839910507202, "loss_scale": 4.0, "retries": 0, "seconds": 198.1, "max_memory_gb": 24.64, "pair": 0.561967259272933, "sign": 0.5590012390166521, "keep": 0.37930495105683804, "leak": 1.0756798153743148, "replacement": 1.3000403344631195}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 3.622976779937744, "loss_scale": 4.0, "retries": 0, "seconds": 199.3, "max_memory_gb": 24.64, "pair": 0.42623631563037634, "sign": 0.48129898216575384, "keep": 0.07891935366205871, "leak": 0.4096008352935314, "replacement": 0.8087457120418549}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 2.0459372997283936, "loss_scale": 4.0, "retries": 0, "seconds": 200.5, "max_memory_gb": 24.64, "pair": 0.38684236630797386, "sign": 0.2798174424096942, "keep": 0.07038868311792612, "leak": 0.8179361894726753, "replacement": 0.7664294540882111}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 3.6645233631134033, "loss_scale": 4.0, "retries": 0, "seconds": 201.7, "max_memory_gb": 24.64, "pair": 0.5790451429784298, "sign": 0.7608612300828099, "keep": 0.42061295732855797, "leak": 0.6960661970078945, "replacement": 1.3437987565994263}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 3.520925760269165, "loss_scale": 4.0, "retries": 0, "seconds": 203.0, "max_memory_gb": 24.64, "pair": 0.3557602073997259, "sign": 0.4603890236467123, "keep": 0.11800299445167184, "leak": 1.9832439720630646, "replacement": 1.1407671719789505}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 2.9523520469665527, "loss_scale": 4.0, "retries": 0, "seconds": 204.3, "max_memory_gb": 24.64, "pair": 0.4072051905095577, "sign": 0.4527057483792305, "keep": 0.24588235327973962, "leak": 1.6349385139765218, "replacement": 1.165233850479126}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 4.15639591217041, "loss_scale": 4.0, "retries": 0, "seconds": 205.5, "max_memory_gb": 24.64, "pair": 0.3232186632230878, "sign": 0.3779039429500699, "keep": 0.5079952646046877, "leak": 1.737910585012287, "replacement": 1.2006458789110184}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.728435516357422, "loss_scale": 4.0, "retries": 0, "seconds": 206.7, "max_memory_gb": 24.64, "pair": 0.37793710082769394, "sign": 0.33758228830993176, "keep": 0.10588885494507849, "leak": 1.3999112620949745, "replacement": 0.9496504813432693}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 4.741367340087891, "loss_scale": 4.0, "retries": 0, "seconds": 208.0, "max_memory_gb": 24.64, "pair": 0.32528917118906975, "sign": 0.35972458124160767, "keep": 0.6346862595528364, "leak": 2.3592881057411432, "replacement": 1.4123166054487228}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 1.9583475589752197, "loss_scale": 4.0, "retries": 0, "seconds": 209.2, "max_memory_gb": 24.64, "pair": 0.21185285225510597, "sign": 0.0701525085605681, "keep": 0.03895923402160406, "leak": 1.4168378561735153, "replacement": 0.6206181719899178}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 6.020623207092285, "loss_scale": 4.0, "retries": 0, "seconds": 210.4, "max_memory_gb": 24.64, "pair": 0.3904363214969635, "sign": 0.35697205178439617, "keep": 0.05708013125695288, "leak": 1.669107623398304, "replacement": 1.01473930478096}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 3.8445844650268555, "loss_scale": 4.0, "retries": 0, "seconds": 211.7, "max_memory_gb": 24.64, "pair": 0.20797372981905937, "sign": 0.06773186614736915, "keep": 0.1840117583051324, "leak": 4.580797176808119, "replacement": 1.4790448099374771}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 5.259735107421875, "loss_scale": 4.0, "retries": 0, "seconds": 212.9, "max_memory_gb": 24.64, "pair": 0.2482957672327757, "sign": 0.37354837637394667, "keep": 0.31488459184765816, "leak": 2.4911713749170303, "replacement": 1.215305119752884}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 2.955785036087036, "loss_scale": 4.0, "retries": 0, "seconds": 214.1, "max_memory_gb": 24.64, "pair": 0.3640485778450966, "sign": 0.24952848255634308, "keep": 0.1767154596745968, "leak": 1.2738532074727118, "replacement": 0.8956338465213776}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.735069751739502, "loss_scale": 4.0, "retries": 0, "seconds": 215.3, "max_memory_gb": 24.64, "pair": 0.3536608014255762, "sign": 0.4405980724841356, "keep": 0.14339741691946983, "leak": 2.438864446245134, "replacement": 1.2553746551275253}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 2.812203884124756, "loss_scale": 4.0, "retries": 0, "seconds": 216.6, "max_memory_gb": 24.64, "pair": 0.3340962827205658, "sign": 0.13109183497726917, "keep": 0.10173531621694565, "leak": 2.168934065848589, "replacement": 0.9927433729171753}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 6.592394828796387, "loss_scale": 4.0, "retries": 0, "seconds": 217.8, "max_memory_gb": 24.64, "pair": 0.4141952469944954, "sign": 0.3492975337430835, "keep": 0.4611940081231296, "leak": 0.6857572812587023, "replacement": 0.9908803552389145}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.8509631156921387, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 24.64, "pair": 0.4666382670402527, "sign": 0.47411611676216125, "keep": 0.17979053687304258, "leak": 1.6140298140235245, "replacement": 1.197099044919014}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 2.9416723251342773, "loss_scale": 4.0, "retries": 0, "seconds": 220.2, "max_memory_gb": 24.64, "pair": 0.6861540917307138, "sign": 0.7516560303047299, "keep": 0.6007682359777391, "leak": 0.7813183928374201, "replacement": 1.5576958507299423}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 8.328899383544922, "loss_scale": 4.0, "retries": 0, "seconds": 221.4, "max_memory_gb": 24.64, "pair": 0.8294994309544563, "sign": 0.7093346919864416, "keep": 1.130076976493001, "leak": 0.6518286601640284, "replacement": 1.9121625125408173}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 4.076695919036865, "loss_scale": 4.0, "retries": 0, "seconds": 222.6, "max_memory_gb": 24.64, "pair": 0.4495644085109234, "sign": 0.15607355860993266, "keep": 0.16276702843606472, "leak": 2.0974805429577827, "replacement": 1.1333548352122307}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 5.765696048736572, "loss_scale": 4.0, "retries": 0, "seconds": 223.9, "max_memory_gb": 24.64, "pair": 0.5824651010334492, "sign": 0.602997750043869, "keep": 0.6124641709029675, "leak": 0.8730992041528225, "replacement": 1.4084708243608475}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.7421586513519287, "loss_scale": 4.0, "retries": 0, "seconds": 225.1, "max_memory_gb": 24.64, "pair": 0.23210406675934792, "sign": 0.12267295271158218, "keep": 0.039241813006810844, "leak": 1.3863132297992706, "replacement": 0.6596397385001183}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 6.242918491363525, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 24.64, "pair": 0.4331485964357853, "sign": 0.31796457432210445, "keep": 0.1627357049146667, "leak": 0.6332357283681631, "replacement": 0.8318076580762863}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 3.272516965866089, "loss_scale": 4.0, "retries": 0, "seconds": 227.6, "max_memory_gb": 24.64, "pair": 0.3752014935016632, "sign": 0.23768740985542536, "keep": 0.09236993454396725, "leak": 2.3545758561813273, "replacement": 1.128874123096466}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 3.606642484664917, "loss_scale": 4.0, "retries": 0, "seconds": 228.8, "max_memory_gb": 24.64, "pair": 0.3544107712805271, "sign": 0.24109426233917475, "keep": 0.14669416099786758, "leak": 0.7559540721122175, "replacement": 0.7372934967279434}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 4.221617221832275, "loss_scale": 4.0, "retries": 0, "seconds": 230.1, "max_memory_gb": 24.64, "pair": 0.3929584138095379, "sign": 0.08224542625248432, "keep": 0.18112103082239628, "leak": 1.2450153846293688, "replacement": 0.8358955085277557}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 2.67952299118042, "loss_scale": 4.0, "retries": 0, "seconds": 231.3, "max_memory_gb": 24.64, "pair": 0.49916721135377884, "sign": 0.48263959027826786, "keep": 0.2212598966434598, "leak": 0.3706099997507408, "replacement": 0.9437694549560547}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.3872339725494385, "loss_scale": 4.0, "retries": 0, "seconds": 232.6, "max_memory_gb": 24.64, "pair": 0.27928798273205757, "sign": 0.19205919466912746, "keep": 0.0925078191794455, "leak": 0.9634698674199171, "replacement": 0.6624389737844467}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 2.703464984893799, "loss_scale": 4.0, "retries": 0, "seconds": 233.8, "max_memory_gb": 24.64, "pair": 0.25781236588954926, "sign": 0.17712450213730335, "keep": 0.07666970230638981, "leak": 1.3209842965006828, "replacement": 0.7149555534124374}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 6.542234897613525, "loss_scale": 4.0, "retries": 0, "seconds": 234.9, "max_memory_gb": 24.64, "pair": 0.5216112434864044, "sign": 0.5900417435914278, "keep": 0.7820960143581033, "leak": 0.8927324414253235, "replacement": 1.4308632612228394}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 2.0278141498565674, "loss_scale": 4.0, "retries": 0, "seconds": 236.2, "max_memory_gb": 24.64, "pair": 0.5951542593538761, "sign": 0.4616649355739355, "keep": 0.34209061600267887, "leak": 0.4584819398005493, "replacement": 1.1116525083780289}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 2.6662817001342773, "loss_scale": 4.0, "retries": 0, "seconds": 237.4, "max_memory_gb": 24.64, "pair": 0.3797071985900402, "sign": 0.16525812447071075, "keep": 0.12369129899889231, "leak": 0.42449801415205, "replacement": 0.6303064078092575}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 11.246149063110352, "loss_scale": 4.0, "retries": 0, "seconds": 238.6, "max_memory_gb": 24.64, "pair": 0.6661289036273956, "sign": 0.8791776150465012, "keep": 1.4441770184785128, "leak": 1.0479447692632675, "replacement": 2.0897924602031708}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 5.168399333953857, "loss_scale": 4.0, "retries": 0, "seconds": 239.9, "max_memory_gb": 24.64, "pair": 0.49246664717793465, "sign": 0.3751003462821245, "keep": 0.4612130634486675, "leak": 1.8450498450547457, "replacement": 1.3718858510255814}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 1.9159598350524902, "loss_scale": 4.0, "retries": 0, "seconds": 241.1, "max_memory_gb": 24.64, "pair": 0.27678127586841583, "sign": 0.09649667330086231, "keep": 0.07616374175995588, "leak": 0.5060760453343391, "replacement": 0.48963047564029694}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 4.18179178237915, "loss_scale": 4.0, "retries": 0, "seconds": 242.4, "max_memory_gb": 24.64, "pair": 0.4473321959376335, "sign": 0.44907655473798513, "keep": 0.1371318008750677, "leak": 0.34108314011245966, "replacement": 0.8257071673870087}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 3.7914297580718994, "loss_scale": 4.0, "retries": 0, "seconds": 243.6, "max_memory_gb": 24.64, "pair": 0.438077624887228, "sign": 0.5177557580173016, "keep": 0.2586920987814665, "leak": 1.0394033272241359, "replacement": 1.0861523896455765}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.235074520111084, "loss_scale": 4.0, "retries": 0, "seconds": 244.8, "max_memory_gb": 24.64, "pair": 0.2722692899405956, "sign": 0.11826909519731998, "keep": 0.3334302350413054, "leak": 1.9934364557266235, "replacement": 0.9964780807495117}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 2.3355696201324463, "loss_scale": 4.0, "retries": 0, "seconds": 246.0, "max_memory_gb": 24.64, "pair": 0.44730832427740097, "sign": 0.549638818949461, "keep": 0.4218271626159549, "leak": 0.7578872963786125, "replacement": 1.1225131750106812}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 3.888089895248413, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 24.64, "pair": 0.42515629529953003, "sign": 0.5365144181996584, "keep": 0.17971021635457873, "leak": 0.7627174034714699, "replacement": 0.9739479571580887}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 3.9263341426849365, "loss_scale": 4.0, "retries": 0, "seconds": 248.5, "max_memory_gb": 24.64, "pair": 0.3962128013372421, "sign": 0.28489959985017776, "keep": 0.013042792212218046, "leak": 0.8116983771324158, "replacement": 0.7481085956096649}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 5.00254487991333, "loss_scale": 4.0, "retries": 0, "seconds": 249.7, "max_memory_gb": 24.64, "pair": 0.42436186224222183, "sign": 0.4318388430401683, "keep": 0.2756017167121172, "leak": 1.18406777898781, "replacement": 1.0740990862250328}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 4.8509111404418945, "loss_scale": 4.0, "retries": 0, "seconds": 250.9, "max_memory_gb": 24.64, "pair": 0.5740282908082008, "sign": 0.6071161832660437, "keep": 0.2832713481038809, "leak": 1.427058756351471, "replacement": 1.375986710190773}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 3.8288521766662598, "loss_scale": 4.0, "retries": 0, "seconds": 252.2, "max_memory_gb": 24.64, "pair": 0.1773979254066944, "sign": 0.2345193475484848, "keep": 0.08454438322223723, "leak": 1.8748230412602425, "replacement": 0.8056355565786362}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 3.302182912826538, "loss_scale": 4.0, "retries": 0, "seconds": 253.4, "max_memory_gb": 24.64, "pair": 0.3458001874387264, "sign": 0.28851368464529514, "keep": 0.15128557209391147, "leak": 2.287065800279379, "replacement": 1.1374662965536118}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 8.669873237609863, "loss_scale": 4.0, "retries": 0, "seconds": 254.7, "max_memory_gb": 24.64, "pair": 0.8092356994748116, "sign": 0.9450322985649109, "keep": 1.550886817742139, "leak": 0.4614629689604044, "replacement": 2.1725610345602036}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 3.31646466255188, "loss_scale": 4.0, "retries": 0, "seconds": 255.9, "max_memory_gb": 24.64, "pair": 0.3403087928891182, "sign": 0.08460068562999368, "keep": 0.04637384647503495, "leak": 1.554810754954815, "replacement": 0.7944987416267395}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 3.091001510620117, "loss_scale": 4.0, "retries": 0, "seconds": 257.2, "max_memory_gb": 24.64, "pair": 0.4622556008398533, "sign": 0.5251444894820452, "keep": 0.3100135698914528, "leak": 0.7728511928580701, "replacement": 1.0730474516749382}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 3.2503790855407715, "loss_scale": 4.0, "retries": 0, "seconds": 258.4, "max_memory_gb": 24.64, "pair": 0.38925874419510365, "sign": 0.46695668809115887, "keep": 0.7029963927343488, "leak": 2.2847841195762157, "replacement": 1.5454313158988953}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 7.398282527923584, "loss_scale": 4.0, "retries": 0, "seconds": 259.7, "max_memory_gb": 24.64, "pair": 0.6298281531780958, "sign": 0.6332353197503835, "keep": 0.48018124513328075, "leak": 0.6643917053006589, "replacement": 1.352634385228157}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 10.073511123657227, "loss_scale": 4.0, "retries": 0, "seconds": 260.9, "max_memory_gb": 24.64, "pair": 0.4534429460763931, "sign": 0.6760380491614342, "keep": 0.6651079095900059, "leak": 0.83311374229379, "replacement": 1.3322943598031998}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 4.199394226074219, "loss_scale": 4.0, "retries": 0, "seconds": 262.1, "max_memory_gb": 24.64, "pair": 0.3662573844194412, "sign": 0.45811319071799517, "keep": 0.9992965906858444, "leak": 1.6266389023512602, "replacement": 1.501621961593628}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 3.180858850479126, "loss_scale": 4.0, "retries": 0, "seconds": 263.4, "max_memory_gb": 24.64, "pair": 0.5547406934201717, "sign": 0.5520026497542858, "keep": 0.4938690452836454, "leak": 0.40984749193739844, "replacement": 1.1801383942365646}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 6.105751991271973, "loss_scale": 4.0, "retries": 0, "seconds": 264.6, "max_memory_gb": 24.64, "pair": 0.7751568481326103, "sign": 0.9701446108520031, "keep": 1.7658962635323405, "leak": 0.4970134915820381, "replacement": 2.2674305737018585}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 2.5872325897216797, "loss_scale": 4.0, "retries": 0, "seconds": 265.9, "max_memory_gb": 24.64, "pair": 0.26585812121629715, "sign": 0.250078771263361, "keep": 0.2727901563048363, "leak": 0.7435148507356644, "replacement": 0.7131712883710861}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 6.125205039978027, "loss_scale": 4.0, "retries": 0, "seconds": 267.1, "max_memory_gb": 24.64, "pair": 0.18190541584044695, "sign": 0.1507452456280589, "keep": 0.04716466763056815, "leak": 1.8107244074344635, "replacement": 0.7335414588451385}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 5.509424209594727, "loss_scale": 4.0, "retries": 0, "seconds": 268.3, "max_memory_gb": 24.64, "pair": 0.7218995243310928, "sign": 0.9209218099713326, "keep": 0.4388826403301209, "leak": 0.17669581811060198, "replacement": 1.4459757208824158}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 5.629812240600586, "loss_scale": 4.0, "retries": 0, "seconds": 269.5, "max_memory_gb": 24.64, "pair": 0.4049297831952572, "sign": 0.48609352856874466, "keep": 0.22869490436278284, "leak": 0.38794916908227606, "replacement": 0.8593112677335739}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 4.678666591644287, "loss_scale": 4.0, "retries": 0, "seconds": 270.7, "max_memory_gb": 24.64, "pair": 0.3456161804497242, "sign": 0.47426630556583405, "keep": 0.3654860110254958, "leak": 0.7917134314775467, "replacement": 0.9634206891059875}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 2.7450006008148193, "loss_scale": 4.0, "retries": 0, "seconds": 271.9, "max_memory_gb": 24.64, "pair": 0.3273338507860899, "sign": 0.2650614380836487, "keep": 0.21432541031390429, "leak": 0.9519101196201518, "replacement": 0.8050047904253006}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 4.65254020690918, "loss_scale": 4.0, "retries": 0, "seconds": 273.2, "max_memory_gb": 24.64, "pair": 0.38187699764966965, "sign": 0.21760953590273857, "keep": 0.44137427024543285, "leak": 1.1345909163355827, "replacement": 0.9950166419148445}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 6.953762054443359, "loss_scale": 4.0, "retries": 0, "seconds": 274.4, "max_memory_gb": 24.64, "pair": 0.5085338950157166, "sign": 0.6069733276963234, "keep": 0.2773349637282081, "leak": 0.21216533984988928, "replacement": 1.003729410469532}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 3.4722821712493896, "loss_scale": 4.0, "retries": 0, "seconds": 275.7, "max_memory_gb": 24.64, "pair": 0.3019139803946018, "sign": 0.33821018785238266, "keep": 0.11825208773370832, "leak": 1.2246201522648335, "replacement": 0.8363001644611359}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 3.221243381500244, "loss_scale": 4.0, "retries": 0, "seconds": 276.9, "max_memory_gb": 24.64, "pair": 0.4804528867825866, "sign": 0.5343441450968385, "keep": 0.11325820069760084, "leak": 1.12928706407547, "replacement": 1.0865758284926414}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 4.39917516708374, "loss_scale": 4.0, "retries": 0, "seconds": 278.1, "max_memory_gb": 24.64, "pair": 0.26552916690707207, "sign": 0.2466687597334385, "keep": 0.19613883108831942, "leak": 0.7636576443910599, "replacement": 0.6778473556041718}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.544987201690674, "loss_scale": 4.0, "retries": 0, "seconds": 279.3, "max_memory_gb": 24.64, "pair": 0.4308161661028862, "sign": 0.19340970180928707, "keep": 0.04334302979987115, "leak": 0.5980473421514034, "replacement": 0.6987043917179108}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 3.7067761421203613, "loss_scale": 4.0, "retries": 0, "seconds": 280.5, "max_memory_gb": 24.64, "pair": 0.16815987415611744, "sign": 0.25634484365582466, "keep": 0.027992668095976114, "leak": 1.1712611019611359, "replacement": 0.6031439006328583}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 5.272933006286621, "loss_scale": 4.0, "retries": 0, "seconds": 281.7, "max_memory_gb": 24.64, "pair": 0.42860952764749527, "sign": 0.6231113323010504, "keep": 0.672514129197225, "leak": 1.0899515450000763, "replacement": 1.3489101976156235}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 5.391844272613525, "loss_scale": 4.0, "retries": 0, "seconds": 282.9, "max_memory_gb": 24.64, "pair": 0.6681457832455635, "sign": 0.8049769438803196, "keep": 0.6549057774245739, "leak": 0.46703996509313583, "replacement": 1.5148471742868423}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 5.487743854522705, "loss_scale": 4.0, "retries": 0, "seconds": 284.1, "max_memory_gb": 24.64, "pair": 0.5981009751558304, "sign": 0.7466532252728939, "keep": 0.6641790340654552, "leak": 0.3511345610022545, "replacement": 1.3913007974624634}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 3.5541024208068848, "loss_scale": 4.0, "retries": 0, "seconds": 285.4, "max_memory_gb": 24.64, "pair": 0.24874770268797874, "sign": 0.2732718186452985, "keep": 0.04837704077363014, "leak": 0.5613372614607215, "replacement": 0.5499064326286316}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 7.367796897888184, "loss_scale": 4.0, "retries": 0, "seconds": 286.6, "max_memory_gb": 24.64, "pair": 0.28449852392077446, "sign": 0.16772799100726843, "keep": 0.5181754347868264, "leak": 1.3115344885736704, "replacement": 0.9553338289260864}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.707461357116699, "loss_scale": 4.0, "retries": 0, "seconds": 287.8, "max_memory_gb": 24.64, "pair": 0.33453774079680443, "sign": 0.3986380733549595, "keep": 0.26648914883844554, "leak": 1.765806145966053, "replacement": 1.1085528880357742}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 5.21925163269043, "loss_scale": 4.0, "retries": 0, "seconds": 289.0, "max_memory_gb": 24.64, "pair": 0.5049232505261898, "sign": 0.5887105949223042, "keep": 0.7673599477857351, "leak": 0.8704933718472603, "replacement": 1.4005818665027618}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 4.510007858276367, "loss_scale": 4.0, "retries": 0, "seconds": 290.2, "max_memory_gb": 24.64, "pair": 0.5424472242593765, "sign": 0.46806490793824196, "keep": 0.2695036481600255, "leak": 0.197266710922122, "replacement": 0.9605481177568436}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.7080814838409424, "loss_scale": 4.0, "retries": 0, "seconds": 291.4, "max_memory_gb": 24.64, "pair": 0.49501047283411026, "sign": 0.4280724972486496, "keep": 0.3594819721765816, "leak": 0.3676566946087405, "replacement": 0.9807018935680389}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 2.8919320106506348, "loss_scale": 4.0, "retries": 0, "seconds": 292.6, "max_memory_gb": 24.64, "pair": 0.189845971763134, "sign": 0.11724903620779514, "keep": 0.10827383981086314, "leak": 1.2139606215059757, "replacement": 0.6060975566506386}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 4.629153251647949, "loss_scale": 4.0, "retries": 0, "seconds": 293.9, "max_memory_gb": 24.64, "pair": 0.39569105952978134, "sign": 0.4897681586444378, "keep": 0.2976044421084225, "leak": 3.0841240007430315, "replacement": 1.5604083687067032}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.340348482131958, "loss_scale": 4.0, "retries": 0, "seconds": 295.1, "max_memory_gb": 24.64, "pair": 0.29050976410508156, "sign": 0.2513200640678406, "keep": 0.0791864845668897, "leak": 0.5387090966105461, "replacement": 0.5904403105378151}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 5.621215343475342, "loss_scale": 4.0, "retries": 0, "seconds": 296.3, "max_memory_gb": 24.64, "pair": 0.4132286310195923, "sign": 0.5959870740771294, "keep": 0.1381469606421888, "leak": 0.608140388445463, "replacement": 0.9323307424783707}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 2.558274269104004, "loss_scale": 4.0, "retries": 0, "seconds": 297.5, "max_memory_gb": 24.64, "pair": 0.4037310592830181, "sign": 0.5094495443627238, "keep": 0.20693439280148596, "leak": 0.7409224510192871, "replacement": 0.9471536129713058}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 3.582058906555176, "loss_scale": 4.0, "retries": 0, "seconds": 298.8, "max_memory_gb": 24.64, "pair": 0.2582041062414646, "sign": 0.21670990251004696, "keep": 0.05166441958863288, "leak": 2.7700310200452805, "replacement": 1.0848990231752396}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 11.552026748657227, "loss_scale": 4.0, "retries": 0, "seconds": 300.0, "max_memory_gb": 24.64, "pair": 0.7941894084215164, "sign": 1.014326037839055, "keep": 1.2497756276279688, "leak": 0.7088514911010861, "replacement": 2.1034531593322754}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 5.237351417541504, "loss_scale": 4.0, "retries": 0, "seconds": 301.2, "max_memory_gb": 24.64, "pair": 0.8232208751142025, "sign": 1.0000839605927467, "keep": 0.8064132556319237, "leak": 0.4401292288675904, "replacement": 1.8365017473697662}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 2.542551040649414, "loss_scale": 4.0, "retries": 0, "seconds": 302.5, "max_memory_gb": 24.64, "pair": 0.1941712684929371, "sign": 0.1599949486553669, "keep": 0.05997188237961382, "leak": 1.0015836209058762, "replacement": 0.5545505881309509}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 3.359196662902832, "loss_scale": 4.0, "retries": 0, "seconds": 303.7, "max_memory_gb": 24.64, "pair": 0.2794030122458935, "sign": 0.31715793162584305, "keep": 0.2055703052901663, "leak": 0.6556303352117538, "replacement": 0.704674705862999}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.3851311206817627, "loss_scale": 4.0, "retries": 0, "seconds": 304.9, "max_memory_gb": 24.64, "pair": 0.2262212038040161, "sign": 0.12958829873241484, "keep": 0.16057902574539185, "leak": 0.5069750434486195, "replacement": 0.4980486184358597}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 5.624207019805908, "loss_scale": 4.0, "retries": 0, "seconds": 306.2, "max_memory_gb": 24.64, "pair": 0.4863439239561558, "sign": 0.7356211468577385, "keep": 0.08256945060566068, "leak": 0.7795899378834292, "replacement": 1.0903367102146149}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 4.479263782501221, "loss_scale": 4.0, "retries": 0, "seconds": 307.4, "max_memory_gb": 24.64, "pair": 0.20762128569185734, "sign": 0.1927070114761591, "keep": 0.07688742026221007, "leak": 1.7152359150350094, "replacement": 0.7712274789810181}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 3.8067405223846436, "loss_scale": 4.0, "retries": 0, "seconds": 308.6, "max_memory_gb": 24.64, "pair": 0.3777921088039875, "sign": 0.403202660381794, "keep": 0.17783430370036513, "leak": 1.279506069200579, "replacement": 0.9881870970129967}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 2.4068355560302734, "loss_scale": 4.0, "retries": 0, "seconds": 309.9, "max_memory_gb": 24.64, "pair": 0.23219433054327965, "sign": 0.13676822697743773, "keep": 0.0265404237434268, "leak": 0.8856215104460716, "replacement": 0.5352540388703346}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 8.143760681152344, "loss_scale": 4.0, "retries": 0, "seconds": 311.1, "max_memory_gb": 24.64, "pair": 0.2505214177072048, "sign": 0.2948283152654767, "keep": 0.208802250912413, "leak": 1.9495140835642815, "replacement": 0.9897152036428452}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 4.269198417663574, "loss_scale": 4.0, "retries": 0, "seconds": 312.3, "max_memory_gb": 24.64, "pair": 0.22252260521054268, "sign": 0.24950047209858894, "keep": 0.41640233248472214, "leak": 1.396523803472519, "replacement": 0.904604971408844}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 4.130025863647461, "loss_scale": 4.0, "retries": 0, "seconds": 313.5, "max_memory_gb": 24.64, "pair": 0.44020558148622513, "sign": 0.6063981093466282, "keep": 0.08567087270785123, "leak": 0.6357384920120239, "replacement": 0.9451746940612793}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 3.4791128635406494, "loss_scale": 4.0, "retries": 0, "seconds": 314.7, "max_memory_gb": 24.64, "pair": 0.4030994139611721, "sign": 0.4881774913519621, "keep": 0.10404661577194929, "leak": 0.930028572678566, "replacement": 0.9317186176776886}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 9.24095344543457, "loss_scale": 4.0, "retries": 0, "seconds": 315.9, "max_memory_gb": 24.64, "pair": 0.41562519967556, "sign": 0.44420411018654704, "keep": 0.16751492081675678, "leak": 0.9615640416741371, "replacement": 0.961875669658184}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 2.582826614379883, "loss_scale": 4.0, "retries": 0, "seconds": 317.2, "max_memory_gb": 24.64, "pair": 0.19683909509330988, "sign": 0.1426504822447896, "keep": 0.2978625927353278, "leak": 1.2435822486877441, "replacement": 0.7279911860823631}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.0233378410339355, "loss_scale": 4.0, "retries": 0, "seconds": 318.4, "max_memory_gb": 24.64, "pair": 0.19678529724478722, "sign": 0.17521585151553154, "keep": 0.006900632695760578, "leak": 0.8014966007322073, "replacement": 0.48821768164634705}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 6.427338600158691, "loss_scale": 4.0, "retries": 0, "seconds": 319.6, "max_memory_gb": 24.64, "pair": 0.24970905017107725, "sign": 0.35778323374688625, "keep": 0.16723559983074665, "leak": 1.2711385637521744, "replacement": 0.8300031200051308}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.7257769107818604, "loss_scale": 4.0, "retries": 0, "seconds": 320.8, "max_memory_gb": 24.64, "pair": 0.32317399233579636, "sign": 0.31068160105496645, "keep": 0.288282360881567, "leak": 0.4804426613263786, "replacement": 0.7427665963768959}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 6.913936138153076, "loss_scale": 4.0, "retries": 0, "seconds": 322.0, "max_memory_gb": 24.64, "pair": 0.5175524316728115, "sign": 0.5977928917855024, "keep": 0.5196229799184948, "leak": 0.37389610009267926, "replacement": 1.1697343587875366}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 5.109992980957031, "loss_scale": 4.0, "retries": 0, "seconds": 323.3, "max_memory_gb": 24.64, "pair": 0.3224816769361496, "sign": 0.34083304926753044, "keep": 0.3224107678979635, "leak": 1.7401677183806896, "replacement": 1.0891455262899399}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 8.205364227294922, "loss_scale": 4.0, "retries": 0, "seconds": 324.5, "max_memory_gb": 24.64, "pair": 0.28562915697693825, "sign": 0.1625716662965715, "keep": 0.8789562123129144, "leak": 2.6898380515922327, "replacement": 1.4788525849580765}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 3.2803757190704346, "loss_scale": 4.0, "retries": 0, "seconds": 325.8, "max_memory_gb": 24.64, "pair": 0.23694545403122902, "sign": 0.26314255502074957, "keep": 0.08352960110642016, "leak": 1.6922796051949263, "replacement": 0.8333514332771301}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 11.656960487365723, "loss_scale": 4.0, "retries": 0, "seconds": 327.0, "max_memory_gb": 24.64, "pair": 0.563665084540844, "sign": 0.5941417599096894, "keep": 1.1486469050869346, "leak": 0.6427015215158463, "replacement": 1.595734789967537}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 4.128020286560059, "loss_scale": 4.0, "retries": 0, "seconds": 328.2, "max_memory_gb": 24.64, "pair": 0.4238581694662571, "sign": 0.4696906600147486, "keep": 0.43035683513153344, "leak": 0.49241000786423683, "replacement": 0.9969844333827496}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 2.9473447799682617, "loss_scale": 4.0, "retries": 0, "seconds": 329.4, "max_memory_gb": 24.64, "pair": 0.5994997043162584, "sign": 0.7168640680611134, "keep": 0.3950399092864245, "leak": 0.28479276978760026, "replacement": 1.226649910211563}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.077256202697754, "loss_scale": 4.0, "retries": 0, "seconds": 330.7, "max_memory_gb": 24.64, "pair": 0.22205574624240398, "sign": 0.31761561892926693, "keep": 1.2534371475921944, "leak": 0.9397413320839405, "replacement": 1.242517463862896}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 9.327681541442871, "loss_scale": 4.0, "retries": 0, "seconds": 331.9, "max_memory_gb": 24.64, "pair": 0.24667297676205635, "sign": 0.4566840915940702, "keep": 0.6807095427066088, "leak": 1.4441650062799454, "replacement": 1.1764110252261162}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 3.730133533477783, "loss_scale": 4.0, "retries": 0, "seconds": 333.2, "max_memory_gb": 24.64, "pair": 0.2473885975778103, "sign": 0.21285587176680565, "keep": 0.07334041129797697, "leak": 2.0530944615602493, "replacement": 0.9037603437900543}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 3.376443862915039, "loss_scale": 4.0, "retries": 0, "seconds": 334.4, "max_memory_gb": 24.64, "pair": 0.12089938670396805, "sign": 0.044637734070420265, "keep": 0.05230483552441001, "leak": 0.7936525456607342, "replacement": 0.3677838072180748}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 4.4242048263549805, "loss_scale": 4.0, "retries": 0, "seconds": 335.6, "max_memory_gb": 24.64, "pair": 0.447122685611248, "sign": 0.19409264996647835, "keep": 0.04604209330864251, "leak": 0.08241730474401265, "replacement": 0.5877943634986877}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 7.44521951675415, "loss_scale": 4.0, "retries": 0, "seconds": 336.8, "max_memory_gb": 24.64, "pair": 0.4723663702607155, "sign": 0.7031051367521286, "keep": 0.15325832413509488, "leak": 0.3272617382463068, "replacement": 0.9823635518550873}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 4.03382682800293, "loss_scale": 4.0, "retries": 0, "seconds": 338.1, "max_memory_gb": 24.64, "pair": 0.44706887751817703, "sign": 0.47149498388171196, "keep": 1.549437121488154, "leak": 0.2937964992597699, "replacement": 1.5309840738773346}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 2.3952884674072266, "loss_scale": 4.0, "retries": 0, "seconds": 339.3, "max_memory_gb": 24.64, "pair": 0.16466017719358206, "sign": 0.08641912788152695, "keep": 0.053715186135377735, "leak": 0.5447931094095111, "replacement": 0.3709256201982498}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 3.0696420669555664, "loss_scale": 4.0, "retries": 0, "seconds": 340.5, "max_memory_gb": 24.64, "pair": 0.3546082451939583, "sign": 0.26039216853678226, "keep": 0.03271649044472724, "leak": 0.2276013819500804, "replacement": 0.5580629333853722}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 12.848724365234375, "loss_scale": 4.0, "retries": 0, "seconds": 341.7, "max_memory_gb": 24.64, "pair": 0.4136595241725445, "sign": 0.5965378805994987, "keep": 0.6950326077640057, "leak": 0.47504333034157753, "replacement": 1.178205594420433}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 4.3636393547058105, "loss_scale": 4.0, "retries": 0, "seconds": 342.9, "max_memory_gb": 24.64, "pair": 0.37106627598404884, "sign": 0.24760036915540695, "keep": 0.14161643583793193, "leak": 1.1962904930114746, "replacement": 0.8647473007440567}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 2.279808521270752, "loss_scale": 4.0, "retries": 0, "seconds": 344.2, "max_memory_gb": 24.64, "pair": 0.17486291751265526, "sign": 0.11534801591187716, "keep": 0.05643553449772298, "leak": 0.6115149133838713, "replacement": 0.41363342478871346}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 3.055145025253296, "loss_scale": 4.0, "retries": 0, "seconds": 345.4, "max_memory_gb": 24.64, "pair": 0.21902714110910892, "sign": 0.2269250974059105, "keep": 0.32171207771170884, "leak": 0.8020161017775536, "replacement": 0.6938497349619865}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.7168056964874268, "loss_scale": 4.0, "retries": 0, "seconds": 346.6, "max_memory_gb": 24.64, "pair": 0.34064967278391123, "sign": 0.3168438198044896, "keep": 0.14539907779544592, "leak": 0.1973597938194871, "replacement": 0.6211110763251781}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 4.449910640716553, "loss_scale": 4.0, "retries": 0, "seconds": 347.9, "max_memory_gb": 24.64, "pair": 0.18240522779524326, "sign": 0.09126912988722324, "keep": 0.1521582929417491, "leak": 2.647205173969269, "replacement": 0.9659202396869659}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.382866144180298, "loss_scale": 4.0, "retries": 0, "seconds": 349.1, "max_memory_gb": 24.64, "pair": 0.20789199881255627, "sign": 0.1786463689059019, "keep": 0.15883032069541514, "leak": 0.8057622984051704, "replacement": 0.5780709162354469}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 4.938361167907715, "loss_scale": 4.0, "retries": 0, "seconds": 350.3, "max_memory_gb": 24.64, "pair": 0.3137946296483278, "sign": 0.503610011190176, "keep": 0.18425441486760974, "leak": 2.383213177556172, "replacement": 1.2535300999879837}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 7.661339282989502, "loss_scale": 4.0, "retries": 0, "seconds": 351.6, "max_memory_gb": 24.64, "pair": 0.34927586652338505, "sign": 0.495275616645813, "keep": 0.04071805579587817, "leak": 2.4250416308641434, "replacement": 1.2235330939292908}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 9.589069366455078, "loss_scale": 4.0, "retries": 0, "seconds": 352.9, "max_memory_gb": 24.64, "pair": 0.3289601467549801, "sign": 0.433548413682729, "keep": 0.5053812023252249, "leak": 0.13739070255542174, "replacement": 0.8327725976705551}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 5.384693145751953, "loss_scale": 4.0, "retries": 0, "seconds": 354.1, "max_memory_gb": 24.64, "pair": 0.33503001369535923, "sign": 0.2996293529868126, "keep": 0.05696692690253258, "leak": 0.7845037169754505, "replacement": 0.7094541043043137}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 4.853271961212158, "loss_scale": 4.0, "retries": 0, "seconds": 355.4, "max_memory_gb": 24.64, "pair": 0.19861546158790588, "sign": 0.15157930925488472, "keep": 0.06906990054994822, "leak": 1.649939285358414, "replacement": 0.7214248850941658}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 5.504306316375732, "loss_scale": 4.0, "retries": 0, "seconds": 356.6, "max_memory_gb": 24.64, "pair": 0.24289087019860744, "sign": 0.32063448266126215, "keep": 0.25240699015557766, "leak": 0.7522255033254623, "replacement": 0.7174679934978485}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 4.471199035644531, "loss_scale": 4.0, "retries": 0, "seconds": 357.8, "max_memory_gb": 24.64, "pair": 0.11267943121492863, "sign": 0.0537288268096745, "keep": 0.03645373065955937, "leak": 2.050969921052456, "replacement": 0.6705131828784943}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 3.179490804672241, "loss_scale": 4.0, "retries": 0, "seconds": 359.1, "max_memory_gb": 24.64, "pair": 0.1667046807706356, "sign": 0.12745835026726127, "keep": 0.02948125684633851, "leak": 0.9797951355576515, "replacement": 0.49012327194213867}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 4.899161338806152, "loss_scale": 4.0, "retries": 0, "seconds": 360.3, "max_memory_gb": 24.64, "pair": 0.2904748944565654, "sign": 0.2837401553988457, "keep": 0.054678106564097106, "leak": 0.7222667410969734, "replacement": 0.6402507349848747}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 3.757917881011963, "loss_scale": 4.0, "retries": 0, "seconds": 361.5, "max_memory_gb": 24.64, "pair": 0.18801143020391464, "sign": 0.1633797981776297, "keep": 0.36775854229927063, "leak": 1.1795722767710686, "replacement": 0.748473696410656}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 11.296014785766602, "loss_scale": 4.0, "retries": 0, "seconds": 362.7, "max_memory_gb": 24.64, "pair": 0.2963842675089836, "sign": 0.41837965697050095, "keep": 0.3530036836164072, "leak": 1.759393049404025, "replacement": 1.1219241619110107}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 9.679522514343262, "loss_scale": 4.0, "retries": 0, "seconds": 363.9, "max_memory_gb": 24.64, "pair": 0.42262718081474304, "sign": 0.715282492339611, "keep": 0.6993371876887977, "leak": 1.2360122569371015, "replacement": 1.4389400854706764}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 7.168638706207275, "loss_scale": 4.0, "retries": 0, "seconds": 365.1, "max_memory_gb": 24.64, "pair": 0.483707157894969, "sign": 0.8308789506554604, "keep": 0.5678896734607406, "leak": 1.061233401298523, "replacement": 1.4483998492360115}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 5.859986305236816, "loss_scale": 4.0, "retries": 0, "seconds": 366.4, "max_memory_gb": 24.64, "pair": 0.5307767689228058, "sign": 0.6335274800658226, "keep": 0.7780263158492744, "leak": 0.2672501417109743, "replacement": 1.3033661618828773}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 3.947233200073242, "loss_scale": 4.0, "retries": 0, "seconds": 367.6, "max_memory_gb": 24.64, "pair": 0.32837837748229504, "sign": 0.1084990045055747, "keep": 0.06020756531506777, "leak": 1.3436383791267872, "replacement": 0.7486412525177002}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 13.37845230102539, "loss_scale": 4.0, "retries": 0, "seconds": 368.8, "max_memory_gb": 24.64, "pair": 0.5837555006146431, "sign": 0.8410493731498718, "keep": 0.7097630314528942, "leak": 1.625974199268967, "replacement": 1.765655279159546}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 7.040998458862305, "loss_scale": 4.0, "retries": 0, "seconds": 370.0, "max_memory_gb": 24.64, "pair": 0.2608465626835823, "sign": 0.36037836130708456, "keep": 0.20815254282206297, "leak": 0.6558581963181496, "replacement": 0.7090765684843063}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 4.063337802886963, "loss_scale": 4.0, "retries": 0, "seconds": 371.3, "max_memory_gb": 24.64, "pair": 0.17234346643090248, "sign": 0.05615101847797632, "keep": 0.13099748454988003, "leak": 0.33485740050673485, "replacement": 0.34963206201791763}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 7.046201229095459, "loss_scale": 4.0, "retries": 0, "seconds": 372.5, "max_memory_gb": 24.64, "pair": 0.5155705315992236, "sign": 0.37467085756361485, "keep": 0.19420526502653956, "leak": 1.2198081370443106, "replacement": 1.1049606204032898}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 6.78245210647583, "loss_scale": 4.0, "retries": 0, "seconds": 373.7, "max_memory_gb": 24.64, "pair": 0.42057467298582196, "sign": 0.6153517626225948, "keep": 0.046940967440605164, "leak": 1.4878571182489395, "replacement": 1.123685333877802}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 5.80695915222168, "loss_scale": 4.0, "retries": 0, "seconds": 375.0, "max_memory_gb": 24.64, "pair": 0.25908657163381577, "sign": 0.16341423615813255, "keep": 0.06536206277087331, "leak": 1.667964480817318, "replacement": 0.790465846657753}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 5.533140182495117, "loss_scale": 4.0, "retries": 0, "seconds": 376.2, "max_memory_gb": 24.64, "pair": 0.15542923472821712, "sign": 0.040770338382571936, "keep": 0.05057995347306132, "leak": 0.9134490787982941, "replacement": 0.4294666573405266}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 13.290053367614746, "loss_scale": 4.0, "retries": 0, "seconds": 377.4, "max_memory_gb": 24.64, "pair": 0.4149181991815567, "sign": 0.45358372665941715, "keep": 0.521836772793904, "leak": 0.6523052006959915, "replacement": 1.0657047107815742}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 4.193551063537598, "loss_scale": 4.0, "retries": 0, "seconds": 378.7, "max_memory_gb": 24.64, "pair": 0.18473886838182807, "sign": 0.28666550293564796, "keep": 0.1072345677530393, "leak": 1.3336296826601028, "replacement": 0.7150963135063648}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 21.233186721801758, "loss_scale": 4.0, "retries": 0, "seconds": 379.8, "max_memory_gb": 24.64, "pair": 0.5817571505904198, "sign": 0.7955160045530647, "keep": 0.8102958146482706, "leak": 1.283982016146183, "replacement": 1.7056585550308228}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 1.8253616094589233, "loss_scale": 4.0, "retries": 0, "seconds": 381.1, "max_memory_gb": 24.64, "pair": 0.2584094926714897, "sign": 0.11051615560427308, "keep": 0.040236382745206356, "leak": 1.1679425425827503, "replacement": 0.6257713884115219}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 6.163174152374268, "loss_scale": 4.0, "retries": 0, "seconds": 382.3, "max_memory_gb": 24.64, "pair": 0.3433313649147749, "sign": 0.46761852875351906, "keep": 0.518841480021365, "leak": 0.7636203691363335, "replacement": 1.0274664536118507}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 2.801654100418091, "loss_scale": 4.0, "retries": 0, "seconds": 383.5, "max_memory_gb": 24.64, "pair": 0.20385047048330307, "sign": 0.15689788106828928, "keep": 0.07153119577560574, "leak": 0.19777920516207814, "replacement": 0.36750980466604233}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 14.146116256713867, "loss_scale": 4.0, "retries": 0, "seconds": 384.7, "max_memory_gb": 24.64, "pair": 0.4073759000748396, "sign": 0.6545832008123398, "keep": 0.8859059989918023, "leak": 1.0774112790822983, "replacement": 1.4469733536243439}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 12.481894493103027, "loss_scale": 4.0, "retries": 0, "seconds": 386.0, "max_memory_gb": 24.64, "pair": 0.33032059110701084, "sign": 0.43827975098975, "keep": 0.6279783195350319, "leak": 1.7197082936763763, "replacement": 1.2933767139911652}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 6.512212753295898, "loss_scale": 4.0, "retries": 0, "seconds": 387.2, "max_memory_gb": 24.64, "pair": 0.2016638945788145, "sign": 0.22242029663175344, "keep": 0.0775973042473197, "leak": 1.2261816691607237, "replacement": 0.6582181230187416}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 4.774430274963379, "loss_scale": 4.0, "retries": 0, "seconds": 388.5, "max_memory_gb": 24.64, "pair": 0.5909659005701542, "sign": 0.7668833108618855, "keep": 0.37521587684750557, "leak": 0.7724720342084765, "replacement": 1.3551334589719772}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 5.642606735229492, "loss_scale": 4.0, "retries": 0, "seconds": 389.7, "max_memory_gb": 24.64, "pair": 0.2839240711182356, "sign": 0.36076652631163597, "keep": 0.22678707167506218, "leak": 1.6326975040137768, "replacement": 0.9858752340078354}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 6.011115074157715, "loss_scale": 4.0, "retries": 0, "seconds": 390.8, "max_memory_gb": 24.64, "pair": 0.2631388492882252, "sign": 0.24794461205601692, "keep": 0.07926814304664731, "leak": 1.1578638553619385, "replacement": 0.716211199760437}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 6.599488735198975, "loss_scale": 4.0, "retries": 0, "seconds": 392.1, "max_memory_gb": 24.64, "pair": 0.46004527248442173, "sign": 0.3865894805639982, "keep": 0.1359183844178915, "leak": 0.5033958214335144, "replacement": 0.8471481204032898}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 6.580092906951904, "loss_scale": 4.0, "retries": 0, "seconds": 393.3, "max_memory_gb": 24.64, "pair": 0.32710741087794304, "sign": 0.3863512189127505, "keep": 0.36560076219029725, "leak": 0.6547171995043755, "replacement": 0.8667627274990082}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 4.335235595703125, "loss_scale": 4.0, "retries": 0, "seconds": 394.5, "max_memory_gb": 24.64, "pair": 0.26135212555527687, "sign": 0.17988965287804604, "keep": 0.02101325721014291, "leak": 0.874292807187885, "replacement": 0.5803768038749695}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 5.930700778961182, "loss_scale": 4.0, "retries": 0, "seconds": 395.7, "max_memory_gb": 24.64, "pair": 0.3662766255438328, "sign": 0.44387220591306686, "keep": 0.05726440902799368, "leak": 0.1510158554156078, "replacement": 0.6545989066362381}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 9.263214111328125, "loss_scale": 4.0, "retries": 0, "seconds": 397.0, "max_memory_gb": 24.64, "pair": 0.06276482529938221, "sign": 0.0714269494637847, "keep": 0.08641726337373257, "leak": 4.222851395606995, "replacement": 1.1973998248577118}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 4.301860332489014, "loss_scale": 4.0, "retries": 0, "seconds": 398.2, "max_memory_gb": 24.64, "pair": 0.14605052629485726, "sign": 0.07119597168639302, "keep": 0.08846822683699429, "leak": 1.2375155021436512, "replacement": 0.5352614894509315}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 12.125810623168945, "loss_scale": 4.0, "retries": 0, "seconds": 399.4, "max_memory_gb": 24.64, "pair": 0.8482825309038162, "sign": 0.7671795608475804, "keep": 0.7547667435137555, "leak": 0.14597987523302436, "replacement": 1.6457506641745567}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 8.875834465026855, "loss_scale": 4.0, "retries": 0, "seconds": 400.7, "max_memory_gb": 24.64, "pair": 0.4432070702314377, "sign": 0.5295189507305622, "keep": 0.4359608514350839, "leak": 0.7045530879441344, "replacement": 1.1020852774381638}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 6.957437038421631, "loss_scale": 4.0, "retries": 0, "seconds": 401.9, "max_memory_gb": 24.64, "pair": 0.1587429828941822, "sign": 0.09519871743395925, "keep": 0.44265822460874915, "leak": 0.24738698732107878, "replacement": 0.4895182028412819}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 6.073185443878174, "loss_scale": 4.0, "retries": 0, "seconds": 403.1, "max_memory_gb": 24.64, "pair": 0.4087369590997696, "sign": 0.5214109420776367, "keep": 0.1291953360196203, "leak": 0.2253258605778683, "replacement": 0.7903715670108795}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 8.079675674438477, "loss_scale": 4.0, "retries": 0, "seconds": 404.3, "max_memory_gb": 24.64, "pair": 0.3032900672405958, "sign": 0.4083014652132988, "keep": 0.3281415373785421, "leak": 0.9706692285835743, "replacement": 0.914178878068924}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 9.967828750610352, "loss_scale": 4.0, "retries": 0, "seconds": 405.6, "max_memory_gb": 24.64, "pair": 0.5399548280984163, "sign": 0.5737432977184653, "keep": 0.4185491222888231, "leak": 0.5101561924730049, "replacement": 1.1636401265859604}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 7.854190349578857, "loss_scale": 4.0, "retries": 0, "seconds": 406.8, "max_memory_gb": 24.64, "pair": 0.5914216935634613, "sign": 0.6229524090886116, "keep": 0.6176653082948178, "leak": 1.06868387148279, "replacement": 1.4789015501737595}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 4.470209121704102, "loss_scale": 4.0, "retries": 0, "seconds": 408.1, "max_memory_gb": 24.64, "pair": 0.20780275017023087, "sign": 0.22605353966355324, "keep": 0.5297356519149616, "leak": 0.6981502920389175, "replacement": 0.7602349445223808}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 6.78782320022583, "loss_scale": 4.0, "retries": 0, "seconds": 409.3, "max_memory_gb": 24.64, "pair": 0.5543850995600224, "sign": 0.7172207552939653, "keep": 0.5797117985785007, "leak": 0.17802089359611273, "replacement": 1.2473565749824047}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 7.284957408905029, "loss_scale": 4.0, "retries": 0, "seconds": 410.5, "max_memory_gb": 24.64, "pair": 0.4836205989122391, "sign": 0.5677016079425812, "keep": 0.6632360033690929, "leak": 1.5558097207671153, "replacement": 1.4880418851971626}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 5.667888164520264, "loss_scale": 4.0, "retries": 0, "seconds": 411.8, "max_memory_gb": 24.64, "pair": 0.45674365758895874, "sign": 0.5536281168460846, "keep": 1.0787927452474833, "leak": 0.3312184242531657, "replacement": 1.3557586967945099}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 6.096125602722168, "loss_scale": 4.0, "retries": 0, "seconds": 413.0, "max_memory_gb": 24.64, "pair": 0.25882523879408836, "sign": 0.13787857443094254, "keep": 0.057530858728569, "leak": 1.063321864232421, "replacement": 0.6223604083061218}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 6.854059219360352, "loss_scale": 4.0, "retries": 0, "seconds": 414.3, "max_memory_gb": 24.64, "pair": 0.4754802882671356, "sign": 0.6863376488909125, "keep": 0.2781422380357981, "leak": 1.0863038976676762, "replacement": 1.229296252131462}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 8.559856414794922, "loss_scale": 4.0, "retries": 0, "seconds": 415.5, "max_memory_gb": 24.64, "pair": 0.37662739865481853, "sign": 0.5092719849199057, "keep": 0.5470712940441445, "leak": 0.8866369351744652, "replacement": 1.1264582574367523}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 6.017101287841797, "loss_scale": 4.0, "retries": 0, "seconds": 416.8, "max_memory_gb": 24.64, "pair": 0.38191291876137257, "sign": 0.6882307138293982, "keep": 0.33568842289969325, "leak": 1.6800491120666265, "replacement": 1.3138847164809704}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 7.592253684997559, "loss_scale": 4.0, "retries": 0, "seconds": 418.0, "max_memory_gb": 24.64, "pair": 0.365750752389431, "sign": 0.3671249570325017, "keep": 0.7238027099519968, "leak": 0.7251619584858418, "replacement": 1.0925050899386406}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 7.4758620262146, "loss_scale": 4.0, "retries": 0, "seconds": 419.2, "max_memory_gb": 24.64, "pair": 0.20674683898687363, "sign": 0.31316478038206697, "keep": 0.3819376062601805, "leak": 1.3203786313533783, "replacement": 0.8843926712870598}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 9.182093620300293, "loss_scale": 4.0, "retries": 0, "seconds": 420.5, "max_memory_gb": 24.64, "pair": 0.5675384681671858, "sign": 0.6954230293631554, "keep": 0.7477068402804434, "leak": 1.3297393079847097, "replacement": 1.6215382516384125}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 5.2810797691345215, "loss_scale": 4.0, "retries": 0, "seconds": 421.7, "max_memory_gb": 24.64, "pair": 0.4637654088437557, "sign": 0.6621728204190731, "keep": 0.02271491871215403, "leak": 0.3971038442105055, "replacement": 0.9054852202534676}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 11.9539213180542, "loss_scale": 4.0, "retries": 0, "seconds": 422.9, "max_memory_gb": 24.64, "pair": 0.24763737991452217, "sign": 0.21249188110232353, "keep": 0.05191539484076202, "leak": 1.4118537977337837, "replacement": 0.7328044772148132}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 6.950410842895508, "loss_scale": 4.0, "retries": 0, "seconds": 424.1, "max_memory_gb": 24.64, "pair": 0.27180673740804195, "sign": 0.40008106315508485, "keep": 0.34169750288128853, "leak": 0.827711827121675, "replacement": 0.8496239930391312}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 9.388181686401367, "loss_scale": 4.0, "retries": 0, "seconds": 425.3, "max_memory_gb": 24.64, "pair": 0.3757154531776905, "sign": 0.5138735622167587, "keep": 0.5382191882235929, "leak": 1.5750360041856766, "replacement": 1.295520856976509}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 3.9743294715881348, "loss_scale": 4.0, "retries": 0, "seconds": 426.6, "max_memory_gb": 24.64, "pair": 0.3983964826911688, "sign": 0.3892614278011024, "keep": 0.32233922462910414, "leak": 0.5634621974313632, "replacement": 0.8950623273849487}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 3.4372522830963135, "loss_scale": 4.0, "retries": 0, "seconds": 427.8, "max_memory_gb": 24.64, "pair": 0.23103591054677963, "sign": 0.12270733434706926, "keep": 0.019264305126853287, "leak": 2.1790598777442938, "replacement": 0.8467867001891136}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 4.571974277496338, "loss_scale": 4.0, "retries": 0, "seconds": 429.1, "max_memory_gb": 24.64, "pair": 0.30827758088707924, "sign": 0.34011593647301197, "keep": 0.1154010365717113, "leak": 1.1670734016224742, "replacement": 0.8278044238686562}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 3.9640164375305176, "loss_scale": 4.0, "retries": 0, "seconds": 430.3, "max_memory_gb": 24.64, "pair": 0.3528936989605427, "sign": 0.5845690220594406, "keep": 0.09818761609494686, "leak": 0.850111149251461, "replacement": 0.9067998081445694}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 3.665149688720703, "loss_scale": 4.0, "retries": 0, "seconds": 431.5, "max_memory_gb": 24.64, "pair": 0.21345039457082748, "sign": 0.10826714197173715, "keep": 0.07450227322988212, "leak": 0.39206443447619677, "replacement": 0.40285119228065014}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 10.425925254821777, "loss_scale": 4.0, "retries": 0, "seconds": 432.7, "max_memory_gb": 24.64, "pair": 0.2847777884453535, "sign": 0.4943509898148477, "keep": 0.412491760449484, "leak": 1.2227437943220139, "replacement": 1.043885089457035}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 5.9844069480896, "loss_scale": 4.0, "retries": 0, "seconds": 433.9, "max_memory_gb": 24.64, "pair": 0.39260632917284966, "sign": 0.4849361330270767, "keep": 0.5732074100524187, "leak": 0.6823010221123695, "replacement": 1.0922533348202705}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 2.1466104984283447, "loss_scale": 4.0, "retries": 0, "seconds": 435.1, "max_memory_gb": 24.64, "pair": 0.45903003960847855, "sign": 0.441348634660244, "keep": 0.307529307436198, "leak": 0.1378395891515538, "replacement": 0.8679288774728775}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 5.082710266113281, "loss_scale": 4.0, "retries": 0, "seconds": 436.4, "max_memory_gb": 24.64, "pair": 0.4243185855448246, "sign": 0.41406722739338875, "keep": 0.21019170759245753, "leak": 1.9964200984686613, "replacement": 1.2355530858039856}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 5.971079349517822, "loss_scale": 4.0, "retries": 0, "seconds": 437.6, "max_memory_gb": 24.64, "pair": 0.12090154737234116, "sign": 0.13373976666480303, "keep": 0.03014745470136404, "leak": 2.456339031457901, "replacement": 0.8169299364089966}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 8.587355613708496, "loss_scale": 4.0, "retries": 0, "seconds": 438.9, "max_memory_gb": 24.64, "pair": 0.2013720078393817, "sign": 0.12411941681057215, "keep": 0.19549017306417227, "leak": 2.315304048359394, "replacement": 0.940002828836441}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 5.537611961364746, "loss_scale": 4.0, "retries": 0, "seconds": 440.1, "max_memory_gb": 24.64, "pair": 0.5086630471050739, "sign": 0.6419076398015022, "keep": 0.46160406502895057, "leak": 0.5212498679757118, "replacement": 1.1907313913106918}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 4.839224338531494, "loss_scale": 4.0, "retries": 0, "seconds": 441.4, "max_memory_gb": 24.64, "pair": 0.346700482070446, "sign": 0.5923540331423283, "keep": 0.04898690298432484, "leak": 1.1127828061580658, "replacement": 0.9455666989088058}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 5.4862189292907715, "loss_scale": 4.0, "retries": 0, "seconds": 442.6, "max_memory_gb": 24.64, "pair": 0.22679126262664795, "sign": 0.19045941904187202, "keep": 0.05466713535133749, "leak": 1.4539390876889229, "replacement": 0.7128393203020096}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 7.859208583831787, "loss_scale": 4.0, "retries": 0, "seconds": 443.8, "max_memory_gb": 24.64, "pair": 0.2617366686463356, "sign": 0.4558953670784831, "keep": 0.02231887204106897, "leak": 1.3571602329611778, "replacement": 0.8401338383555412}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 5.469351291656494, "loss_scale": 4.0, "retries": 0, "seconds": 445.0, "max_memory_gb": 24.64, "pair": 0.46387194097042084, "sign": 0.4278218112885952, "keep": 0.9370182168204337, "leak": 0.15644331322982907, "replacement": 1.185402773320675}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 7.146161079406738, "loss_scale": 4.0, "retries": 0, "seconds": 446.2, "max_memory_gb": 24.64, "pair": 0.38220145739614964, "sign": 0.7089252434670925, "keep": 0.05503099085763097, "leak": 0.6556794839561917, "replacement": 0.9280994236469269}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 4.082403182983398, "loss_scale": 4.0, "retries": 0, "seconds": 447.5, "max_memory_gb": 24.64, "pair": 0.16175554785877466, "sign": 0.046533909160643816, "keep": 0.05242332024499774, "leak": 1.5422887429594994, "replacement": 0.5968063548207283}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 3.919902801513672, "loss_scale": 4.0, "retries": 0, "seconds": 448.7, "max_memory_gb": 24.64, "pair": 0.35266223549842834, "sign": 0.36418974213302135, "keep": 0.24982246162835509, "leak": 0.3308668569661677, "replacement": 0.742385059595108}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 6.877034664154053, "loss_scale": 4.0, "retries": 0, "seconds": 449.9, "max_memory_gb": 24.64, "pair": 0.2632740493863821, "sign": 0.2607409982010722, "keep": 0.07085452973842621, "leak": 2.098688066005707, "replacement": 0.9537438303232193}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 4.5113844871521, "loss_scale": 4.0, "retries": 0, "seconds": 451.1, "max_memory_gb": 24.64, "pair": 0.46764151379466057, "sign": 0.43462288938462734, "keep": 0.4471364507917315, "leak": 0.5572100430727005, "replacement": 1.047823689877987}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 2.847681760787964, "loss_scale": 4.0, "retries": 0, "seconds": 452.4, "max_memory_gb": 24.64, "pair": 0.1931113861501217, "sign": 0.12342548510059714, "keep": 0.056839474476873875, "leak": 0.5076967803761363, "replacement": 0.4101680517196655}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 4.681257247924805, "loss_scale": 4.0, "retries": 0, "seconds": 453.6, "max_memory_gb": 24.64, "pair": 0.4764277823269367, "sign": 0.3855586922727525, "keep": 0.041946325334720314, "leak": 0.6561112026683986, "replacement": 0.8542081341147423}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 11.152511596679688, "loss_scale": 4.0, "retries": 0, "seconds": 454.8, "max_memory_gb": 24.64, "pair": 0.22387776896357536, "sign": 0.28468428552150726, "keep": 1.2539409957826138, "leak": 0.689636368304491, "replacement": 1.165599524974823}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.694817543029785, "loss_scale": 4.0, "retries": 0, "seconds": 456.1, "max_memory_gb": 24.64, "pair": 0.274040287360549, "sign": 0.08499720972031355, "keep": 0.29770396649837494, "leak": 0.5373178599402308, "replacement": 0.5997203513979912}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 7.289041042327881, "loss_scale": 4.0, "retries": 0, "seconds": 457.3, "max_memory_gb": 24.64, "pair": 0.28459150437265635, "sign": 0.2640380533412099, "keep": 0.08121006248984486, "leak": 0.9189646299928427, "replacement": 0.6869566887617111}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 3.7241246700286865, "loss_scale": 4.0, "retries": 0, "seconds": 458.5, "max_memory_gb": 24.64, "pair": 0.5027460604906082, "sign": 0.7247167304158211, "keep": 0.05622034030966461, "leak": 0.45504117105156183, "replacement": 1.0069748684763908}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 5.102712154388428, "loss_scale": 4.0, "retries": 0, "seconds": 459.7, "max_memory_gb": 24.64, "pair": 0.3142929133027792, "sign": 0.2895661061629653, "keep": 0.5631463006138802, "leak": 0.5166733730584383, "replacement": 0.8698174580931664}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 3.872950315475464, "loss_scale": 4.0, "retries": 0, "seconds": 461.0, "max_memory_gb": 24.64, "pair": 0.47174302116036415, "sign": 0.5032322034239769, "keep": 0.3763754736282863, "leak": 0.3532132867258042, "replacement": 0.9998501464724541}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.8729240894317627, "loss_scale": 4.0, "retries": 0, "seconds": 462.2, "max_memory_gb": 24.64, "pair": 0.2500730976462364, "sign": 0.2025986723601818, "keep": 0.10666898847557604, "leak": 0.18314565438777208, "replacement": 0.45049335435032845}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 4.0088677406311035, "loss_scale": 4.0, "retries": 0, "seconds": 463.4, "max_memory_gb": 24.64, "pair": 0.22596803680062294, "sign": 0.2971402918919921, "keep": 0.03800569428130984, "leak": 0.9195586591959, "replacement": 0.6234307028353214}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 4.819126605987549, "loss_scale": 4.0, "retries": 0, "seconds": 464.6, "max_memory_gb": 24.64, "pair": 0.21104050986468792, "sign": 0.15296326763927937, "keep": 0.1268180445767939, "leak": 1.2201738245785236, "replacement": 0.6559746116399765}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 10.667627334594727, "loss_scale": 4.0, "retries": 0, "seconds": 465.8, "max_memory_gb": 24.64, "pair": 0.23963830713182688, "sign": 0.34298874251544476, "keep": 0.278363729827106, "leak": 0.5790339382365346, "replacement": 0.6950730383396149}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 9.21497917175293, "loss_scale": 4.0, "retries": 0, "seconds": 467.1, "max_memory_gb": 24.64, "pair": 0.19379839021712542, "sign": 0.2774707153439522, "keep": 0.052017577923834324, "leak": 2.862179756164551, "replacement": 1.0740875005722046}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 4.866377353668213, "loss_scale": 4.0, "retries": 0, "seconds": 468.3, "max_memory_gb": 24.64, "pair": 0.5815784782171249, "sign": 0.792377769947052, "keep": 0.4579930556938052, "leak": 0.34283349104225636, "replacement": 1.2924723029136658}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 5.599793434143066, "loss_scale": 4.0, "retries": 0, "seconds": 469.5, "max_memory_gb": 24.64, "pair": 0.37545037968084216, "sign": 0.5610681620892137, "keep": 1.050501761958003, "leak": 0.9459661655128002, "replacement": 1.4177268296480179}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 3.455789804458618, "loss_scale": 4.0, "retries": 0, "seconds": 470.7, "max_memory_gb": 24.64, "pair": 0.2426340337842703, "sign": 0.2687184177339077, "keep": 0.12895409110933542, "leak": 1.056787507608533, "replacement": 0.7056671679019928}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 2.7255539894104004, "loss_scale": 4.0, "retries": 0, "seconds": 471.9, "max_memory_gb": 24.64, "pair": 0.09964176267385483, "sign": 0.06091052945703268, "keep": 0.03550196858122945, "leak": 0.5007630251348019, "replacement": 0.2730387672781944}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 5.464074611663818, "loss_scale": 4.0, "retries": 0, "seconds": 473.2, "max_memory_gb": 24.64, "pair": 0.43791074864566326, "sign": 0.49537365324795246, "keep": 0.27507929876446724, "leak": 0.39377114339731634, "replacement": 0.921579971909523}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 5.411831378936768, "loss_scale": 4.0, "retries": 0, "seconds": 474.4, "max_memory_gb": 24.64, "pair": 0.2631442192941904, "sign": 0.17251168563961983, "keep": 0.04535852826666087, "leak": 0.2877345560118556, "replacement": 0.44401296228170395}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 4.515036582946777, "loss_scale": 4.0, "retries": 0, "seconds": 475.6, "max_memory_gb": 24.64, "pair": 0.21524294093251228, "sign": 0.4168564397841692, "keep": 0.09313372313044965, "leak": 0.9079525684937835, "replacement": 0.6972261667251587}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 13.86878776550293, "loss_scale": 4.0, "retries": 0, "seconds": 476.9, "max_memory_gb": 24.64, "pair": 0.27156424010172486, "sign": 0.3084770254790783, "keep": 0.20468168513616547, "leak": 1.0672506024129689, "replacement": 0.7949562519788742}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 3.7413406372070312, "loss_scale": 4.0, "retries": 0, "seconds": 478.1, "max_memory_gb": 24.64, "pair": 0.2003064537420869, "sign": 0.10053101181983948, "keep": 0.17410130985081196, "leak": 1.624938104301691, "replacement": 0.743857130408287}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 14.211465835571289, "loss_scale": 4.0, "retries": 0, "seconds": 479.4, "max_memory_gb": 24.64, "pair": 0.3771788515150547, "sign": 0.5225803125649691, "keep": 0.3076585347298533, "leak": 0.8545602345839143, "replacement": 1.0059383064508438}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 4.41076135635376, "loss_scale": 4.0, "retries": 0, "seconds": 480.6, "max_memory_gb": 24.64, "pair": 0.5349932732060552, "sign": 0.7500015702098608, "keep": 0.4956260919570923, "leak": 0.5184843675233424, "replacement": 1.2874281853437424}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 2.672271728515625, "loss_scale": 4.0, "retries": 0, "seconds": 481.8, "max_memory_gb": 24.64, "pair": 0.15482931025326252, "sign": 0.09004291705787182, "keep": 0.06127195921726525, "leak": 0.23858107719570398, "replacement": 0.2901320159435272}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 4.943633556365967, "loss_scale": 4.0, "retries": 0, "seconds": 483.1, "max_memory_gb": 24.64, "pair": 0.10930927097797394, "sign": 0.14022641815245152, "keep": 0.10854116454720497, "leak": 0.37636118970112875, "replacement": 0.32778334990143776}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 13.934488296508789, "loss_scale": 4.0, "retries": 0, "seconds": 484.3, "max_memory_gb": 24.64, "pair": 0.47069244645535946, "sign": 0.487155357375741, "keep": 0.18818112509325147, "leak": 0.9388905139639974, "replacement": 1.0430833101272583}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 6.499880790710449, "loss_scale": 4.0, "retries": 0, "seconds": 485.6, "max_memory_gb": 24.64, "pair": 0.10555347613990307, "sign": 0.1333033125847578, "keep": 0.04465227777836844, "leak": 1.966440111398697, "replacement": 0.6861413046717644}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 3.4146149158477783, "loss_scale": 4.0, "retries": 0, "seconds": 486.7, "max_memory_gb": 24.64, "pair": 0.15382485650479794, "sign": 0.06699883425608277, "keep": 0.10268665105104446, "leak": 0.40141049306839705, "replacement": 0.3390202224254608}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 7.797201156616211, "loss_scale": 4.0, "retries": 0, "seconds": 487.9, "max_memory_gb": 24.64, "pair": 0.3128843680024147, "sign": 0.4613123135641217, "keep": 0.20338601991534233, "leak": 0.8772979862987995, "replacement": 0.8645580485463142}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 6.393459320068359, "loss_scale": 4.0, "retries": 0, "seconds": 489.2, "max_memory_gb": 24.64, "pair": 0.34999603033065796, "sign": 0.36273745261132717, "keep": 0.24424888310022652, "leak": 0.7336094006896019, "replacement": 0.8368915468454361}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 18.310314178466797, "loss_scale": 4.0, "retries": 0, "seconds": 490.4, "max_memory_gb": 24.64, "pair": 0.5108075197786093, "sign": 0.8098209081217647, "keep": 0.9713280093856156, "leak": 0.9209154918789864, "replacement": 1.6316107772290707}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 9.635557174682617, "loss_scale": 4.0, "retries": 0, "seconds": 491.7, "max_memory_gb": 24.64, "pair": 0.35693322122097015, "sign": 0.266651613637805, "keep": 0.2734469845890999, "leak": 1.3289528039786092, "replacement": 0.9592207074165344}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 15.688467025756836, "loss_scale": 4.0, "retries": 0, "seconds": 493.0, "max_memory_gb": 24.64, "pair": 0.32302276976406574, "sign": 0.32910457625985146, "keep": 0.22203599917702377, "leak": 1.0849903039634228, "replacement": 0.869840681552887}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 12.487990379333496, "loss_scale": 4.0, "retries": 0, "seconds": 494.2, "max_memory_gb": 24.64, "pair": 0.7934832479804754, "sign": 0.9047587318345904, "keep": 0.41862957552075386, "leak": 0.48656993680924643, "replacement": 1.5768198519945145}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 4.3862714767456055, "loss_scale": 4.0, "retries": 0, "seconds": 495.5, "max_memory_gb": 24.64, "pair": 0.18857427313923836, "sign": 0.12143462523818016, "keep": 0.020106431911699474, "leak": 1.5695301964879036, "replacement": 0.6517273485660553}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 16.747140884399414, "loss_scale": 4.0, "retries": 0, "seconds": 496.7, "max_memory_gb": 24.64, "pair": 0.3512703161686659, "sign": 0.4891888638958335, "keep": 0.454763725399971, "leak": 0.6421802863478661, "replacement": 0.9837916269898415}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 4.601583003997803, "loss_scale": 4.0, "retries": 0, "seconds": 497.9, "max_memory_gb": 24.64, "pair": 0.20289148949086666, "sign": 0.17128816805779934, "keep": 0.02503934781998396, "leak": 0.6512944512069225, "replacement": 0.4638788551092148}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 5.457673072814941, "loss_scale": 4.0, "retries": 0, "seconds": 499.2, "max_memory_gb": 24.64, "pair": 0.6133492365479469, "sign": 0.6963440999388695, "keep": 0.16685205698013306, "leak": 0.5174126282799989, "replacement": 1.174300454556942}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 11.569441795349121, "loss_scale": 4.0, "retries": 0, "seconds": 500.4, "max_memory_gb": 24.64, "pair": 0.6200052760541439, "sign": 0.9172424739226699, "keep": 2.1618956020101905, "leak": 0.7107963226735592, "replacement": 2.337273508310318}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 4.990551948547363, "loss_scale": 4.0, "retries": 0, "seconds": 501.6, "max_memory_gb": 24.64, "pair": 0.34259460866451263, "sign": 0.4661468295380473, "keep": 0.5040573509177193, "leak": 0.8242586916312575, "replacement": 1.0337613299489021}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 5.612207889556885, "loss_scale": 4.0, "retries": 0, "seconds": 502.9, "max_memory_gb": 24.64, "pair": 0.373541459441185, "sign": 0.4233748810365796, "keep": 0.38605935871601105, "leak": 1.4608792294748127, "replacement": 1.1434783339500427}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 9.82453727722168, "loss_scale": 4.0, "retries": 0, "seconds": 504.1, "max_memory_gb": 24.64, "pair": 0.19308466464281082, "sign": 0.12385186553001404, "keep": 0.783002617827151, "leak": 1.0011665783822536, "replacement": 0.8968035653233528}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 12.535893440246582, "loss_scale": 4.0, "retries": 0, "seconds": 505.4, "max_memory_gb": 24.64, "pair": 0.45787282288074493, "sign": 0.5722454078495502, "keep": 0.7990784983849153, "leak": 1.0281287971884012, "replacement": 1.4005670547485352}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 4.66025447845459, "loss_scale": 4.0, "retries": 0, "seconds": 506.6, "max_memory_gb": 24.64, "pair": 0.48092273622751236, "sign": 0.5786561220884323, "keep": 0.27240039338357747, "leak": 0.16412738105282187, "replacement": 0.9474828243255615}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 6.312607765197754, "loss_scale": 4.0, "retries": 0, "seconds": 507.8, "max_memory_gb": 24.64, "pair": 0.16324103064835072, "sign": 0.1268764389678836, "keep": 0.03848548233509064, "leak": 1.3248893171548843, "replacement": 0.5771443098783493}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 23.372461318969727, "loss_scale": 4.0, "retries": 0, "seconds": 509.0, "max_memory_gb": 24.64, "pair": 0.39487345330417156, "sign": 0.43176800245419145, "keep": 0.9795328741893172, "leak": 0.6100075095891953, "replacement": 1.2530256882309914}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 7.197345733642578, "loss_scale": 4.0, "retries": 0, "seconds": 510.2, "max_memory_gb": 24.64, "pair": 0.36474689096212387, "sign": 0.3810563301667571, "keep": 1.4244703105650842, "leak": 0.6379355597600807, "replacement": 1.4269940853118896}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 2.6756253242492676, "loss_scale": 4.0, "retries": 0, "seconds": 511.4, "max_memory_gb": 24.64, "pair": 0.1851344984024763, "sign": 0.1222349121235311, "keep": 0.08803730586078018, "leak": 0.6715708654373884, "replacement": 0.4581633284687996}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 12.203543663024902, "loss_scale": 4.0, "retries": 0, "seconds": 512.6, "max_memory_gb": 24.64, "pair": 0.8201992325484753, "sign": 1.1957006603479385, "keep": 1.506661754567176, "leak": 0.7180412635207176, "replacement": 2.3508908450603485}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 2.6163148880004883, "loss_scale": 4.0, "retries": 0, "seconds": 513.8, "max_memory_gb": 24.64, "pair": 0.1328652873635292, "sign": 0.1816117432899773, "keep": 0.045561953098513186, "leak": 0.4417857131411438, "replacement": 0.35689856484532356}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 9.12234878540039, "loss_scale": 4.0, "retries": 0, "seconds": 515.0, "max_memory_gb": 24.64, "pair": 0.44645993411540985, "sign": 0.6680823490023613, "keep": 0.06528042652644217, "leak": 1.5434258729219437, "replacement": 1.1989977657794952}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 16.740055084228516, "loss_scale": 4.0, "retries": 0, "seconds": 516.2, "max_memory_gb": 24.64, "pair": 0.21061836928129196, "sign": 0.4125720467418432, "keep": 0.6661626566201448, "leak": 2.105660706758499, "replacement": 1.2764009162783623}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 8.345179557800293, "loss_scale": 4.0, "retries": 0, "seconds": 517.5, "max_memory_gb": 24.64, "pair": 0.20989786740392447, "sign": 0.24813578743487597, "keep": 0.09719373076222837, "leak": 2.0033443747088313, "replacement": 0.883398711681366}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 7.262589454650879, "loss_scale": 4.0, "retries": 0, "seconds": 518.7, "max_memory_gb": 24.64, "pair": 0.5727942939847708, "sign": 0.5594499912112951, "keep": 0.5433502653613687, "leak": 0.23976101157677476, "replacement": 1.1841346509754658}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 11.42624568939209, "loss_scale": 4.0, "retries": 0, "seconds": 519.9, "max_memory_gb": 24.64, "pair": 0.5253961011767387, "sign": 0.7912591174244881, "keep": 0.8980637639760971, "leak": 0.5014656070852652, "replacement": 1.4954240284860134}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 2.543114423751831, "loss_scale": 4.0, "retries": 0, "seconds": 521.1, "max_memory_gb": 24.64, "pair": 0.1591871976852417, "sign": 0.07924837432801723, "keep": 0.06944990158081055, "leak": 0.6336574032902718, "replacement": 0.39195068925619125}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 6.312501430511475, "loss_scale": 4.0, "retries": 0, "seconds": 522.4, "max_memory_gb": 24.64, "pair": 0.23242627829313278, "sign": 0.14051131065934896, "keep": 0.061320280423387885, "leak": 0.5674372985959053, "replacement": 0.4752013832330704}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 8.494545936584473, "loss_scale": 4.0, "retries": 0, "seconds": 523.6, "max_memory_gb": 24.64, "pair": 0.6536932364106178, "sign": 0.6981230527162552, "keep": 0.3714708797633648, "leak": 0.2848746273375582, "replacement": 1.2597088068723679}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 5.636983394622803, "loss_scale": 4.0, "retries": 0, "seconds": 524.9, "max_memory_gb": 24.64, "pair": 0.3002062328159809, "sign": 0.24074299447238445, "keep": 0.07734371814876795, "leak": 2.122120081912726, "replacement": 0.989779606461525}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 10.719558715820312, "loss_scale": 4.0, "retries": 0, "seconds": 526.1, "max_memory_gb": 24.64, "pair": 0.3418569676578045, "sign": 0.14667190611362457, "keep": 0.063713533221744, "leak": 0.2144311647862196, "replacement": 0.5006574839353561}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 13.275468826293945, "loss_scale": 4.0, "retries": 0, "seconds": 527.3, "max_memory_gb": 24.64, "pair": 0.3730385862290859, "sign": 0.09259833581745625, "keep": 0.042043220018967986, "leak": 0.21045631408924237, "replacement": 0.4929734468460083}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 3.5113863945007324, "loss_scale": 4.0, "retries": 0, "seconds": 528.6, "max_memory_gb": 24.64, "pair": 0.26051221042871475, "sign": 0.24066914478316903, "keep": 0.21567688544746488, "leak": 0.5434855185449123, "replacement": 0.6245565935969353}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 5.206751346588135, "loss_scale": 4.0, "retries": 0, "seconds": 529.8, "max_memory_gb": 24.64, "pair": 0.15333322808146477, "sign": 0.06784902838990092, "keep": 0.10785800986923277, "leak": 1.2164714497630484, "replacement": 0.545304611325264}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 3.3237662315368652, "loss_scale": 4.0, "retries": 0, "seconds": 531.1, "max_memory_gb": 24.64, "pair": 0.12337969616055489, "sign": 0.15228486713021994, "keep": 0.08478497574105859, "leak": 0.3982586532365531, "replacement": 0.3414792772382498}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 5.053582191467285, "loss_scale": 4.0, "retries": 0, "seconds": 532.3, "max_memory_gb": 24.64, "pair": 0.14841767586767673, "sign": 0.21556999068707228, "keep": 0.06202894542366266, "leak": 1.4019078984856606, "replacement": 0.6376940906047821}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 3.0654683113098145, "loss_scale": 4.0, "retries": 0, "seconds": 533.5, "max_memory_gb": 24.64, "pair": 0.20342259854078293, "sign": 0.14777521789073944, "keep": 0.02350389934144914, "leak": 0.15718447719700634, "replacement": 0.3283582776784897}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 16.596879959106445, "loss_scale": 4.0, "retries": 0, "seconds": 534.7, "max_memory_gb": 24.64, "pair": 0.49300423241220415, "sign": 0.489755738992244, "keep": 0.6455595009028912, "leak": 0.6142615973949432, "replacement": 1.2142272293567657}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 6.508694171905518, "loss_scale": 4.0, "retries": 0, "seconds": 535.9, "max_memory_gb": 24.64, "pair": 0.27655032835900784, "sign": 0.27786949649453163, "keep": 0.10878503881394863, "leak": 1.0957926139235497, "replacement": 0.7438257187604904}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 11.523204803466797, "loss_scale": 4.0, "retries": 0, "seconds": 537.2, "max_memory_gb": 24.64, "pair": 0.281644981354475, "sign": 0.5048913350328803, "keep": 0.6483191344887018, "leak": 1.2327236831188202, "replacement": 1.1664311438798904}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 3.1114771366119385, "loss_scale": 4.0, "retries": 0, "seconds": 538.4, "max_memory_gb": 24.64, "pair": 0.12585727497935295, "sign": 0.08443846367299557, "keep": 0.09359783516265452, "leak": 0.5008695940487087, "replacement": 0.34009283035993576}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 7.272437572479248, "loss_scale": 4.0, "retries": 0, "seconds": 539.6, "max_memory_gb": 24.64, "pair": 0.4166786000132561, "sign": 0.4617631481960416, "keep": 0.5800861958414316, "leak": 0.35296236211434007, "replacement": 1.0258438736200333}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 3.9324474334716797, "loss_scale": 4.0, "retries": 0, "seconds": 540.9, "max_memory_gb": 24.64, "pair": 0.16651427000761032, "sign": 0.28543660044670105, "keep": 0.013780252018477768, "leak": 0.8009648472070694, "replacement": 0.5163638889789581}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 15.635940551757812, "loss_scale": 4.0, "retries": 0, "seconds": 542.2, "max_memory_gb": 24.64, "pair": 0.7351566851139069, "sign": 0.8581709247082472, "keep": 1.2431467122223694, "leak": 1.4590761959552765, "replacement": 2.1505845487117767}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 3.765800714492798, "loss_scale": 4.0, "retries": 0, "seconds": 543.4, "max_memory_gb": 24.64, "pair": 0.4101465009152889, "sign": 0.4175212811678648, "keep": 0.10604453762061894, "leak": 1.3979140012525022, "replacement": 1.0214079171419144}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 4.851339340209961, "loss_scale": 4.0, "retries": 0, "seconds": 544.6, "max_memory_gb": 24.64, "pair": 0.2519134059548378, "sign": 0.29133183509111404, "keep": 0.04474365233909339, "leak": 0.7012213855050504, "replacement": 0.595256507396698}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 8.714685440063477, "loss_scale": 4.0, "retries": 0, "seconds": 545.8, "max_memory_gb": 24.64, "pair": 0.2668985743075609, "sign": 0.4061567385215312, "keep": 0.23362290335353464, "leak": 0.5624791979789734, "replacement": 0.7274081856012344}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 6.856657028198242, "loss_scale": 4.0, "retries": 0, "seconds": 547.1, "max_memory_gb": 24.64, "pair": 0.6504211463034153, "sign": 0.8331835307180882, "keep": 0.8535705115064047, "leak": 0.7517308189999312, "replacement": 1.6817309111356735}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 9.064830780029297, "loss_scale": 4.0, "retries": 0, "seconds": 548.3, "max_memory_gb": 24.64, "pair": 0.18421903625130653, "sign": 0.22807441093027592, "keep": 0.24998850421980023, "leak": 1.1095110550522804, "replacement": 0.700628250837326}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 5.70883846282959, "loss_scale": 4.0, "retries": 0, "seconds": 549.5, "max_memory_gb": 24.64, "pair": 0.4128056913614273, "sign": 0.5484941359609365, "keep": 0.31316392961889505, "leak": 1.632846686614812, "replacement": 1.251846432685852}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 4.2925591468811035, "loss_scale": 4.0, "retries": 0, "seconds": 550.7, "max_memory_gb": 24.64, "pair": 0.15996852330863476, "sign": 0.21357025392353535, "keep": 0.1369953490793705, "leak": 0.6067673712968826, "replacement": 0.48694314807653427}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 4.234861373901367, "loss_scale": 4.0, "retries": 0, "seconds": 551.9, "max_memory_gb": 24.64, "pair": 0.1859890427440405, "sign": 0.16173379495739937, "keep": 0.05056118778884411, "leak": 0.8486132360994816, "replacement": 0.504289835691452}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 4.147482395172119, "loss_scale": 4.0, "retries": 0, "seconds": 553.1, "max_memory_gb": 24.64, "pair": 0.26250865682959557, "sign": 0.14806834049522877, "keep": 0.22993954340927303, "leak": 0.9276795005425811, "replacement": 0.6834324672818184}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 4.856068134307861, "loss_scale": 4.0, "retries": 0, "seconds": 554.4, "max_memory_gb": 24.64, "pair": 0.14899531938135624, "sign": 0.12009076587855816, "keep": 0.02870974875986576, "leak": 1.2189547345042229, "replacement": 0.5281342640519142}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 5.529736042022705, "loss_scale": 4.0, "retries": 0, "seconds": 555.6, "max_memory_gb": 24.64, "pair": 0.31023186072707176, "sign": 0.31600244902074337, "keep": 0.4918120251968503, "leak": 1.2398663237690926, "replacement": 1.0241056829690933}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 21.287778854370117, "loss_scale": 4.0, "retries": 0, "seconds": 556.9, "max_memory_gb": 24.64, "pair": 0.311267364770174, "sign": 0.32175695057958364, "keep": 0.1049409705447033, "leak": 1.2704284461215138, "replacement": 0.8422233983874321}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 18.3742618560791, "loss_scale": 4.0, "retries": 0, "seconds": 558.1, "max_memory_gb": 24.64, "pair": 0.428222831338644, "sign": 0.2780712563544512, "keep": 0.3824716731905937, "leak": 0.9103313107043505, "replacement": 0.986077070236206}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 4.632754802703857, "loss_scale": 4.0, "retries": 0, "seconds": 559.3, "max_memory_gb": 24.64, "pair": 0.1421564668416977, "sign": 0.17630359157919884, "keep": 0.24957378720864654, "leak": 2.0161485001444817, "replacement": 0.8591322898864746}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 6.117729663848877, "loss_scale": 4.0, "retries": 0, "seconds": 560.5, "max_memory_gb": 24.64, "pair": 0.36051399260759354, "sign": 0.6766923302784562, "keep": 0.05287023354321718, "leak": 1.1776820812374353, "replacement": 1.0197157561779022}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 2.7021913528442383, "loss_scale": 4.0, "retries": 0, "seconds": 561.7, "max_memory_gb": 24.64, "pair": 0.14847546815872192, "sign": 0.21954751387238503, "keep": 0.022050845494959503, "leak": 0.3726674630306661, "replacement": 0.3624415248632431}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 8.753437042236328, "loss_scale": 4.0, "retries": 0, "seconds": 563.0, "max_memory_gb": 24.64, "pair": 0.32260674238204956, "sign": 0.4873523749411106, "keep": 0.3138528857380152, "leak": 0.9160746255828371, "replacement": 0.9522280395030975}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 2.8914456367492676, "loss_scale": 4.0, "retries": 0, "seconds": 564.2, "max_memory_gb": 24.64, "pair": 0.2558785695582628, "sign": 0.259190471842885, "keep": 0.055662248749285936, "leak": 0.23445047717541456, "replacement": 0.4719175472855568}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 2.7092134952545166, "loss_scale": 4.0, "retries": 0, "seconds": 565.5, "max_memory_gb": 24.64, "pair": 0.2018601205199957, "sign": 0.22101913578808308, "keep": 0.02697361819446087, "leak": 0.2222992079332471, "replacement": 0.38143130391836166}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 11.928271293640137, "loss_scale": 4.0, "retries": 0, "seconds": 566.7, "max_memory_gb": 24.64, "pair": 0.256646478548646, "sign": 0.5753154307603836, "keep": 0.6604749499820173, "leak": 1.1529393047094345, "replacement": 1.1627764850854874}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 3.5142126083374023, "loss_scale": 4.0, "retries": 0, "seconds": 568.0, "max_memory_gb": 24.64, "pair": 0.15269838087260723, "sign": 0.210683717392385, "keep": 0.03145324181241449, "leak": 0.5047693898668513, "replacement": 0.3999591991305351}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 3.575293779373169, "loss_scale": 4.0, "retries": 0, "seconds": 569.2, "max_memory_gb": 24.64, "pair": 0.22362649254500866, "sign": 0.18939408659934998, "keep": 0.022735767590347677, "leak": 0.18904208950698376, "replacement": 0.3769519403576851}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 6.073878765106201, "loss_scale": 4.0, "retries": 0, "seconds": 570.4, "max_memory_gb": 24.64, "pair": 0.1559290513396263, "sign": 0.1488033877685666, "keep": 0.017200999951455742, "leak": 1.5996204055845737, "replacement": 0.6388363614678383}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 5.204129219055176, "loss_scale": 4.0, "retries": 0, "seconds": 571.7, "max_memory_gb": 24.64, "pair": 0.12320092041045427, "sign": 0.13410132937133312, "keep": 0.141161777661182, "leak": 1.3042713552713394, "replacement": 0.586900319904089}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 10.664632797241211, "loss_scale": 4.0, "retries": 0, "seconds": 572.9, "max_memory_gb": 24.64, "pair": 0.1412164643406868, "sign": 0.2128316224552691, "keep": 0.08497096272185445, "leak": 1.7219772785902023, "replacement": 0.720612071454525}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 4.483395099639893, "loss_scale": 4.0, "retries": 0, "seconds": 574.1, "max_memory_gb": 24.64, "pair": 0.16429996117949486, "sign": 0.12178068701177835, "keep": 0.04620687570422888, "leak": 0.7692094501107931, "replacement": 0.4405961111187935}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 4.293485641479492, "loss_scale": 4.0, "retries": 0, "seconds": 575.4, "max_memory_gb": 24.64, "pair": 0.181201565079391, "sign": 0.1571407769806683, "keep": 0.01603916662861593, "leak": 0.36678942234721035, "replacement": 0.3594888783991337}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 13.212162017822266, "loss_scale": 4.0, "retries": 0, "seconds": 576.6, "max_memory_gb": 24.64, "pair": 0.4820855278521776, "sign": 0.1887708082795143, "keep": 0.3283698436571285, "leak": 0.9281965750415111, "replacement": 0.9727049991488457}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 4.196152687072754, "loss_scale": 4.0, "retries": 0, "seconds": 577.8, "max_memory_gb": 24.64, "pair": 0.2279902994632721, "sign": 0.0921152988448739, "keep": 0.019880696025211364, "leak": 0.2096275220392272, "replacement": 0.33639518544077873}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 15.53243350982666, "loss_scale": 4.0, "retries": 0, "seconds": 579.0, "max_memory_gb": 24.64, "pair": 0.3937729597091675, "sign": 0.6124807894229889, "keep": 0.7898773038759828, "leak": 0.9130389355123043, "replacement": 1.3232117891311646}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 10.75537395477295, "loss_scale": 4.0, "retries": 0, "seconds": 580.2, "max_memory_gb": 24.64, "pair": 0.3330031605437398, "sign": 0.3757958812639117, "keep": 0.14611413015518337, "leak": 0.7514880681410432, "replacement": 0.7818301990628242}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 9.86918830871582, "loss_scale": 4.0, "retries": 0, "seconds": 581.4, "max_memory_gb": 24.64, "pair": 0.2099636010825634, "sign": 0.13779371418058872, "keep": 0.32116037979722023, "leak": 0.12995836790651083, "replacement": 0.47193023562431335}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 11.854084968566895, "loss_scale": 4.0, "retries": 0, "seconds": 582.7, "max_memory_gb": 24.64, "pair": 0.3543781153857708, "sign": 0.4902700036764145, "keep": 0.5878239932353608, "leak": 0.6934291981160641, "replacement": 1.0667823925614357}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 12.692441940307617, "loss_scale": 4.0, "retries": 0, "seconds": 583.9, "max_memory_gb": 24.64, "pair": 0.3407485857605934, "sign": 0.4515110142529011, "keep": 0.46253243496175855, "leak": 0.5302619721114752, "replacement": 0.9303358495235443}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 4.677561283111572, "loss_scale": 4.0, "retries": 0, "seconds": 585.2, "max_memory_gb": 24.64, "pair": 0.21567754354327917, "sign": 0.18961854465305805, "keep": 0.05412619619164616, "leak": 1.833789512515068, "replacement": 0.7959972769021988}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 6.0527567863464355, "loss_scale": 4.0, "retries": 0, "seconds": 586.4, "max_memory_gb": 24.64, "pair": 0.294288270175457, "sign": 0.4424978932365775, "keep": 0.026055773138068616, "leak": 0.48438679799437523, "replacement": 0.6496618241071701}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 7.985474586486816, "loss_scale": 4.0, "retries": 0, "seconds": 587.6, "max_memory_gb": 24.64, "pair": 0.34119547717273235, "sign": 0.27288617426529527, "keep": 0.035560388350859284, "leak": 0.3527688618050888, "replacement": 0.5836109891533852}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 8.79236888885498, "loss_scale": 4.0, "retries": 0, "seconds": 588.9, "max_memory_gb": 24.64, "pair": 0.5162448547780514, "sign": 0.5171555355191231, "keep": 0.5234454700257629, "leak": 0.4831211178097874, "replacement": 1.1573256626725197}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 5.275903701782227, "loss_scale": 4.0, "retries": 0, "seconds": 590.1, "max_memory_gb": 24.64, "pair": 0.35888537485152483, "sign": 0.4970719553530216, "keep": 0.2900441661477089, "leak": 0.8941706900950521, "replacement": 0.9759861454367638}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 4.100348949432373, "loss_scale": 4.0, "retries": 0, "seconds": 591.3, "max_memory_gb": 24.64, "pair": 0.13544126972556114, "sign": 0.11853834055364132, "keep": 0.17045443638926372, "leak": 0.5771804843097925, "replacement": 0.42423277720808983}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 4.880549430847168, "loss_scale": 4.0, "retries": 0, "seconds": 592.5, "max_memory_gb": 24.64, "pair": 0.1804082179442048, "sign": 0.3214128911495209, "keep": 0.12167974247131497, "leak": 0.5805017473176122, "replacement": 0.5470799803733826}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 14.791101455688477, "loss_scale": 4.0, "retries": 0, "seconds": 593.7, "max_memory_gb": 24.64, "pair": 0.5040410310029984, "sign": 0.7828711867332458, "keep": 0.5519787967205048, "leak": 0.34158913558349013, "replacement": 1.2568633407354355}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 19.050159454345703, "loss_scale": 4.0, "retries": 0, "seconds": 594.9, "max_memory_gb": 24.64, "pair": 0.841932013630867, "sign": 1.1144610177725554, "keep": 1.7261481918394566, "leak": 0.7409885744564235, "replacement": 2.4474836885929108}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 14.09290885925293, "loss_scale": 4.0, "retries": 0, "seconds": 596.2, "max_memory_gb": 24.64, "pair": 0.27838306687772274, "sign": 0.5438339225947857, "keep": 0.02749011293053627, "leak": 0.9316820539534092, "replacement": 0.7969656139612198}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 11.026484489440918, "loss_scale": 4.0, "retries": 0, "seconds": 597.3, "max_memory_gb": 24.64, "pair": 0.33323627104982734, "sign": 0.5816329009830952, "keep": 0.6771187935955822, "leak": 0.6301164571195841, "replacement": 1.1201412007212639}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 3.9845242500305176, "loss_scale": 4.0, "retries": 0, "seconds": 598.5, "max_memory_gb": 24.64, "pair": 0.15585742890834808, "sign": 0.14465260389260948, "keep": 0.0695202253991738, "leak": 0.5668444798211567, "replacement": 0.4046549443155527}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 5.263822555541992, "loss_scale": 4.0, "retries": 0, "seconds": 599.8, "max_memory_gb": 24.64, "pair": 0.2608781699091196, "sign": 0.2866546716541052, "keep": 0.11215205769985914, "leak": 1.63410335034132, "replacement": 0.8688073754310608}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 10.543790817260742, "loss_scale": 4.0, "retries": 0, "seconds": 601.0, "max_memory_gb": 24.64, "pair": 0.11572024691849947, "sign": 0.2267571073025465, "keep": 0.7078039260813966, "leak": 2.0491844452917576, "replacement": 1.095296859741211}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 11.463717460632324, "loss_scale": 4.0, "retries": 0, "seconds": 602.2, "max_memory_gb": 24.64, "pair": 0.16816718596965075, "sign": 0.23966665752232075, "keep": 0.032595102151390165, "leak": 0.9670211896300316, "replacement": 0.5460533648729324}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 13.55606746673584, "loss_scale": 4.0, "retries": 0, "seconds": 603.4, "max_memory_gb": 24.64, "pair": 0.22418133728206158, "sign": 0.3731855694204569, "keep": 0.28744514257414266, "leak": 1.3350460454821587, "replacement": 0.8882582113146782}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 4.406633377075195, "loss_scale": 4.0, "retries": 0, "seconds": 604.6, "max_memory_gb": 24.64, "pair": 0.16574815660715103, "sign": 0.14216794446110725, "keep": 0.031441399711184204, "leak": 0.8742539659142494, "replacement": 0.4711163267493248}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 12.548946380615234, "loss_scale": 4.0, "retries": 0, "seconds": 605.9, "max_memory_gb": 24.64, "pair": 0.557478740811348, "sign": 0.8789912536740303, "keep": 0.5054897627560422, "leak": 0.9936337918043137, "replacement": 1.4981276541948318}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 6.324710369110107, "loss_scale": 4.0, "retries": 0, "seconds": 607.1, "max_memory_gb": 24.64, "pair": 0.09457219252362847, "sign": 0.21032768860459328, "keep": 0.030486226547509432, "leak": 1.2282252482837066, "replacement": 0.5220354571938515}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 5.7347917556762695, "loss_scale": 4.0, "retries": 0, "seconds": 608.4, "max_memory_gb": 24.64, "pair": 0.19233534671366215, "sign": 0.24004840524867177, "keep": 0.03433768975082785, "leak": 0.7455164566636086, "replacement": 0.5159075036644936}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 3.4043571949005127, "loss_scale": 4.0, "retries": 0, "seconds": 609.6, "max_memory_gb": 24.64, "pair": 0.21466616913676262, "sign": 0.07952882815152407, "keep": 0.08285741275176406, "leak": 0.6009448704717215, "replacement": 0.4460955187678337}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 18.83669090270996, "loss_scale": 4.0, "retries": 0, "seconds": 610.8, "max_memory_gb": 24.64, "pair": 0.4647381976246834, "sign": 0.6551977433264256, "keep": 0.9692845717654563, "leak": 1.2855253047309816, "replacement": 1.5983606427907944}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 5.639629364013672, "loss_scale": 4.0, "retries": 0, "seconds": 612.0, "max_memory_gb": 24.64, "pair": 0.4148948863148689, "sign": 0.38778193946927786, "keep": 0.26913384278304875, "leak": 0.6408970542252064, "replacement": 0.9035770446062088}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 20.00026512145996, "loss_scale": 4.0, "retries": 0, "seconds": 613.2, "max_memory_gb": 24.64, "pair": 0.5227822754532099, "sign": 0.48665722645819187, "keep": 0.7290170646738261, "leak": 0.9854736207053065, "replacement": 1.376987785100937}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 20.243513107299805, "loss_scale": 4.0, "retries": 0, "seconds": 614.4, "max_memory_gb": 24.64, "pair": 0.2995935417711735, "sign": 0.5111198592931032, "keep": 0.7871538982144557, "leak": 1.1523776911199093, "replacement": 1.2368248328566551}
diff --git a/runs/single_llava/replacement/L14_s13/adapter/README.md b/runs/single_llava/replacement/L14_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L14_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L14_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L14_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..39ffd83c5c09a7b7e87eefd89efc68de3b1e1b9b
--- /dev/null
+++ b/runs/single_llava/replacement/L14_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.14.self_attn.k_proj",
+    "model.language_model.layers.14.self_attn.v_proj",
+    "model.language_model.layers.14.self_attn.q_proj",
+    "model.language_model.layers.14.mlp.up_proj",
+    "model.language_model.layers.14.mlp.gate_proj",
+    "model.language_model.layers.14.self_attn.o_proj",
+    "model.language_model.layers.14.mlp.down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L14_s13/config.json b/runs/single_llava/replacement/L14_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..a0bff80b664c28824b14ffc8e58c7e11ed909a22
--- /dev/null
+++ b/runs/single_llava/replacement/L14_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 14,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L14_s13/steps.jsonl b/runs/single_llava/replacement/L14_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2c52e0ec418cd932b13cc352a3b173b4d7dd0660
--- /dev/null
+++ b/runs/single_llava/replacement/L14_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.5991926193237305, "loss_scale": 4.0, "retries": 0, "seconds": 1.3, "max_memory_gb": 22.82, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.9142074584960938, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 22.82, "pair": 1.0985039174556732, "sign": 0.5366511680185795, "keep": 5.372033937368542, "leak": 3.402146679665208e-05, "replacement": 4.052854955196381}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.3787580728530884, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 22.88, "pair": 1.0966259837150574, "sign": 0.5406990805640817, "keep": 10.804192781448364, "leak": 5.791018384115887e-05, "replacement": 6.769086480140686}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 1.085868000984192, "loss_scale": 4.0, "retries": 0, "seconds": 4.8, "max_memory_gb": 22.89, "pair": 1.0895086526870728, "sign": 0.4702874980866909, "keep": 7.673395658843219, "leak": 7.246398399729514e-05, "replacement": 5.16136834025383}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.6853387355804443, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 22.89, "pair": 1.0956873893737793, "sign": 0.7755896300077438, "keep": 0.1583336926996708, "leak": 1.869013320998647e-06, "replacement": 1.562649518251419}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.7392383813858032, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 22.89, "pair": 1.093985915184021, "sign": 0.7363318353891373, "keep": 2.9338733702898026, "leak": 7.792526605499006e-05, "replacement": 2.9291079938411713}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 1.2835757732391357, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 22.89, "pair": 1.0743320286273956, "sign": 0.545250590890646, "keep": 6.879353540018201, "leak": 0.0005140772482263856, "replacement": 4.786762684583664}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 1.1388108730316162, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 22.89, "pair": 1.0806901156902313, "sign": 0.5854592025279999, "keep": 7.386370852589607, "leak": 0.0003177746630171896, "replacement": 5.0666844844818115}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 1.2130314111709595, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 22.89, "pair": 1.069080412387848, "sign": 0.6164807924069464, "keep": 5.705232873558998, "leak": 0.0010711166937760197, "replacement": 4.230204939842224}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.9354776740074158, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 22.89, "pair": 1.0731874406337738, "sign": 0.7774146199226379, "keep": 3.273608475923538, "leak": 0.0007291097590496065, "replacement": 3.098881274461746}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 1.5040595531463623, "loss_scale": 4.0, "retries": 0, "seconds": 12.8, "max_memory_gb": 22.89, "pair": 1.0438829064369202, "sign": 0.44932313822209835, "keep": 5.262729149311781, "leak": 0.0025010499193740543, "replacement": 3.9005341827869415}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 1.150463581085205, "loss_scale": 4.0, "retries": 0, "seconds": 13.9, "max_memory_gb": 22.89, "pair": 1.0596527457237244, "sign": 0.6058684289455414, "keep": 3.0533935176208615, "leak": 0.00195355485084292, "replacement": 2.8897719979286194}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 1.3146607875823975, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 22.89, "pair": 1.057394415140152, "sign": 0.502945814281702, "keep": 5.346154639497399, "leak": 0.0017450564803631607, "replacement": 3.9823808670043945}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.7388596534729004, "loss_scale": 4.0, "retries": 0, "seconds": 16.1, "max_memory_gb": 22.89, "pair": 1.0257053673267365, "sign": 0.3927103206515312, "keep": 7.789699614048004, "leak": 0.0038454140594694763, "replacement": 5.117871701717377}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 1.1958190202713013, "loss_scale": 4.0, "retries": 0, "seconds": 17.2, "max_memory_gb": 22.89, "pair": 1.0441310703754425, "sign": 0.685480672866106, "keep": 5.176168979378417, "leak": 0.0036622003244701773, "replacement": 3.9758715331554413}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 1.1804537773132324, "loss_scale": 4.0, "retries": 0, "seconds": 18.4, "max_memory_gb": 22.89, "pair": 1.0586864650249481, "sign": 0.7830073237419128, "keep": 2.7618035776540637, "leak": 0.0024065888283075765, "replacement": 2.83169361948967}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 1.002394199371338, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 22.89, "pair": 1.0499733090400696, "sign": 0.7035036683082581, "keep": 3.184420813806355, "leak": 0.00362557345215464, "replacement": 2.9948418736457825}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.731688141822815, "loss_scale": 4.0, "retries": 0, "seconds": 20.6, "max_memory_gb": 22.95, "pair": 1.0181877315044403, "sign": 0.5229737497866154, "keep": 7.1801096349954605, "leak": 0.004492501029744744, "replacement": 4.870852559804916}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 2.6807494163513184, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 22.95, "pair": 0.9570626020431519, "sign": 0.5136829130351543, "keep": 10.693816184997559, "leak": 0.015872953226789832, "replacement": 6.564780354499817}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 1.1203669309616089, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 22.95, "pair": 1.0496065616607666, "sign": 0.421068012714386, "keep": 0.02129824855364859, "leak": 0.0027164602652192116, "replacement": 1.2714687883853912}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.7617837190628052, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 22.95, "pair": 0.9848789572715759, "sign": 0.35962190479040146, "keep": 5.5437145926989615, "leak": 0.005035257228882983, "replacement": 3.9378058314323425}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.731272578239441, "loss_scale": 4.0, "retries": 0, "seconds": 25.2, "max_memory_gb": 22.95, "pair": 0.9886856824159622, "sign": 0.5606896281242371, "keep": 5.910903550684452, "leak": 0.01929728445247747, "replacement": 4.229306608438492}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 2.2061879634857178, "loss_scale": 4.0, "retries": 0, "seconds": 26.4, "max_memory_gb": 22.95, "pair": 0.9004480987787247, "sign": 0.2323848595842719, "keep": 5.302761947270483, "leak": 0.03375839098589495, "replacement": 3.6764609813690186}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 3.044767379760742, "loss_scale": 4.0, "retries": 0, "seconds": 27.5, "max_memory_gb": 22.95, "pair": 0.887681171298027, "sign": 0.2365762572735548, "keep": 10.014763832092285, "leak": 0.017005254281684756, "replacement": 6.0176026821136475}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 2.4240224361419678, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 22.95, "pair": 0.9188453257083893, "sign": 0.37386222556233406, "keep": 8.457780607044697, "leak": 0.03125740083032724, "replacement": 5.342481046915054}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 3.4728429317474365, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 22.95, "pair": 0.8297281861305237, "sign": 0.632890154607594, "keep": 10.455266833305359, "leak": 0.057117938064038754, "replacement": 6.388086199760437}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 2.5275702476501465, "loss_scale": 4.0, "retries": 0, "seconds": 31.0, "max_memory_gb": 22.95, "pair": 0.9146724492311478, "sign": 0.5677588246762753, "keep": 8.390579357743263, "leak": 0.029334087652387097, "replacement": 5.401175081729889}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 2.052077054977417, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 22.95, "pair": 0.9603479504585266, "sign": 0.7574613373726606, "keep": 5.606862485408783, "leak": 0.028567709758817728, "replacement": 4.149651765823364}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.940781831741333, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 22.95, "pair": 0.9519694894552231, "sign": 0.8218666054308414, "keep": 5.191200259141624, "leak": 0.032959854244836606, "replacement": 3.9667429327964783}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 2.236187696456909, "loss_scale": 4.0, "retries": 0, "seconds": 34.4, "max_memory_gb": 22.95, "pair": 0.8179146945476532, "sign": 0.32565889693796635, "keep": 5.495093243196607, "leak": 0.09743231372704031, "replacement": 3.7526489198207855}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 2.1647918224334717, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 22.95, "pair": 0.9335073977708817, "sign": 0.7883361950516701, "keep": 5.634904190897942, "leak": 0.04635948629527675, "replacement": 4.156717509031296}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 3.6703710556030273, "loss_scale": 4.0, "retries": 0, "seconds": 36.7, "max_memory_gb": 22.95, "pair": 0.82974012196064, "sign": 0.3506581149995327, "keep": 10.658146858215332, "leak": 0.0701348939910531, "replacement": 6.351676344871521}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 2.2047362327575684, "loss_scale": 4.0, "retries": 0, "seconds": 37.9, "max_memory_gb": 22.95, "pair": 0.8676363825798035, "sign": 0.5892255306243896, "keep": 4.847874394617975, "leak": 0.08816831511899181, "replacement": 3.6082282960414886}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 2.5436699390411377, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 22.95, "pair": 0.8691514581441879, "sign": 0.49062757194042206, "keep": 5.215794896706939, "leak": 0.03575900546275079, "replacement": 3.7313024699687958}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 2.6035003662109375, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 22.95, "pair": 0.7658327221870422, "sign": 0.4039963958784938, "keep": 5.385564690455794, "leak": 0.1676868472713977, "replacement": 3.702535003423691}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 2.523458957672119, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 22.95, "pair": 0.764580100774765, "sign": 0.39714038372039795, "keep": 5.121713884174824, "leak": 0.1327051913831383, "replacement": 3.5571835041046143}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 2.634589672088623, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 22.95, "pair": 0.7919788062572479, "sign": 0.550264947116375, "keep": 4.848756022751331, "leak": 0.14153326319501502, "replacement": 3.5268725752830505}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 4.0713114738464355, "loss_scale": 4.0, "retries": 0, "seconds": 43.6, "max_memory_gb": 22.95, "pair": 0.716557189822197, "sign": 0.7211557114496827, "keep": 9.344773173332214, "leak": 0.15938520757481456, "replacement": 5.789367914199829}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.9790477752685547, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 22.95, "pair": 0.8756834864616394, "sign": 0.5224080979824066, "keep": 2.8244060673750937, "leak": 0.1247925260104239, "replacement": 2.580288738012314}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 3.0090179443359375, "loss_scale": 4.0, "retries": 0, "seconds": 45.9, "max_memory_gb": 22.95, "pair": 0.7733039706945419, "sign": 0.4413276258856058, "keep": 5.459298424422741, "leak": 0.09429560275748372, "replacement": 3.747190982103348}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.510615348815918, "loss_scale": 4.0, "retries": 0, "seconds": 47.0, "max_memory_gb": 22.95, "pair": 0.9710676670074463, "sign": 0.9838375151157379, "keep": 2.926800824701786, "leak": 0.04215434932393691, "replacement": 2.9369254112243652}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 3.5685558319091797, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 22.95, "pair": 0.7958808541297913, "sign": 0.3599077630788088, "keep": 4.460013517178595, "leak": 0.08933502528816462, "replacement": 3.2281752228736877}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 3.5136637687683105, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 22.95, "pair": 0.7244508117437363, "sign": 0.45679790899157524, "keep": 4.655455057974905, "leak": 0.12714619003236294, "replacement": 3.3123638033866882}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.843137264251709, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 22.95, "pair": 0.7884533181786537, "sign": 0.5180503688752651, "keep": 2.6816541478037834, "leak": 0.1895736800506711, "replacement": 2.435698986053467}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 2.106980085372925, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 22.95, "pair": 0.8948422372341156, "sign": 0.8375783115625381, "keep": 2.694028116762638, "leak": 0.08083888189867139, "replacement": 2.6808550357818604}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 3.647311210632324, "loss_scale": 4.0, "retries": 0, "seconds": 52.7, "max_memory_gb": 22.95, "pair": 0.7632662504911423, "sign": 0.6500815488398075, "keep": 4.315598294138908, "leak": 0.19721874000970274, "replacement": 3.295410990715027}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 3.9803333282470703, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 22.95, "pair": 0.6807746216654778, "sign": 0.5807494483888149, "keep": 3.802183199673891, "leak": 0.2247648129123263, "replacement": 2.928432196378708}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 6.297522068023682, "loss_scale": 4.0, "retries": 0, "seconds": 55.0, "max_memory_gb": 22.95, "pair": 0.5416627526283264, "sign": 0.1612535109743476, "keep": 7.0028295293450356, "leak": 0.33425976894795895, "replacement": 4.207269132137299}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 5.636756420135498, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 22.95, "pair": 0.6077638603746891, "sign": 0.6041108816862106, "keep": 6.56655627861619, "leak": 0.4827355341985822, "replacement": 4.313781350851059}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 6.6783552169799805, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 22.95, "pair": 0.6865140944719315, "sign": 0.36217620968818665, "keep": 6.593759838491678, "leak": 0.16599907912313938, "replacement": 4.205981761217117}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 2.0692105293273926, "loss_scale": 4.0, "retries": 0, "seconds": 58.4, "max_memory_gb": 22.95, "pair": 0.7573558539152145, "sign": 0.5279144998639822, "keep": 2.5622575394809246, "leak": 0.2574426995124668, "replacement": 2.366802453994751}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 5.158361911773682, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 22.95, "pair": 0.6682372055947781, "sign": 0.6333186542615294, "keep": 4.210429944097996, "leak": 0.43903268221765757, "replacement": 3.1998696327209473}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 6.838608741760254, "loss_scale": 4.0, "retries": 0, "seconds": 60.7, "max_memory_gb": 22.95, "pair": 0.6588774025440216, "sign": 0.628235325217247, "keep": 6.733122251927853, "leak": 0.35088627063669264, "replacement": 4.42727792263031}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 8.22369384765625, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 22.95, "pair": 0.5933775156736374, "sign": 0.5394351035356522, "keep": 6.776638001203537, "leak": 0.5471752025187016, "replacement": 4.388208031654358}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 6.086239814758301, "loss_scale": 4.0, "retries": 0, "seconds": 63.0, "max_memory_gb": 22.95, "pair": 0.578855998814106, "sign": 0.3168327659368515, "keep": 3.9942664878908545, "leak": 0.3957134885713458, "replacement": 2.833334118127823}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 12.991168975830078, "loss_scale": 4.0, "retries": 0, "seconds": 64.1, "max_memory_gb": 22.95, "pair": 0.2985336408019066, "sign": 0.10165905393660069, "keep": 7.687397718429565, "leak": 0.6520022749900818, "replacement": 4.356062531471252}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 3.311541795730591, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 22.95, "pair": 0.7704325392842293, "sign": 0.7821719907224178, "keep": 2.071586008183658, "leak": 0.13939259760081768, "replacement": 2.2321596145629883}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 6.755146503448486, "loss_scale": 4.0, "retries": 0, "seconds": 66.3, "max_memory_gb": 22.95, "pair": 0.6316450238227844, "sign": 0.22843232192099094, "keep": 3.7279791076434776, "leak": 0.5283942278474569, "replacement": 2.7419492453336716}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 4.785445213317871, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 22.95, "pair": 0.7098532393574715, "sign": 0.5652378052473068, "keep": 2.2432964891195297, "leak": 0.11024988444478367, "replacement": 2.141682878136635}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 7.47664737701416, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 22.95, "pair": 0.5582337416708469, "sign": 0.575217429548502, "keep": 3.796162521466613, "leak": 0.4111012648791075, "replacement": 2.846698969602585}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 12.720844268798828, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 22.95, "pair": 0.5233310833573341, "sign": 0.6348795164376497, "keep": 4.593783438205719, "leak": 0.894899081758922, "replacement": 3.361387252807617}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 5.4037580490112305, "loss_scale": 4.0, "retries": 0, "seconds": 70.8, "max_memory_gb": 22.95, "pair": 0.7230567708611488, "sign": 0.63825305365026, "keep": 1.9964863592758775, "leak": 0.2183572077192366, "replacement": 2.095015823841095}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 15.261811256408691, "loss_scale": 4.0, "retries": 0, "seconds": 72.0, "max_memory_gb": 22.95, "pair": 0.5682177543640137, "sign": 0.4998480938374996, "keep": 4.950845065526664, "leak": 0.2313218671188224, "replacement": 3.351394772529602}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 15.52970027923584, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 23.0, "pair": 0.39073391631245613, "sign": 0.3023370548617095, "keep": 3.921626629307866, "leak": 1.3067841120064259, "replacement": 2.829411804676056}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 15.38177490234375, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 23.0, "pair": 0.5807005763053894, "sign": 0.47494390048086643, "keep": 4.515529990196228, "leak": 0.48173303657677025, "replacement": 3.196370840072632}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 3.827949285507202, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 23.0, "pair": 0.7879573255777359, "sign": 0.521391898393631, "keep": 1.3970900662243366, "leak": 0.13100378937087953, "replacement": 1.7799492329359055}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 14.376801490783691, "loss_scale": 4.0, "retries": 0, "seconds": 76.6, "max_memory_gb": 23.0, "pair": 0.454935897141695, "sign": 0.345485083758831, "keep": 3.0160153210163116, "leak": 1.265708970488049, "replacement": 2.4521133303642273}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 7.194621562957764, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 23.0, "pair": 0.7360508069396019, "sign": 0.44628299959003925, "keep": 1.3804957140237093, "leak": 0.26741164783015847, "replacement": 1.7162930369377136}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 11.6781005859375, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 23.0, "pair": 0.641843918710947, "sign": 0.7027354519814253, "keep": 1.7554625645279884, "leak": 0.3924789685988799, "replacement": 1.9690627753734589}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 8.832134246826172, "loss_scale": 4.0, "retries": 0, "seconds": 80.1, "max_memory_gb": 23.0, "pair": 0.5371589511632919, "sign": 0.4408652186393738, "keep": 1.3731255466118455, "leak": 1.0217658132314682, "replacement": 1.699595794081688}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 15.136100769042969, "loss_scale": 4.0, "retries": 0, "seconds": 81.2, "max_memory_gb": 23.0, "pair": 0.36094391718506813, "sign": 0.11051464825868607, "keep": 1.6612492315471172, "leak": 0.7861612390261143, "replacement": 1.443366140127182}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 17.625207901000977, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 23.0, "pair": 0.2592059653252363, "sign": 0.18314606323838234, "keep": 1.6291894689202309, "leak": 1.2260165959596634, "replacement": 1.4718778431415558}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 2.0018224716186523, "loss_scale": 4.0, "retries": 0, "seconds": 83.5, "max_memory_gb": 23.0, "pair": 0.8006908297538757, "sign": 0.7107764333486557, "keep": 0.3629901425447315, "leak": 0.09073562716366723, "replacement": 1.3602580577135086}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 6.327085494995117, "loss_scale": 4.0, "retries": 0, "seconds": 84.6, "max_memory_gb": 23.0, "pair": 0.6278930380940437, "sign": 0.45541360694915056, "keep": 0.6393865603022277, "leak": 0.4330939212231897, "replacement": 1.2835665792226791}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 4.457554817199707, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 23.0, "pair": 0.7381109111011028, "sign": 0.8507675589062274, "keep": 1.5992349870502949, "leak": 0.8060512243537232, "replacement": 2.1646249890327454}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 5.737955570220947, "loss_scale": 4.0, "retries": 0, "seconds": 86.9, "max_memory_gb": 23.0, "pair": 0.43200791999697685, "sign": 0.3540608137845993, "keep": 0.46603906713426113, "leak": 1.1953289173543453, "replacement": 1.140890121459961}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 6.939674377441406, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 23.0, "pair": 0.42989322915673256, "sign": 0.40602551586925983, "keep": 1.0506245791912079, "leak": 2.5218658531084657, "replacement": 1.7886847257614136}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 4.109147071838379, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 23.0, "pair": 0.6322058066725731, "sign": 0.702902015298605, "keep": 0.4161738622933626, "leak": 0.9782185219228268, "replacement": 1.4362983703613281}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.7130595445632935, "loss_scale": 4.0, "retries": 0, "seconds": 90.2, "max_memory_gb": 23.0, "pair": 0.6501416265964508, "sign": 0.49838123098015785, "keep": 0.14740054169669747, "leak": 0.4216448049992323, "replacement": 1.0784437209367752}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 3.772092580795288, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 23.0, "pair": 0.5302796792238951, "sign": 0.3423946648836136, "keep": 0.18712210259400308, "leak": 1.522246716544032, "replacement": 1.1755997389554977}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 2.5819783210754395, "loss_scale": 4.0, "retries": 0, "seconds": 92.5, "max_memory_gb": 23.0, "pair": 0.3499018605798483, "sign": 0.3457353888079524, "keep": 0.26826485246419907, "leak": 1.9734706906601787, "replacement": 1.1502696424722672}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 3.1631999015808105, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 23.0, "pair": 0.31753494311124086, "sign": 0.3037772518582642, "keep": 0.2417164221405983, "leak": 1.8197708736624918, "replacement": 1.0452244654297829}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 2.496525287628174, "loss_scale": 4.0, "retries": 0, "seconds": 94.8, "max_memory_gb": 23.0, "pair": 0.3842424005270004, "sign": 0.24719313718378544, "keep": 0.188898635096848, "leak": 1.6578942239284515, "replacement": 1.0167618095874786}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 2.294902801513672, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 23.0, "pair": 0.6666110977530479, "sign": 0.5817927569150925, "keep": 0.271462420001626, "leak": 1.1811205472331494, "replacement": 1.3885188102722168}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.979337453842163, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 23.0, "pair": 0.8173751682043076, "sign": 0.5273902863264084, "keep": 0.07910771970637143, "leak": 0.11920069344341755, "replacement": 1.1504243165254593}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 2.825814962387085, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 23.0, "pair": 0.4097523204982281, "sign": 0.549413550645113, "keep": 0.1919347457587719, "leak": 0.8098230499890633, "replacement": 0.9828822016716003}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 3.142972230911255, "loss_scale": 4.0, "retries": 0, "seconds": 99.4, "max_memory_gb": 23.0, "pair": 0.3313651829957962, "sign": 0.25066259503364563, "keep": 0.08466752897948027, "leak": 3.000754593133024, "replacement": 1.2492188811302185}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 1.9137898683547974, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 23.0, "pair": 0.6063133589923382, "sign": 0.5734921619296074, "keep": 0.3259077207185328, "leak": 0.6678292500291718, "replacement": 1.222970649600029}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 2.3084378242492676, "loss_scale": 4.0, "retries": 0, "seconds": 101.7, "max_memory_gb": 23.0, "pair": 0.8722847700119019, "sign": 0.6794251576066017, "keep": 0.21774376858957112, "leak": 0.08654700755141675, "replacement": 1.3425060212612152}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.2325384616851807, "loss_scale": 4.0, "retries": 0, "seconds": 102.8, "max_memory_gb": 23.0, "pair": 0.45904706977307796, "sign": 0.4166562557220459, "keep": 0.2656323058763519, "leak": 0.8216844536364079, "replacement": 1.0056124925613403}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 2.3183999061584473, "loss_scale": 4.0, "retries": 0, "seconds": 104.0, "max_memory_gb": 23.0, "pair": 0.7987313866615295, "sign": 0.886356383562088, "keep": 0.5448111956939101, "leak": 0.14945800270652398, "replacement": 1.5516797304153442}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.9888471364974976, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 23.0, "pair": 0.6549823507666588, "sign": 0.4372578635811806, "keep": 0.14956387097481638, "leak": 1.7260711677372456, "replacement": 1.3799110054969788}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 1.5555006265640259, "loss_scale": 4.0, "retries": 0, "seconds": 106.2, "max_memory_gb": 23.0, "pair": 0.4118071533739567, "sign": 0.22090348321944475, "keep": 0.08268375718034804, "leak": 0.44596800580620766, "replacement": 0.6750927716493607}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 3.426060199737549, "loss_scale": 4.0, "retries": 0, "seconds": 107.4, "max_memory_gb": 23.0, "pair": 0.3014066256582737, "sign": 0.31170262303203344, "keep": 0.2203919105231762, "leak": 1.7043894015951082, "replacement": 0.9935512393712997}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 2.0691254138946533, "loss_scale": 4.0, "retries": 0, "seconds": 108.5, "max_memory_gb": 23.0, "pair": 0.41624450497329235, "sign": 0.39271922409534454, "keep": 0.40589028084650636, "leak": 0.5607081837952137, "replacement": 0.9557262510061264}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 2.7868363857269287, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 23.0, "pair": 0.39645521715283394, "sign": 0.3707710145972669, "keep": 0.2194709349423647, "leak": 3.1862665958469734, "replacement": 1.4881428927183151}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.590911626815796, "loss_scale": 4.0, "retries": 0, "seconds": 110.8, "max_memory_gb": 23.0, "pair": 0.4083494432270527, "sign": 0.3470996804535389, "keep": 0.29983846843242645, "leak": 1.2256217896938324, "replacement": 1.0382239669561386}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 2.9768688678741455, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 23.0, "pair": 0.4922635406255722, "sign": 0.452635258436203, "keep": 0.049926349776797, "leak": 1.2263434901833534, "replacement": 1.0501301884651184}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 2.525010585784912, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 23.0, "pair": 0.722545437514782, "sign": 0.8116815835237503, "keep": 0.252126733539626, "leak": 0.19501370377838612, "replacement": 1.3032030165195465}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 1.6461273431777954, "loss_scale": 4.0, "retries": 0, "seconds": 114.2, "max_memory_gb": 23.0, "pair": 0.5506421588361263, "sign": 0.5871308641508222, "keep": 0.20840616710484028, "leak": 1.7811089163638485, "replacement": 1.3936879336833954}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 1.8342664241790771, "loss_scale": 4.0, "retries": 0, "seconds": 115.4, "max_memory_gb": 23.0, "pair": 0.29542306438088417, "sign": 0.10814460809342563, "keep": 0.13582156226038933, "leak": 0.7345763593912125, "replacement": 0.6010502427816391}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 3.302241325378418, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 23.0, "pair": 0.3504321537911892, "sign": 0.2732486268505454, "keep": 0.5266593843698502, "leak": 1.7715701828710735, "replacement": 1.193278729915619}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 2.270451307296753, "loss_scale": 4.0, "retries": 0, "seconds": 117.7, "max_memory_gb": 23.0, "pair": 0.39779260754585266, "sign": 0.25361840799450874, "keep": 0.06258316920138896, "leak": 1.6232524886727333, "replacement": 0.9617065042257309}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 4.167227745056152, "loss_scale": 4.0, "retries": 0, "seconds": 118.8, "max_memory_gb": 23.0, "pair": 0.5798591896891594, "sign": 0.7828239500522614, "keep": 0.048995057644788176, "leak": 0.4246210381388664, "replacement": 1.1019239872694016}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 2.56734037399292, "loss_scale": 4.0, "retries": 0, "seconds": 119.9, "max_memory_gb": 23.0, "pair": 0.42894659750163555, "sign": 0.28308194503188133, "keep": 0.1937780026346445, "leak": 1.8813036680221558, "replacement": 1.1377024799585342}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 2.8269588947296143, "loss_scale": 4.0, "retries": 0, "seconds": 121.1, "max_memory_gb": 23.0, "pair": 0.3738119937479496, "sign": 0.3702046824619174, "keep": 0.19454450905323029, "leak": 2.32061941572465, "replacement": 1.2363414019346237}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.78939151763916, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 23.0, "pair": 0.5556404665112495, "sign": 0.5436156541109085, "keep": 0.14184310985729098, "leak": 0.2121117552742362, "replacement": 0.9513977766036987}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 3.292299509048462, "loss_scale": 4.0, "retries": 0, "seconds": 123.4, "max_memory_gb": 23.0, "pair": 0.6395691931247711, "sign": 0.663409287109971, "keep": 0.13790662202518433, "leak": 2.212921499973163, "replacement": 1.5934575498104095}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 2.889655113220215, "loss_scale": 4.0, "retries": 0, "seconds": 124.5, "max_memory_gb": 23.0, "pair": 0.378737136721611, "sign": 0.40588719584047794, "keep": 0.3872980121523142, "leak": 2.01774545526132, "replacement": 1.2797661125659943}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 3.912116527557373, "loss_scale": 4.0, "retries": 0, "seconds": 125.7, "max_memory_gb": 23.0, "pair": 0.3712043594568968, "sign": 0.3948164046742022, "keep": 0.21737244445830584, "leak": 1.9623581934720278, "replacement": 1.1678882986307144}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 2.5355756282806396, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 23.0, "pair": 0.39128438755869865, "sign": 0.4076456050388515, "keep": 0.06479910854250193, "leak": 0.36429351987317204, "replacement": 0.7185801342129707}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 3.399611711502075, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 23.0, "pair": 0.36274369060993195, "sign": 0.30138746462762356, "keep": 0.14901404874399304, "leak": 0.7418088763952255, "replacement": 0.7733966559171677}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 4.200378894805908, "loss_scale": 4.0, "retries": 0, "seconds": 129.1, "max_memory_gb": 23.0, "pair": 0.4588193818926811, "sign": 0.45535503700375557, "keep": 0.2647802480496466, "leak": 0.5858125053346157, "replacement": 0.9653401672840118}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 4.012279033660889, "loss_scale": 4.0, "retries": 0, "seconds": 130.3, "max_memory_gb": 23.0, "pair": 0.3683602660894394, "sign": 0.4390281196683645, "keep": 0.5626496113836765, "leak": 1.536660198122263, "replacement": 1.2533642277121544}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 3.5851645469665527, "loss_scale": 4.0, "retries": 0, "seconds": 131.4, "max_memory_gb": 23.0, "pair": 0.41687189787626266, "sign": 0.46910784766077995, "keep": 0.38301308918744326, "leak": 1.2719524009153247, "replacement": 1.1609204411506653}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 2.155299425125122, "loss_scale": 4.0, "retries": 0, "seconds": 132.6, "max_memory_gb": 23.0, "pair": 0.5360839068889618, "sign": 0.5427481196820736, "keep": 0.17953753843903542, "leak": 0.48495032312348485, "replacement": 1.018464282155037}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 4.831405162811279, "loss_scale": 4.0, "retries": 0, "seconds": 133.7, "max_memory_gb": 23.0, "pair": 0.5756589695811272, "sign": 0.39156661555171013, "keep": 0.49639163352549076, "leak": 0.6734504103660583, "replacement": 1.1880006939172745}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 1.7613312005996704, "loss_scale": 4.0, "retries": 0, "seconds": 134.9, "max_memory_gb": 23.0, "pair": 0.5752444192767143, "sign": 0.4617976304143667, "keep": 0.2623897389275953, "leak": 0.4562673328910023, "replacement": 1.0514049530029297}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.481865644454956, "loss_scale": 4.0, "retries": 0, "seconds": 136.0, "max_memory_gb": 23.0, "pair": 0.6241571605205536, "sign": 0.6868472695350647, "keep": 0.09167451108805835, "leak": 0.3401981247588992, "replacement": 1.098467618227005}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 6.2948384284973145, "loss_scale": 4.0, "retries": 0, "seconds": 137.2, "max_memory_gb": 23.0, "pair": 0.4951492138206959, "sign": 0.4992402824573219, "keep": 0.6622838436160237, "leak": 0.8827376605477184, "replacement": 1.2965956255793571}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 3.9115583896636963, "loss_scale": 4.0, "retries": 0, "seconds": 138.4, "max_memory_gb": 23.0, "pair": 0.5558429509401321, "sign": 0.5976805686950684, "keep": 0.5227643286343664, "leak": 2.1059749238193035, "replacement": 1.6425590813159943}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 3.987809658050537, "loss_scale": 4.0, "retries": 0, "seconds": 139.5, "max_memory_gb": 23.0, "pair": 0.644853763282299, "sign": 0.6438653133809566, "keep": 0.16853664524387568, "leak": 0.6035978097934276, "replacement": 1.2019542306661606}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.341758966445923, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 23.0, "pair": 0.625364538282156, "sign": 0.7256865128874779, "keep": 0.9442212544381618, "leak": 0.562992843799293, "replacement": 1.6010666191577911}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 4.287633419036865, "loss_scale": 4.0, "retries": 0, "seconds": 141.8, "max_memory_gb": 23.0, "pair": 0.7512713447213173, "sign": 1.0058216154575348, "keep": 0.6531404200941324, "leak": 0.26550679840147495, "replacement": 1.647129088640213}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.6827757358551025, "loss_scale": 4.0, "retries": 0, "seconds": 142.9, "max_memory_gb": 23.0, "pair": 0.5184833407402039, "sign": 0.6258656233549118, "keep": 0.1466143480502069, "leak": 0.2850640071555972, "replacement": 0.9759893417358398}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 2.4150314331054688, "loss_scale": 4.0, "retries": 0, "seconds": 144.0, "max_memory_gb": 23.0, "pair": 0.645573191344738, "sign": 0.5875515015795827, "keep": 0.46338244900107384, "leak": 0.39236602996243164, "replacement": 1.2691317200660706}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 1.3876265287399292, "loss_scale": 4.0, "retries": 0, "seconds": 145.2, "max_memory_gb": 23.0, "pair": 0.34246115386486053, "sign": 0.23465658444911242, "keep": 0.03250073140952736, "leak": 0.5442615784704685, "replacement": 0.612105205655098}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 2.50209903717041, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 23.0, "pair": 0.23729033023118973, "sign": 0.16444512596353889, "keep": 0.17425782087957487, "leak": 1.3874698877334595, "replacement": 0.7535092830657959}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 2.4542460441589355, "loss_scale": 4.0, "retries": 0, "seconds": 147.5, "max_memory_gb": 23.0, "pair": 0.308734817430377, "sign": 0.19269755855202675, "keep": 0.0919836200773716, "leak": 1.580550393089652, "replacement": 0.8462130054831505}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 7.176722526550293, "loss_scale": 4.0, "retries": 0, "seconds": 148.6, "max_memory_gb": 23.0, "pair": 0.6055649351328611, "sign": 0.9737078528851271, "keep": 0.6389829718973488, "leak": 1.482891283929348, "replacement": 1.7826331853866577}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 5.406215190887451, "loss_scale": 4.0, "retries": 0, "seconds": 149.7, "max_memory_gb": 23.0, "pair": 0.4030433800071478, "sign": 0.5589439626783133, "keep": 0.45526510290801525, "leak": 0.7165423259139061, "replacement": 1.0892834663391113}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 3.0974202156066895, "loss_scale": 4.0, "retries": 0, "seconds": 150.9, "max_memory_gb": 23.0, "pair": 0.3569257389754057, "sign": 0.10459102317690849, "keep": 0.7571634368505329, "leak": 1.1188770160079002, "replacement": 1.0675222128629684}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 4.516090393066406, "loss_scale": 4.0, "retries": 0, "seconds": 152.0, "max_memory_gb": 23.0, "pair": 0.41684337705373764, "sign": 0.38965793512761593, "keep": 0.33869290724396706, "leak": 2.475586209446192, "replacement": 1.3999153450131416}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 3.3852429389953613, "loss_scale": 4.0, "retries": 0, "seconds": 153.2, "max_memory_gb": 23.0, "pair": 0.5262540280818939, "sign": 0.5455060005187988, "keep": 0.17740810371469706, "leak": 0.4245542846620083, "replacement": 0.9938496798276901}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.9982521533966064, "loss_scale": 4.0, "retries": 0, "seconds": 154.3, "max_memory_gb": 23.0, "pair": 0.36438993737101555, "sign": 0.3776875454932451, "keep": 0.1740665677934885, "leak": 2.705177373095921, "replacement": 1.316561333835125}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 5.990110874176025, "loss_scale": 4.0, "retries": 0, "seconds": 155.5, "max_memory_gb": 23.0, "pair": 0.5136913768947124, "sign": 0.44898158125579357, "keep": 0.5272207199595869, "leak": 0.5297061149030924, "replacement": 1.1342190951108932}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 5.549549102783203, "loss_scale": 4.0, "retries": 0, "seconds": 156.7, "max_memory_gb": 23.0, "pair": 0.4966253936290741, "sign": 0.4127358254045248, "keep": 0.2737730769440532, "leak": 1.9745384119451046, "replacement": 1.3335145115852356}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 2.509514093399048, "loss_scale": 4.0, "retries": 0, "seconds": 157.8, "max_memory_gb": 23.0, "pair": 0.2550509460270405, "sign": 0.16700103785842657, "keep": 0.1389150203904137, "leak": 0.5771010969765484, "replacement": 0.5522842407226562}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 8.518630027770996, "loss_scale": 4.0, "retries": 0, "seconds": 159.0, "max_memory_gb": 23.0, "pair": 0.4105739016085863, "sign": 0.5272934157401323, "keep": 0.42144165001809597, "leak": 1.4816514700651169, "replacement": 1.2553543001413345}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 2.7231717109680176, "loss_scale": 4.0, "retries": 0, "seconds": 160.2, "max_memory_gb": 23.0, "pair": 0.3420987520366907, "sign": 0.38139364402741194, "keep": 0.08120470447465777, "leak": 1.0564759373664856, "replacement": 0.8375168889760971}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 4.005566596984863, "loss_scale": 4.0, "retries": 0, "seconds": 161.3, "max_memory_gb": 23.0, "pair": 0.3818762004375458, "sign": 0.45211077481508255, "keep": 0.41129848564742133, "leak": 1.2642158269882202, "replacement": 1.129634827375412}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.0571272373199463, "loss_scale": 4.0, "retries": 0, "seconds": 162.5, "max_memory_gb": 23.0, "pair": 0.3315920829772949, "sign": 0.3201957307755947, "keep": 0.17259328183718026, "leak": 0.664903998374939, "replacement": 0.7442125901579857}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 3.1402089595794678, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 23.0, "pair": 0.6764119267463684, "sign": 0.9405867867171764, "keep": 0.19693536800332367, "leak": 0.5642305891960859, "replacement": 1.3862306773662567}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 4.123229503631592, "loss_scale": 4.0, "retries": 0, "seconds": 164.8, "max_memory_gb": 23.0, "pair": 0.2850338201969862, "sign": 0.2734814118593931, "keep": 0.19204913428984582, "leak": 2.1073296703398228, "replacement": 1.0446315556764603}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 2.655587911605835, "loss_scale": 4.0, "retries": 0, "seconds": 165.9, "max_memory_gb": 23.0, "pair": 0.43212952092289925, "sign": 0.23335048276931047, "keep": 0.03634372213855386, "leak": 1.4697719011455774, "replacement": 0.934419572353363}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.35349440574646, "loss_scale": 4.0, "retries": 0, "seconds": 167.1, "max_memory_gb": 23.0, "pair": 0.41845909878611565, "sign": 0.4882390624843538, "keep": 0.12896962277591228, "leak": 1.3186675868928432, "replacement": 1.0567303448915482}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 4.265872001647949, "loss_scale": 4.0, "retries": 0, "seconds": 168.2, "max_memory_gb": 23.0, "pair": 0.4313304089009762, "sign": 0.4915440855547786, "keep": 0.5558687768061645, "leak": 1.5122458636760712, "replacement": 1.333098292350769}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.7518136501312256, "loss_scale": 4.0, "retries": 0, "seconds": 169.4, "max_memory_gb": 23.0, "pair": 0.31437333673238754, "sign": 0.18849874660372734, "keep": 0.022817636898253113, "leak": 0.7555905583394633, "replacement": 0.6089291721582413}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 2.7599880695343018, "loss_scale": 4.0, "retries": 0, "seconds": 170.5, "max_memory_gb": 23.0, "pair": 0.352647814899683, "sign": 0.3442099019885063, "keep": 0.06480934785213321, "leak": 0.6066542444750667, "replacement": 0.7088209912180901}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 3.0898842811584473, "loss_scale": 4.0, "retries": 0, "seconds": 171.7, "max_memory_gb": 23.0, "pair": 0.39262161403894424, "sign": 0.3064062977209687, "keep": 0.5372011512517929, "leak": 2.0385762453079224, "replacement": 1.3240693807601929}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 3.6963181495666504, "loss_scale": 4.0, "retries": 0, "seconds": 172.8, "max_memory_gb": 23.0, "pair": 0.27330779284238815, "sign": 0.2730714902281761, "keep": 0.41850748425349593, "leak": 0.5351866707205772, "replacement": 0.7528939247131348}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 3.9224703311920166, "loss_scale": 4.0, "retries": 0, "seconds": 174.0, "max_memory_gb": 23.0, "pair": 0.15679893642663956, "sign": 0.12145413178950548, "keep": 0.09605502104386687, "leak": 3.073618173599243, "replacement": 1.0339580476284027}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 3.0164034366607666, "loss_scale": 4.0, "retries": 0, "seconds": 175.2, "max_memory_gb": 23.0, "pair": 0.2004811055958271, "sign": 0.15360246039927006, "keep": 0.04549246106762439, "leak": 1.8066792488098145, "replacement": 0.75169837474823}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 3.9731249809265137, "loss_scale": 4.0, "retries": 0, "seconds": 176.3, "max_memory_gb": 23.0, "pair": 0.2609138870611787, "sign": 0.1607324331998825, "keep": 0.8163792756386101, "leak": 1.0981284864246845, "replacement": 1.024001844227314}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.638017177581787, "loss_scale": 4.0, "retries": 0, "seconds": 177.5, "max_memory_gb": 23.0, "pair": 0.3544723875820637, "sign": 0.4500231258571148, "keep": 1.1585358765441924, "leak": 1.939085803925991, "replacement": 1.643523395061493}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 2.016120433807373, "loss_scale": 4.0, "retries": 0, "seconds": 178.7, "max_memory_gb": 23.0, "pair": 0.47515151649713516, "sign": 0.4378039939329028, "keep": 0.17318345559760928, "leak": 0.35064361058175564, "replacement": 0.868306115269661}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 4.105961322784424, "loss_scale": 4.0, "retries": 0, "seconds": 179.8, "max_memory_gb": 23.0, "pair": 0.5327454172074795, "sign": 1.0229726433753967, "keep": 0.14716262929141521, "leak": 0.6135096028447151, "replacement": 1.271190419793129}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 3.5139968395233154, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 23.0, "pair": 0.44774707220494747, "sign": 0.5574869383126497, "keep": 0.33066984824836254, "leak": 0.7251648377277888, "replacement": 1.0731167197227478}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 3.868262529373169, "loss_scale": 4.0, "retries": 0, "seconds": 182.1, "max_memory_gb": 23.0, "pair": 0.16719089355319738, "sign": 0.12083287350833416, "keep": 0.040250138961710036, "leak": 2.3955962508916855, "replacement": 0.8466314673423767}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 3.849156141281128, "loss_scale": 4.0, "retries": 0, "seconds": 183.2, "max_memory_gb": 23.0, "pair": 0.4076135568320751, "sign": 0.28917089593596756, "keep": 0.17236889898777008, "leak": 1.5882053158493363, "replacement": 1.035434752702713}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 3.901946544647217, "loss_scale": 4.0, "retries": 0, "seconds": 184.4, "max_memory_gb": 23.0, "pair": 0.44878282584249973, "sign": 0.5085149202495813, "keep": 0.38895941409282386, "leak": 1.3063750267028809, "replacement": 1.2241138070821762}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 3.1724560260772705, "loss_scale": 4.0, "retries": 0, "seconds": 185.5, "max_memory_gb": 23.0, "pair": 0.3527827961370349, "sign": 0.4224162707105279, "keep": 0.05985148041509092, "leak": 0.398527173572802, "replacement": 0.6935484632849693}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 3.6218814849853516, "loss_scale": 4.0, "retries": 0, "seconds": 186.6, "max_memory_gb": 23.0, "pair": 0.4476969391107559, "sign": 0.2969706179574132, "keep": 0.0872966330498457, "leak": 0.6108630876988173, "replacement": 0.7925463691353798}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 4.451742649078369, "loss_scale": 4.0, "retries": 0, "seconds": 187.7, "max_memory_gb": 23.0, "pair": 0.5487729944288731, "sign": 0.7451738454401493, "keep": 0.503703884780407, "leak": 0.7049787491559982, "replacement": 1.3494565188884735}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 4.207320213317871, "loss_scale": 4.0, "retries": 0, "seconds": 188.9, "max_memory_gb": 23.0, "pair": 0.30305279418826103, "sign": 0.4161623790860176, "keep": 0.10040356637910008, "leak": 1.7865128591656685, "replacement": 1.007963977754116}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 3.295868158340454, "loss_scale": 4.0, "retries": 0, "seconds": 190.1, "max_memory_gb": 23.0, "pair": 0.35485414043068886, "sign": 0.41483957692980766, "keep": 0.20660315244458616, "leak": 1.8587233481559906, "replacement": 1.1302562952041626}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 3.463388204574585, "loss_scale": 4.0, "retries": 0, "seconds": 191.2, "max_memory_gb": 23.0, "pair": 0.3558414690196514, "sign": 0.38911802135407925, "keep": 0.5058933775871992, "leak": 1.4224644685164094, "replacement": 1.1589633338153362}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.7690062522888184, "loss_scale": 4.0, "retries": 0, "seconds": 192.4, "max_memory_gb": 23.0, "pair": 0.3074823059141636, "sign": 0.255804093554616, "keep": 0.08049077726900578, "leak": 1.3472223430871964, "replacement": 0.8124353513121605}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 5.119472503662109, "loss_scale": 4.0, "retries": 0, "seconds": 193.5, "max_memory_gb": 23.0, "pair": 0.34130996465682983, "sign": 0.42647489067167044, "keep": 0.8383806180208921, "leak": 2.0080810096114874, "replacement": 1.4757579565048218}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 1.424269199371338, "loss_scale": 4.0, "retries": 0, "seconds": 194.7, "max_memory_gb": 23.0, "pair": 0.2413141168653965, "sign": 0.07032856065779924, "keep": 0.04402905027382076, "leak": 1.1976599618792534, "replacement": 0.5979079082608223}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 3.443725824356079, "loss_scale": 4.0, "retries": 0, "seconds": 195.8, "max_memory_gb": 23.0, "pair": 0.3556291237473488, "sign": 0.32687158323824406, "keep": 0.05923513788729906, "leak": 1.1437460370361805, "replacement": 0.8346189856529236}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 3.513850450515747, "loss_scale": 4.0, "retries": 0, "seconds": 197.0, "max_memory_gb": 23.0, "pair": 0.2637270390987396, "sign": 0.07057484425604343, "keep": 0.1477897185832262, "leak": 4.1935112327337265, "replacement": 1.421287178993225}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 4.434851169586182, "loss_scale": 4.0, "retries": 0, "seconds": 198.1, "max_memory_gb": 23.0, "pair": 0.21794753707945347, "sign": 0.3333002105355263, "keep": 0.4011398684233427, "leak": 2.120388090610504, "replacement": 1.1152646094560623}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 2.346876859664917, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 23.0, "pair": 0.3138941079378128, "sign": 0.23303266614675522, "keep": 0.14279288187390193, "leak": 1.1414887346327305, "replacement": 0.7871790453791618}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.8861494064331055, "loss_scale": 4.0, "retries": 0, "seconds": 200.4, "max_memory_gb": 23.0, "pair": 0.36310888081789017, "sign": 0.4463065527379513, "keep": 0.1371347662061453, "leak": 1.8860648376867175, "replacement": 1.1263457238674164}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 2.1628830432891846, "loss_scale": 4.0, "retries": 0, "seconds": 201.5, "max_memory_gb": 23.0, "pair": 0.3125925697386265, "sign": 0.1287028891965747, "keep": 0.10077696852385998, "leak": 2.3385830000042915, "replacement": 1.0119782537221909}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 8.22244930267334, "loss_scale": 4.0, "retries": 0, "seconds": 202.7, "max_memory_gb": 23.0, "pair": 0.36996348574757576, "sign": 0.35140558145940304, "keep": 0.5060418242355809, "leak": 0.6996581557905301, "replacement": 0.9736017063260078}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 3.681478500366211, "loss_scale": 4.0, "retries": 0, "seconds": 203.8, "max_memory_gb": 23.0, "pair": 0.3958763815462589, "sign": 0.4354059137403965, "keep": 0.1403897904092446, "leak": 1.5823800046928227, "replacement": 1.0793692469596863}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 3.461169958114624, "loss_scale": 4.0, "retries": 0, "seconds": 204.9, "max_memory_gb": 23.0, "pair": 0.632381814531982, "sign": 0.7307833023369312, "keep": 0.6289293868467212, "leak": 0.9816383724100888, "replacement": 1.5576477944850922}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 7.541053771972656, "loss_scale": 4.0, "retries": 0, "seconds": 206.0, "max_memory_gb": 23.0, "pair": 0.809996098279953, "sign": 0.6797370370477438, "keep": 1.0489769466221333, "leak": 0.5668945062643616, "replacement": 1.8160766512155533}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 4.6868977546691895, "loss_scale": 4.0, "retries": 0, "seconds": 207.2, "max_memory_gb": 23.0, "pair": 0.4672900065779686, "sign": 0.16896319761872292, "keep": 0.1220891885459423, "leak": 1.790761947631836, "replacement": 1.0605066865682602}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 6.3361945152282715, "loss_scale": 4.0, "retries": 0, "seconds": 208.3, "max_memory_gb": 23.0, "pair": 0.5209188535809517, "sign": 0.5738088488578796, "keep": 0.6456581997917965, "leak": 0.8081889897584915, "replacement": 1.3326996266841888}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.714735269546509, "loss_scale": 4.0, "retries": 0, "seconds": 209.5, "max_memory_gb": 23.0, "pair": 0.2288278192281723, "sign": 0.12001124257221818, "keep": 0.02718790655490011, "leak": 1.1945380680263042, "replacement": 0.6010619178414345}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 2.9513890743255615, "loss_scale": 4.0, "retries": 0, "seconds": 210.6, "max_memory_gb": 23.0, "pair": 0.30900051444768906, "sign": 0.22441670205444098, "keep": 0.07474269345402718, "leak": 0.6602882165461779, "replacement": 0.623652271926403}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 3.1763393878936768, "loss_scale": 4.0, "retries": 0, "seconds": 211.8, "max_memory_gb": 23.0, "pair": 0.33320070058107376, "sign": 0.22906288132071495, "keep": 0.08579470217227936, "leak": 2.197414015448885, "replacement": 1.0399829894304276}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 4.145799160003662, "loss_scale": 4.0, "retries": 0, "seconds": 213.0, "max_memory_gb": 23.0, "pair": 0.2910414468497038, "sign": 0.2188089732080698, "keep": 0.12470578867942095, "leak": 0.5869748361874372, "replacement": 0.60954250395298}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 4.949768543243408, "loss_scale": 4.0, "retries": 0, "seconds": 214.2, "max_memory_gb": 23.0, "pair": 0.4247140437364578, "sign": 0.08674783632159233, "keep": 0.20169359538704157, "leak": 1.0310466885566711, "replacement": 0.8266964331269264}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 3.1676883697509766, "loss_scale": 4.0, "retries": 0, "seconds": 215.3, "max_memory_gb": 23.0, "pair": 0.39948124811053276, "sign": 0.4151771478354931, "keep": 0.21753070526756346, "leak": 0.3273604717105627, "replacement": 0.7976752892136574}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.8084099292755127, "loss_scale": 4.0, "retries": 0, "seconds": 216.4, "max_memory_gb": 23.0, "pair": 0.23164506256580353, "sign": 0.1802601609379053, "keep": 0.06420033262111247, "leak": 1.0144057036523009, "replacement": 0.6074767261743546}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 2.124633312225342, "loss_scale": 4.0, "retries": 0, "seconds": 217.6, "max_memory_gb": 23.0, "pair": 0.23746903985738754, "sign": 0.13924130983650684, "keep": 0.06923296861350536, "leak": 1.1991208270192146, "replacement": 0.6414863839745522}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 8.588860511779785, "loss_scale": 4.0, "retries": 0, "seconds": 218.7, "max_memory_gb": 23.0, "pair": 0.5682097338140011, "sign": 0.6812968775629997, "keep": 1.047620799858123, "leak": 1.3291018903255463, "replacement": 1.7649440467357635}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 1.9309478998184204, "loss_scale": 4.0, "retries": 0, "seconds": 219.8, "max_memory_gb": 23.0, "pair": 0.5552733764052391, "sign": 0.4459144864231348, "keep": 0.33924077078700066, "leak": 0.5952860991237685, "replacement": 1.0966725051403046}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 2.7995848655700684, "loss_scale": 4.0, "retries": 0, "seconds": 220.9, "max_memory_gb": 23.0, "pair": 0.36914782226085663, "sign": 0.1449082624167204, "keep": 0.12450382101815194, "leak": 0.5012797117233276, "replacement": 0.6291738152503967}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 13.584620475769043, "loss_scale": 4.0, "retries": 0, "seconds": 222.1, "max_memory_gb": 23.0, "pair": 0.6023585610091686, "sign": 0.889836985617876, "keep": 1.515456773340702, "leak": 1.21840700507164, "replacement": 2.1096072494983673}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 3.986741781234741, "loss_scale": 4.0, "retries": 0, "seconds": 223.3, "max_memory_gb": 23.0, "pair": 0.4530256651341915, "sign": 0.33616284653544426, "keep": 0.3896463979035616, "leak": 1.499588469043374, "replacement": 1.1908274292945862}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 2.0502264499664307, "loss_scale": 4.0, "retries": 0, "seconds": 224.4, "max_memory_gb": 23.0, "pair": 0.23459943756461143, "sign": 0.08333849301561713, "keep": 0.07224820170085877, "leak": 0.8622368350625038, "replacement": 0.5279520004987717}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 3.335965633392334, "loss_scale": 4.0, "retries": 0, "seconds": 225.6, "max_memory_gb": 23.0, "pair": 0.3613244965672493, "sign": 0.37902766093611717, "keep": 0.09413903020322323, "leak": 0.4573502978309989, "replacement": 0.7122454047203064}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 3.6510424613952637, "loss_scale": 4.0, "retries": 0, "seconds": 226.7, "max_memory_gb": 23.0, "pair": 0.3972374275326729, "sign": 0.4954160675406456, "keep": 0.20702628884464502, "leak": 0.908917504362762, "replacement": 0.9756880402565002}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.058095693588257, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 23.0, "pair": 0.25541727617383003, "sign": 0.12597523257136345, "keep": 0.2621398540213704, "leak": 1.834454357624054, "replacement": 0.9080884009599686}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 2.278174877166748, "loss_scale": 4.0, "retries": 0, "seconds": 229.0, "max_memory_gb": 23.0, "pair": 0.37827832624316216, "sign": 0.503392830491066, "keep": 0.37630787363741547, "leak": 0.9499468980357051, "replacement": 1.055615410208702}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 5.542818069458008, "loss_scale": 4.0, "retries": 0, "seconds": 230.2, "max_memory_gb": 23.0, "pair": 0.42381394654512405, "sign": 0.5371589353308082, "keep": 0.1945495845284313, "leak": 0.7211587652564049, "replacement": 0.9699579253792763}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 3.8547446727752686, "loss_scale": 4.0, "retries": 0, "seconds": 231.3, "max_memory_gb": 23.0, "pair": 0.3484739437699318, "sign": 0.2766244150698185, "keep": 0.008403153624385595, "leak": 0.9505766928195953, "replacement": 0.7286318838596344}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 3.1945998668670654, "loss_scale": 4.0, "retries": 0, "seconds": 232.4, "max_memory_gb": 23.0, "pair": 0.2748805768787861, "sign": 0.34335703402757645, "keep": 0.17870438529644161, "leak": 1.5822416814044118, "replacement": 0.9314717054367065}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 4.2697529792785645, "loss_scale": 4.0, "retries": 0, "seconds": 233.6, "max_memory_gb": 23.0, "pair": 0.46591613441705704, "sign": 0.45783178228884935, "keep": 0.18160137813538313, "leak": 1.3660936006344855, "replacement": 1.127156138420105}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 3.482076644897461, "loss_scale": 4.0, "retries": 0, "seconds": 234.7, "max_memory_gb": 23.0, "pair": 0.16787821799516678, "sign": 0.24848760478198528, "keep": 0.07838471670402214, "leak": 1.6700532138347626, "replacement": 0.748827688395977}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 3.5732452869415283, "loss_scale": 4.0, "retries": 0, "seconds": 235.9, "max_memory_gb": 23.0, "pair": 0.3653775528073311, "sign": 0.31016468070447445, "keep": 0.15840654476778582, "leak": 1.909288004040718, "replacement": 1.0769851580262184}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 8.827729225158691, "loss_scale": 4.0, "retries": 0, "seconds": 237.0, "max_memory_gb": 23.0, "pair": 0.7037072964012623, "sign": 0.8645778279751539, "keep": 1.5985861322842538, "leak": 0.693953575566411, "replacement": 2.108777701854706}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 2.609454870223999, "loss_scale": 4.0, "retries": 0, "seconds": 238.2, "max_memory_gb": 23.0, "pair": 0.3385082669556141, "sign": 0.08296365197747946, "keep": 0.04580990003887564, "leak": 1.2894013077020645, "replacement": 0.7252453714609146}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 3.038038492202759, "loss_scale": 4.0, "retries": 0, "seconds": 239.4, "max_memory_gb": 23.0, "pair": 0.39730963855981827, "sign": 0.47343363519757986, "keep": 0.25691218115389347, "leak": 0.7232240261510015, "replacement": 0.9432885348796844}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 4.061520099639893, "loss_scale": 4.0, "retries": 0, "seconds": 240.5, "max_memory_gb": 23.0, "pair": 0.37338135950267315, "sign": 0.4685988016426563, "keep": 0.7553635570220649, "leak": 2.545781034976244, "replacement": 1.6218077391386032}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 9.14777946472168, "loss_scale": 4.0, "retries": 0, "seconds": 241.7, "max_memory_gb": 23.0, "pair": 0.471384696662426, "sign": 0.5832538437098265, "keep": 0.42924576066434383, "leak": 0.8936067257309332, "replacement": 1.2010361552238464}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 9.76296615600586, "loss_scale": 4.0, "retries": 0, "seconds": 242.9, "max_memory_gb": 23.0, "pair": 0.4486781768500805, "sign": 0.7028502821922302, "keep": 0.7072893790900707, "leak": 0.5841408135602251, "replacement": 1.2997832298278809}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 4.8328351974487305, "loss_scale": 4.0, "retries": 0, "seconds": 244.0, "max_memory_gb": 23.0, "pair": 0.35098670423030853, "sign": 0.4546591634862125, "keep": 1.1013230762910098, "leak": 2.0082168807275593, "replacement": 1.6310321018099785}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 3.7885189056396484, "loss_scale": 4.0, "retries": 0, "seconds": 245.2, "max_memory_gb": 23.0, "pair": 0.48874780908226967, "sign": 0.5424623899161816, "keep": 0.5697849956341088, "leak": 0.46263992085005157, "replacement": 1.1605314686894417}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 8.554731369018555, "loss_scale": 4.0, "retries": 0, "seconds": 246.3, "max_memory_gb": 23.0, "pair": 0.7146798223257065, "sign": 0.948862724006176, "keep": 2.109033510903828, "leak": 0.4735500146634877, "replacement": 2.3620153293013573}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 2.623183012008667, "loss_scale": 4.0, "retries": 0, "seconds": 247.5, "max_memory_gb": 23.0, "pair": 0.19829531572759151, "sign": 0.21110766381025314, "keep": 0.21609298349358141, "leak": 0.806284349411726, "replacement": 0.6134667322039604}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 5.250481605529785, "loss_scale": 4.0, "retries": 0, "seconds": 248.6, "max_memory_gb": 23.0, "pair": 0.15841340739279985, "sign": 0.12939542485401034, "keep": 0.04357483983039856, "leak": 1.489580325782299, "replacement": 0.6172936260700226}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 8.332698822021484, "loss_scale": 4.0, "retries": 0, "seconds": 249.7, "max_memory_gb": 23.0, "pair": 0.6669169217348099, "sign": 0.8849676847457886, "keep": 0.4722177452640608, "leak": 0.19042116323544178, "replacement": 1.393114984035492}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 4.7694830894470215, "loss_scale": 4.0, "retries": 0, "seconds": 250.9, "max_memory_gb": 23.0, "pair": 0.304960360750556, "sign": 0.40216777101159096, "keep": 0.18200738623272628, "leak": 0.44222917314618826, "replacement": 0.7076052129268646}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 3.869429588317871, "loss_scale": 4.0, "retries": 0, "seconds": 252.0, "max_memory_gb": 23.0, "pair": 0.27663104236125946, "sign": 0.3821262754499912, "keep": 0.3181002287892625, "leak": 0.7718571946024895, "replacement": 0.8197086006402969}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 2.245408296585083, "loss_scale": 4.0, "retries": 0, "seconds": 253.2, "max_memory_gb": 23.0, "pair": 0.3243087334558368, "sign": 0.25754818227142096, "keep": 0.19572814740240574, "leak": 0.7942692330107093, "replacement": 0.7495141997933388}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 5.42972993850708, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 23.0, "pair": 0.3867120072245598, "sign": 0.20559485908597708, "keep": 0.29694950953125954, "leak": 1.1332863038405776, "replacement": 0.9213057830929756}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 12.415023803710938, "loss_scale": 4.0, "retries": 0, "seconds": 255.5, "max_memory_gb": 23.0, "pair": 0.6024812757968903, "sign": 0.6877228394150734, "keep": 0.46591752057429403, "leak": 0.24128946010023355, "replacement": 1.2396238669753075}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 2.6227054595947266, "loss_scale": 4.0, "retries": 0, "seconds": 256.7, "max_memory_gb": 23.0, "pair": 0.26564906165003777, "sign": 0.3065656963735819, "keep": 0.09317979100160301, "leak": 0.9817693252116442, "replacement": 0.7109641209244728}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 3.3624818325042725, "loss_scale": 4.0, "retries": 0, "seconds": 257.8, "max_memory_gb": 23.0, "pair": 0.5098389443010092, "sign": 0.5587830152362585, "keep": 0.09851827844977379, "leak": 0.8975543715059757, "replacement": 1.0628781504929066}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 4.615634918212891, "loss_scale": 4.0, "retries": 0, "seconds": 259.0, "max_memory_gb": 23.0, "pair": 0.18254529498517513, "sign": 0.21711840387433767, "keep": 0.18213410722091794, "leak": 0.9059861153364182, "replacement": 0.6086680814623833}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 1.8836652040481567, "loss_scale": 4.0, "retries": 0, "seconds": 260.1, "max_memory_gb": 23.0, "pair": 0.2937803640961647, "sign": 0.14764471538364887, "keep": 0.031215929484460503, "leak": 0.8310219682753086, "replacement": 0.5909661799669266}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 4.078284740447998, "loss_scale": 4.0, "retries": 0, "seconds": 261.2, "max_memory_gb": 23.0, "pair": 0.15243088826537132, "sign": 0.25397539883852005, "keep": 0.01723295426927507, "leak": 1.3798871636390686, "replacement": 0.6330068483948708}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 5.4939188957214355, "loss_scale": 4.0, "retries": 0, "seconds": 262.4, "max_memory_gb": 23.0, "pair": 0.40720631554722786, "sign": 0.6202507428824902, "keep": 0.7548042472917587, "leak": 0.957349568605423, "replacement": 1.334071196615696}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 4.0434041023254395, "loss_scale": 4.0, "retries": 0, "seconds": 263.5, "max_memory_gb": 23.0, "pair": 0.6640072986483574, "sign": 0.8018092438578606, "keep": 0.678442656993866, "leak": 0.27108230628073215, "replacement": 1.4719038382172585}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 7.312046527862549, "loss_scale": 4.0, "retries": 0, "seconds": 264.5, "max_memory_gb": 23.0, "pair": 0.62039665132761, "sign": 0.8120960146188736, "keep": 0.7881292699603364, "leak": 0.4995900606736541, "replacement": 1.545406810939312}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 4.549925327301025, "loss_scale": 4.0, "retries": 0, "seconds": 265.7, "max_memory_gb": 23.0, "pair": 0.18134813755750656, "sign": 0.24921255931258202, "keep": 0.03815606143325567, "leak": 1.2302093878388405, "replacement": 0.6325848177075386}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 3.376627206802368, "loss_scale": 4.0, "retries": 0, "seconds": 266.8, "max_memory_gb": 23.0, "pair": 0.2188985738903284, "sign": 0.12266431748867035, "keep": 0.2791279593948275, "leak": 1.3733366951346397, "replacement": 0.7631288766860962}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.9241461753845215, "loss_scale": 4.0, "retries": 0, "seconds": 268.0, "max_memory_gb": 23.0, "pair": 0.2845726050436497, "sign": 0.36934475041925907, "keep": 0.23271888034651056, "leak": 1.9796773232519627, "replacement": 1.080523744225502}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 4.745080947875977, "loss_scale": 4.0, "retries": 0, "seconds": 269.1, "max_memory_gb": 23.0, "pair": 0.4847721327096224, "sign": 0.5890511237084866, "keep": 0.7590868545230478, "leak": 0.9015792212449014, "replacement": 1.3842359334230423}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 5.991840839385986, "loss_scale": 4.0, "retries": 0, "seconds": 270.2, "max_memory_gb": 23.0, "pair": 0.49942345172166824, "sign": 0.4535788334906101, "keep": 0.3257291375193745, "leak": 0.16537032055202872, "replacement": 0.9304200932383537}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.4052069187164307, "loss_scale": 4.0, "retries": 0, "seconds": 271.3, "max_memory_gb": 23.0, "pair": 0.3871796056628227, "sign": 0.3785907682031393, "keep": 0.31476894032675773, "leak": 0.5376626671495615, "replacement": 0.8682751283049583}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 4.004530906677246, "loss_scale": 4.0, "retries": 0, "seconds": 272.5, "max_memory_gb": 23.0, "pair": 0.24850755371153355, "sign": 0.13397609815001488, "keep": 0.09745165682397783, "leak": 1.396081853657961, "replacement": 0.7132419049739838}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 3.993925094604492, "loss_scale": 4.0, "retries": 0, "seconds": 273.6, "max_memory_gb": 23.0, "pair": 0.4191310815513134, "sign": 0.47474736347794533, "keep": 0.23958921525627375, "leak": 2.2985383719205856, "replacement": 1.3509339690208435}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.6645119190216064, "loss_scale": 4.0, "retries": 0, "seconds": 274.7, "max_memory_gb": 23.0, "pair": 0.29934823885560036, "sign": 0.2470872476696968, "keep": 0.10906311182770878, "leak": 0.38258588314056396, "replacement": 0.5730699002742767}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 3.7984678745269775, "loss_scale": 4.0, "retries": 0, "seconds": 275.9, "max_memory_gb": 23.0, "pair": 0.33674027770757675, "sign": 0.523071214556694, "keep": 0.1036176816560328, "leak": 0.5246541509986855, "replacement": 0.7812482565641403}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 2.755009412765503, "loss_scale": 4.0, "retries": 0, "seconds": 277.0, "max_memory_gb": 23.0, "pair": 0.42161949537694454, "sign": 0.5553559553809464, "keep": 0.2814236163976602, "leak": 0.9183282423764467, "replacement": 1.0695913806557655}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 2.607370376586914, "loss_scale": 4.0, "retries": 0, "seconds": 278.2, "max_memory_gb": 23.0, "pair": 0.3111737538129091, "sign": 0.25175614655017853, "keep": 0.0830026013427414, "leak": 2.389546856051311, "replacement": 1.0759398117661476}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 12.396395683288574, "loss_scale": 4.0, "retries": 0, "seconds": 279.3, "max_memory_gb": 23.0, "pair": 0.6404222995042801, "sign": 0.9728200668469071, "keep": 1.3024247800931334, "leak": 0.7342297513969243, "replacement": 1.9616022258996964}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 6.214446067810059, "loss_scale": 4.0, "retries": 0, "seconds": 280.5, "max_memory_gb": 23.0, "pair": 0.830041378736496, "sign": 1.002443179488182, "keep": 0.8687561387196183, "leak": 0.2774791643023491, "replacement": 1.8350107967853546}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 2.2628018856048584, "loss_scale": 4.0, "retries": 0, "seconds": 281.6, "max_memory_gb": 23.0, "pair": 0.20036179013550282, "sign": 0.16336531564593315, "keep": 0.07261167967226356, "leak": 0.9889114946126938, "replacement": 0.5655781477689743}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 3.001396417617798, "loss_scale": 4.0, "retries": 0, "seconds": 282.7, "max_memory_gb": 23.0, "pair": 0.2538244426250458, "sign": 0.2857191190123558, "keep": 0.1730133083765395, "leak": 0.47983543202281, "replacement": 0.6031495481729507}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.0666849613189697, "loss_scale": 4.0, "retries": 0, "seconds": 283.9, "max_memory_gb": 23.0, "pair": 0.2267312090843916, "sign": 0.12706120405346155, "keep": 0.13813850260339677, "leak": 0.6587942179758102, "replacement": 0.5240296348929405}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 5.227479457855225, "loss_scale": 4.0, "retries": 0, "seconds": 285.0, "max_memory_gb": 23.0, "pair": 0.5151240192353725, "sign": 0.7267732173204422, "keep": 0.06983777182176709, "leak": 0.7733723395504057, "replacement": 1.1067725867033005}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 5.439076900482178, "loss_scale": 4.0, "retries": 0, "seconds": 286.1, "max_memory_gb": 23.0, "pair": 0.19904851727187634, "sign": 0.25786498934030533, "keep": 0.051835260121151805, "leak": 1.9066004380583763, "replacement": 0.8305487334728241}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 3.0171892642974854, "loss_scale": 4.0, "retries": 0, "seconds": 287.3, "max_memory_gb": 23.0, "pair": 0.3574173338711262, "sign": 0.37978423945605755, "keep": 0.16462879278697073, "leak": 1.0597558451336226, "replacement": 0.8945628330111504}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 5.4273176193237305, "loss_scale": 4.0, "retries": 0, "seconds": 288.5, "max_memory_gb": 23.0, "pair": 0.3218783661723137, "sign": 0.18004426080733538, "keep": 0.026972423773258924, "leak": 0.7802699697203934, "replacement": 0.6204542145133018}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 9.04145622253418, "loss_scale": 4.0, "retries": 0, "seconds": 289.6, "max_memory_gb": 23.0, "pair": 0.26265177922323346, "sign": 0.31196127366274595, "keep": 0.24377998034469783, "leak": 1.549666404724121, "replacement": 0.9279389828443527}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 4.014421463012695, "loss_scale": 4.0, "retries": 0, "seconds": 290.7, "max_memory_gb": 23.0, "pair": 0.18539921753108501, "sign": 0.2733860341832042, "keep": 0.24891722295433283, "leak": 1.6549075990915298, "replacement": 0.8602777868509293}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 4.799049377441406, "loss_scale": 4.0, "retries": 0, "seconds": 291.8, "max_memory_gb": 23.0, "pair": 0.41057211346924305, "sign": 0.5842252746224403, "keep": 0.08939477836247534, "leak": 0.574090301990509, "replacement": 0.8909046947956085}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 5.3620924949646, "loss_scale": 4.0, "retries": 0, "seconds": 293.0, "max_memory_gb": 23.0, "pair": 0.43855724297463894, "sign": 0.4967958349734545, "keep": 0.1086569931358099, "leak": 0.8457556853536516, "replacement": 0.952722579240799}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 13.557703018188477, "loss_scale": 4.0, "retries": 0, "seconds": 294.1, "max_memory_gb": 23.0, "pair": 0.46597615256905556, "sign": 0.5262753954157233, "keep": 0.3084590577054769, "leak": 1.4365732297301292, "replacement": 1.2424866929650307}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 3.664546251296997, "loss_scale": 4.0, "retries": 0, "seconds": 295.3, "max_memory_gb": 23.0, "pair": 0.1449588006362319, "sign": 0.12104420457035303, "keep": 0.25002062506973743, "leak": 1.5883976817131042, "replacement": 0.7275906503200531}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.4309566020965576, "loss_scale": 4.0, "retries": 0, "seconds": 296.4, "max_memory_gb": 23.0, "pair": 0.16770266741514206, "sign": 0.1619489397853613, "keep": 0.003916683839634061, "leak": 0.8844465725123882, "replacement": 0.47174713015556335}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 7.609361171722412, "loss_scale": 4.0, "retries": 0, "seconds": 297.5, "max_memory_gb": 23.0, "pair": 0.248977136798203, "sign": 0.3549211136996746, "keep": 0.20084343012422323, "leak": 1.140316516160965, "replacement": 0.8119385465979576}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.6294751167297363, "loss_scale": 4.0, "retries": 0, "seconds": 298.7, "max_memory_gb": 23.0, "pair": 0.2746466640383005, "sign": 0.3027808526530862, "keep": 0.23282601591199636, "leak": 0.9183049253188074, "replacement": 0.7720263190567493}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 6.193593502044678, "loss_scale": 4.0, "retries": 0, "seconds": 299.8, "max_memory_gb": 23.0, "pair": 0.4207396134734154, "sign": 0.5706333415582776, "keep": 0.5507120348047465, "leak": 0.6144582405686378, "replacement": 1.1350268721580505}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 5.107265472412109, "loss_scale": 4.0, "retries": 0, "seconds": 301.0, "max_memory_gb": 23.0, "pair": 0.28023871034383774, "sign": 0.3327009715139866, "keep": 0.2824597842991352, "leak": 1.8416130244731903, "replacement": 1.0482223629951477}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 6.221806049346924, "loss_scale": 4.0, "retries": 0, "seconds": 302.1, "max_memory_gb": 23.0, "pair": 0.2618696019053459, "sign": 0.15674300398677588, "keep": 0.9036859400221147, "leak": 2.1790815449494403, "replacement": 1.336854487657547}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 2.783923387527466, "loss_scale": 4.0, "retries": 0, "seconds": 303.3, "max_memory_gb": 23.0, "pair": 0.22954410314559937, "sign": 0.26255439035594463, "keep": 0.0797152011655271, "leak": 1.8103032000362873, "replacement": 0.8532547056674957}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 10.389420509338379, "loss_scale": 4.0, "retries": 0, "seconds": 304.4, "max_memory_gb": 23.0, "pair": 0.5881539229303598, "sign": 0.6079649524763227, "keep": 1.1899748174473643, "leak": 0.3862255662679672, "replacement": 1.5836801752448082}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 4.480903148651123, "loss_scale": 4.0, "retries": 0, "seconds": 305.5, "max_memory_gb": 23.0, "pair": 0.4112166464328766, "sign": 0.4672810174524784, "keep": 0.4867129873018712, "leak": 0.5495897983200848, "replacement": 1.0256111025810242}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 2.3917934894561768, "loss_scale": 4.0, "retries": 0, "seconds": 306.7, "max_memory_gb": 23.0, "pair": 0.5424625221639872, "sign": 0.6804901733994484, "keep": 0.30853175465017557, "leak": 0.23724475968629122, "replacement": 1.096284680068493}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 3.874519109725952, "loss_scale": 4.0, "retries": 0, "seconds": 307.8, "max_memory_gb": 23.0, "pair": 0.24911881051957607, "sign": 0.3505039978772402, "keep": 1.3796049786615185, "leak": 0.8046101704239845, "replacement": 1.3153258115053177}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 9.63563060760498, "loss_scale": 4.0, "retries": 0, "seconds": 309.0, "max_memory_gb": 23.0, "pair": 0.2599352328106761, "sign": 0.4869196331128478, "keep": 0.7635422255843878, "leak": 1.744860053062439, "replacement": 1.3213812038302422}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 3.6282875537872314, "loss_scale": 4.0, "retries": 0, "seconds": 310.1, "max_memory_gb": 23.0, "pair": 0.22989997640252113, "sign": 0.21387137193232775, "keep": 0.06263144826516509, "leak": 1.9682108163833618, "replacement": 0.8602041006088257}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 4.557618141174316, "loss_scale": 4.0, "retries": 0, "seconds": 311.3, "max_memory_gb": 23.0, "pair": 0.08820135146379471, "sign": 0.040670559741556644, "keep": 0.03346941713243723, "leak": 1.08911338634789, "replacement": 0.39754968136548996}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 2.749295473098755, "loss_scale": 4.0, "retries": 0, "seconds": 312.4, "max_memory_gb": 23.0, "pair": 0.3233995996415615, "sign": 0.15151990950107574, "keep": 0.041587689018342644, "leak": 0.11069908086210489, "replacement": 0.4476281777024269}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 9.195869445800781, "loss_scale": 4.0, "retries": 0, "seconds": 313.5, "max_memory_gb": 23.0, "pair": 0.5308436825871468, "sign": 0.7455193996429443, "keep": 0.19866405986249447, "leak": 0.20974385346562485, "replacement": 1.0553714036941528}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 2.7339603900909424, "loss_scale": 4.0, "retries": 0, "seconds": 314.7, "max_memory_gb": 23.0, "pair": 0.38919142819941044, "sign": 0.46010712534189224, "keep": 1.6452486515045166, "leak": 0.4533907826989889, "replacement": 1.5552169606089592}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 2.2861385345458984, "loss_scale": 4.0, "retries": 0, "seconds": 315.8, "max_memory_gb": 23.0, "pair": 0.20337342470884323, "sign": 0.09675574861466885, "keep": 0.0605602043797262, "leak": 0.4049894521012902, "replacement": 0.3832787722349167}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 1.780908226966858, "loss_scale": 4.0, "retries": 0, "seconds": 317.0, "max_memory_gb": 23.0, "pair": 0.3026423305273056, "sign": 0.25684527307748795, "keep": 0.023320434498600662, "leak": 0.2891805097460747, "replacement": 0.5150203034281731}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 14.405072212219238, "loss_scale": 4.0, "retries": 0, "seconds": 318.1, "max_memory_gb": 23.0, "pair": 0.3916386365890503, "sign": 0.5841216593980789, "keep": 0.6824594068457372, "leak": 0.5071621239185333, "replacement": 1.1517197415232658}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 3.7279562950134277, "loss_scale": 4.0, "retries": 0, "seconds": 319.2, "max_memory_gb": 23.0, "pair": 0.3713834509253502, "sign": 0.25326235592365265, "keep": 0.1001231458503753, "leak": 1.1708178599365056, "replacement": 0.840780682861805}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 2.5914974212646484, "loss_scale": 4.0, "retries": 0, "seconds": 320.3, "max_memory_gb": 23.0, "pair": 0.17074337042868137, "sign": 0.11780459620058537, "keep": 0.03472635766956955, "leak": 0.7853930704295635, "replacement": 0.44335711747407913}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.655454158782959, "loss_scale": 4.0, "retries": 0, "seconds": 321.5, "max_memory_gb": 23.0, "pair": 0.23687946237623692, "sign": 0.22483309358358383, "keep": 0.3437564166961238, "leak": 0.4166352039319463, "replacement": 0.6253330409526825}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.201834201812744, "loss_scale": 4.0, "retries": 0, "seconds": 322.6, "max_memory_gb": 23.0, "pair": 0.3338556978851557, "sign": 0.35434758057817817, "keep": 0.22739297186490148, "leak": 0.2360707763582468, "replacement": 0.6837436631321907}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 3.1990530490875244, "loss_scale": 4.0, "retries": 0, "seconds": 323.8, "max_memory_gb": 23.0, "pair": 0.26375055126845837, "sign": 0.11262627132236958, "keep": 0.17144018784165382, "leak": 1.9160674214363098, "replacement": 0.8848006427288055}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.185725212097168, "loss_scale": 4.0, "retries": 0, "seconds": 324.9, "max_memory_gb": 23.0, "pair": 0.2413907591253519, "sign": 0.179589981213212, "keep": 0.10811584629118443, "leak": 0.9165062513202429, "replacement": 0.6143702268600464}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 4.996580600738525, "loss_scale": 4.0, "retries": 0, "seconds": 326.1, "max_memory_gb": 23.0, "pair": 0.2995796427130699, "sign": 0.4881811700761318, "keep": 0.19796505570411682, "leak": 1.9854691135697067, "replacement": 1.1390200406312943}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 6.944903373718262, "loss_scale": 4.0, "retries": 0, "seconds": 327.3, "max_memory_gb": 23.0, "pair": 0.32091885805130005, "sign": 0.46405675634741783, "keep": 0.039117252454161644, "leak": 1.4856266416609287, "replacement": 0.9439125210046768}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 9.311741828918457, "loss_scale": 4.0, "retries": 0, "seconds": 328.4, "max_memory_gb": 23.0, "pair": 0.31305401399731636, "sign": 0.41475074272602797, "keep": 0.3973661009222269, "leak": 0.12431501084938645, "replacement": 0.7501911744475365}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 5.099273681640625, "loss_scale": 4.0, "retries": 0, "seconds": 329.6, "max_memory_gb": 23.0, "pair": 0.3860047198832035, "sign": 0.33773592626675963, "keep": 0.06563261244446039, "leak": 0.5730915600433946, "replacement": 0.7309618629515171}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 5.552136421203613, "loss_scale": 4.0, "retries": 0, "seconds": 330.8, "max_memory_gb": 23.0, "pair": 0.2756083942949772, "sign": 0.19123393017798662, "keep": 0.12440438941121101, "leak": 1.771641219034791, "replacement": 0.8763378709554672}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 4.9225945472717285, "loss_scale": 4.0, "retries": 0, "seconds": 331.9, "max_memory_gb": 23.0, "pair": 0.3087172843515873, "sign": 0.35504582757130265, "keep": 0.3661956502473913, "leak": 0.5354861654341221, "replacement": 0.8032095432281494}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 3.2494592666625977, "loss_scale": 4.0, "retries": 0, "seconds": 333.0, "max_memory_gb": 23.0, "pair": 0.10422098636627197, "sign": 0.053993046283721924, "keep": 0.023469975916668773, "leak": 1.70870803296566, "replacement": 0.5701295137405396}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 2.826464891433716, "loss_scale": 4.0, "retries": 0, "seconds": 334.2, "max_memory_gb": 23.0, "pair": 0.17362226732075214, "sign": 0.13082365598529577, "keep": 0.03965310472995043, "leak": 0.8278294205665588, "replacement": 0.46581801027059555}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 3.144148826599121, "loss_scale": 4.0, "retries": 0, "seconds": 335.3, "max_memory_gb": 23.0, "pair": 0.22038300056010485, "sign": 0.24017190467566252, "keep": 0.0529929562471807, "leak": 0.5972745157778263, "replacement": 0.5162840634584427}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 4.447323799133301, "loss_scale": 4.0, "retries": 0, "seconds": 336.5, "max_memory_gb": 23.0, "pair": 0.27917185612022877, "sign": 0.2049727849662304, "keep": 0.33265599980950356, "leak": 1.5073732137680054, "replacement": 0.9248295575380325}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 10.873099327087402, "loss_scale": 4.0, "retries": 0, "seconds": 337.6, "max_memory_gb": 23.0, "pair": 0.24885324761271477, "sign": 0.42158010229468346, "keep": 0.4351744062732905, "leak": 1.8376565426588058, "replacement": 1.1366447061300278}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 13.246459007263184, "loss_scale": 4.0, "retries": 0, "seconds": 338.7, "max_memory_gb": 23.0, "pair": 0.463287191465497, "sign": 0.8219549991190434, "keep": 0.9387032960075885, "leak": 1.531460015103221, "replacement": 1.7264813929796219}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 9.63514232635498, "loss_scale": 4.0, "retries": 0, "seconds": 339.8, "max_memory_gb": 23.0, "pair": 0.4585324265062809, "sign": 0.6877760365605354, "keep": 0.37674535217229277, "leak": 0.8124069893965498, "replacement": 1.1938948780298233}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 5.380041599273682, "loss_scale": 4.0, "retries": 0, "seconds": 341.0, "max_memory_gb": 23.0, "pair": 0.49291756749153137, "sign": 0.6464942470192909, "keep": 0.7390589638380334, "leak": 0.21431018318980932, "replacement": 1.2392717152833939}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 3.4251902103424072, "loss_scale": 4.0, "retries": 0, "seconds": 342.1, "max_memory_gb": 23.0, "pair": 0.31448879837989807, "sign": 0.11499857529997826, "keep": 0.04594242852181196, "leak": 1.4570822780951858, "replacement": 0.7592298686504364}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 11.45888900756836, "loss_scale": 4.0, "retries": 0, "seconds": 343.3, "max_memory_gb": 23.0, "pair": 0.6216118074953556, "sign": 0.8512309826910496, "keep": 0.6229223348200321, "leak": 1.4343696041032672, "replacement": 1.7172809541225433}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 3.6293108463287354, "loss_scale": 4.0, "retries": 0, "seconds": 344.4, "max_memory_gb": 23.0, "pair": 0.21933257952332497, "sign": 0.32233694661408663, "keep": 0.06921634241007268, "leak": 0.7520833188082179, "replacement": 0.6031300500035286}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 1.828339695930481, "loss_scale": 4.0, "retries": 0, "seconds": 345.6, "max_memory_gb": 23.0, "pair": 0.12411586567759514, "sign": 0.04763958463445306, "keep": 0.0957481429213658, "leak": 0.5555713977664709, "replacement": 0.3347025830298662}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 4.441531181335449, "loss_scale": 4.0, "retries": 0, "seconds": 346.7, "max_memory_gb": 23.0, "pair": 0.28021733090281487, "sign": 0.24125874252058566, "keep": 0.06655411614337936, "leak": 1.5801160298287868, "replacement": 0.8291527777910233}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 4.811156749725342, "loss_scale": 4.0, "retries": 0, "seconds": 347.9, "max_memory_gb": 23.0, "pair": 0.4512258602771908, "sign": 0.5081263212487102, "keep": 0.05146020743995905, "leak": 1.110128398402594, "replacement": 1.008551231585443}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 3.890117883682251, "loss_scale": 4.0, "retries": 0, "seconds": 349.0, "max_memory_gb": 23.0, "pair": 0.17168531753122807, "sign": 0.10743658803403378, "keep": 0.051454945700243115, "leak": 2.225305959582329, "replacement": 0.8074575662612915}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 3.118349313735962, "loss_scale": 4.0, "retries": 0, "seconds": 350.2, "max_memory_gb": 23.0, "pair": 0.13176198117434978, "sign": 0.04104594932869077, "keep": 0.06764176394790411, "leak": 1.0834214598871768, "replacement": 0.4569611996412277}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 13.07935905456543, "loss_scale": 4.0, "retries": 0, "seconds": 351.3, "max_memory_gb": 23.0, "pair": 0.4256477616727352, "sign": 0.4289118992164731, "keep": 0.4578236830420792, "leak": 0.7338235676288605, "replacement": 1.0524714663624763}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 4.894003868103027, "loss_scale": 4.0, "retries": 0, "seconds": 352.5, "max_memory_gb": 23.0, "pair": 0.18440549075603485, "sign": 0.3266696408390999, "keep": 0.09344322769902647, "leak": 1.7715176176279783, "replacement": 0.8373413048684597}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 18.226831436157227, "loss_scale": 4.0, "retries": 0, "seconds": 353.6, "max_memory_gb": 23.0, "pair": 0.5631049312651157, "sign": 0.7478979388251901, "keep": 0.6798035544343293, "leak": 0.8545737527310848, "replacement": 1.4905991405248642}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 1.7148351669311523, "loss_scale": 4.0, "retries": 0, "seconds": 354.7, "max_memory_gb": 23.0, "pair": 0.3091508373618126, "sign": 0.10507931467145681, "keep": 0.04720705549698323, "leak": 0.889103427529335, "replacement": 0.6075698882341385}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 4.405803680419922, "loss_scale": 4.0, "retries": 0, "seconds": 355.8, "max_memory_gb": 23.0, "pair": 0.3190994206815958, "sign": 0.4251459501683712, "keep": 0.36274751770542935, "leak": 0.613613985478878, "replacement": 0.8664496392011642}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.6625447273254395, "loss_scale": 4.0, "retries": 0, "seconds": 357.0, "max_memory_gb": 23.0, "pair": 0.31687253899872303, "sign": 0.1908800322562456, "keep": 0.08627189282560721, "leak": 0.23322991654276848, "replacement": 0.5137559920549393}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 13.0254487991333, "loss_scale": 4.0, "retries": 0, "seconds": 358.1, "max_memory_gb": 23.0, "pair": 0.45689060539007187, "sign": 0.6510860277339816, "keep": 0.8315455643460155, "leak": 0.9129916056990623, "replacement": 1.426454283297062}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 9.789043426513672, "loss_scale": 4.0, "retries": 0, "seconds": 359.3, "max_memory_gb": 23.0, "pair": 0.34507929906249046, "sign": 0.4430419895797968, "keep": 0.5720962251070887, "leak": 1.3549716621637344, "replacement": 1.1913913041353226}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 4.267920970916748, "loss_scale": 4.0, "retries": 0, "seconds": 360.6, "max_memory_gb": 23.0, "pair": 0.17705596424639225, "sign": 0.19800132187083364, "keep": 0.06183768156915903, "leak": 0.9485773742198944, "replacement": 0.5441198199987411}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 3.0858168601989746, "loss_scale": 4.0, "retries": 0, "seconds": 361.7, "max_memory_gb": 23.0, "pair": 0.5268115289509296, "sign": 0.7153337001800537, "keep": 0.3630844270810485, "leak": 1.2946237062569708, "replacement": 1.3896765112876892}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 5.995314121246338, "loss_scale": 4.0, "retries": 0, "seconds": 362.8, "max_memory_gb": 23.0, "pair": 0.2826489470899105, "sign": 0.36686033196747303, "keep": 0.18177958950400352, "leak": 1.9037235230207443, "replacement": 1.0328997820615768}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 4.511237144470215, "loss_scale": 4.0, "retries": 0, "seconds": 363.9, "max_memory_gb": 23.0, "pair": 0.24419766291975975, "sign": 0.22964700870215893, "keep": 0.04797168879304081, "leak": 1.3455157671123743, "replacement": 0.7193859368562698}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 4.66269588470459, "loss_scale": 4.0, "retries": 0, "seconds": 365.1, "max_memory_gb": 23.0, "pair": 0.34408263862133026, "sign": 0.32270921114832163, "keep": 0.0829122569411993, "leak": 0.536703385412693, "replacement": 0.6810692027211189}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 6.3285322189331055, "loss_scale": 4.0, "retries": 0, "seconds": 366.2, "max_memory_gb": 23.0, "pair": 0.3131205663084984, "sign": 0.370801049284637, "keep": 0.3594488780363463, "leak": 0.7118671536445618, "replacement": 0.8562123253941536}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 5.276731491088867, "loss_scale": 4.0, "retries": 0, "seconds": 367.3, "max_memory_gb": 23.0, "pair": 0.3196646012365818, "sign": 0.2513533839955926, "keep": 0.028168131946586072, "leak": 0.43794841144699603, "replacement": 0.5689124688506126}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 5.4875640869140625, "loss_scale": 4.0, "retries": 0, "seconds": 368.4, "max_memory_gb": 23.0, "pair": 0.4147219955921173, "sign": 0.4858711324632168, "keep": 0.04917985945940018, "leak": 0.09982337799294783, "replacement": 0.7072032913565636}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 6.9769415855407715, "loss_scale": 4.0, "retries": 0, "seconds": 369.6, "max_memory_gb": 23.0, "pair": 0.08524844981729984, "sign": 0.08326566172763705, "keep": 0.07106717582792044, "leak": 3.4919066727161407, "replacement": 1.0353915318846703}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 3.1180763244628906, "loss_scale": 4.0, "retries": 0, "seconds": 370.7, "max_memory_gb": 23.0, "pair": 0.1124331122264266, "sign": 0.06292471056804061, "keep": 0.06185653863940388, "leak": 1.2256888491101563, "replacement": 0.4812459386885166}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 9.809752464294434, "loss_scale": 4.0, "retries": 0, "seconds": 371.9, "max_memory_gb": 23.0, "pair": 0.6727000698447227, "sign": 0.6365289241075516, "keep": 0.5029067663708702, "leak": 0.0761520354917593, "replacement": 1.261455975472927}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 9.08289623260498, "loss_scale": 4.0, "retries": 0, "seconds": 373.0, "max_memory_gb": 23.0, "pair": 0.43089135363698006, "sign": 0.5133284591138363, "keep": 0.3149787295842543, "leak": 0.4758507318329066, "replacement": 0.9640076532959938}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 3.380741834640503, "loss_scale": 4.0, "retries": 0, "seconds": 374.2, "max_memory_gb": 23.0, "pair": 0.15399709530174732, "sign": 0.08889339445158839, "keep": 0.3279125140979886, "leak": 0.7172122485935688, "replacement": 0.54170311242342}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 9.01056957244873, "loss_scale": 4.0, "retries": 0, "seconds": 375.3, "max_memory_gb": 23.0, "pair": 0.41859713196754456, "sign": 0.5723500326275826, "keep": 0.21079998841742054, "leak": 0.2767026738729328, "replacement": 0.8793478459119797}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 9.46341323852539, "loss_scale": 4.0, "retries": 0, "seconds": 376.4, "max_memory_gb": 23.0, "pair": 0.3083978658542037, "sign": 0.43042536452412605, "keep": 0.40534141345415264, "leak": 0.739383190870285, "replacement": 0.9111270606517792}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 6.533886432647705, "loss_scale": 4.0, "retries": 0, "seconds": 377.6, "max_memory_gb": 23.0, "pair": 0.5618988163769245, "sign": 0.6069060075096786, "keep": 0.3561354603152722, "leak": 0.4923899651039392, "replacement": 1.1665170416235924}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 8.53930377960205, "loss_scale": 4.0, "retries": 0, "seconds": 378.7, "max_memory_gb": 23.0, "pair": 0.5731668137013912, "sign": 0.6269158981740475, "keep": 0.6617274014861323, "leak": 1.0268507064320147, "replacement": 1.474201150238514}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 4.625776767730713, "loss_scale": 4.0, "retries": 0, "seconds": 379.9, "max_memory_gb": 23.0, "pair": 0.2544789146631956, "sign": 0.25157036259770393, "keep": 0.36672212940175086, "leak": 0.6989792101085186, "replacement": 0.738369956612587}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 5.377105712890625, "loss_scale": 4.0, "retries": 0, "seconds": 381.0, "max_memory_gb": 23.0, "pair": 0.5501934383064508, "sign": 0.7114479877054691, "keep": 0.5617719078436494, "leak": 0.2746169042366091, "replacement": 1.2554575949907303}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 6.240431308746338, "loss_scale": 4.0, "retries": 0, "seconds": 382.2, "max_memory_gb": 23.0, "pair": 0.5527026355266571, "sign": 0.5956306010484695, "keep": 0.6955610355362296, "leak": 1.2716271643876098, "replacement": 1.5162051767110825}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 3.3818109035491943, "loss_scale": 4.0, "retries": 0, "seconds": 383.3, "max_memory_gb": 23.0, "pair": 0.472503449767828, "sign": 0.5190773662179708, "keep": 0.8061785742174834, "leak": 0.35042653139680624, "replacement": 1.222738079726696}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 4.305243968963623, "loss_scale": 4.0, "retries": 0, "seconds": 384.5, "max_memory_gb": 23.0, "pair": 0.2402423396706581, "sign": 0.1353775802999735, "keep": 0.04715404484886676, "leak": 1.241971306502819, "replacement": 0.6420009657740593}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 6.885233402252197, "loss_scale": 4.0, "retries": 0, "seconds": 385.7, "max_memory_gb": 23.0, "pair": 0.6398495212197304, "sign": 0.7546231523156166, "keep": 0.2978591322898865, "leak": 1.0827038214774802, "replacement": 1.4367666095495224}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 6.110523700714111, "loss_scale": 4.0, "retries": 0, "seconds": 386.8, "max_memory_gb": 23.0, "pair": 0.43499623611569405, "sign": 0.5316115599125624, "keep": 0.45377607218688354, "leak": 0.6852402407675982, "replacement": 1.099000096321106}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 5.9092020988464355, "loss_scale": 4.0, "retries": 0, "seconds": 388.0, "max_memory_gb": 23.0, "pair": 0.4392872378230095, "sign": 0.754046868532896, "keep": 0.33345615572761744, "leak": 1.2906097993254662, "replacement": 1.3056912012398243}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 6.05919075012207, "loss_scale": 4.0, "retries": 0, "seconds": 389.1, "max_memory_gb": 23.0, "pair": 0.34361843951046467, "sign": 0.3503338675945997, "keep": 0.644663107173983, "leak": 0.5668659843504429, "replacement": 0.9828334078192711}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 5.53217887878418, "loss_scale": 4.0, "retries": 0, "seconds": 390.2, "max_memory_gb": 23.0, "pair": 0.29214961640536785, "sign": 0.3265683678910136, "keep": 0.2647063718177378, "leak": 1.320682030171156, "replacement": 0.9179575070738792}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 6.4315924644470215, "loss_scale": 4.0, "retries": 0, "seconds": 391.4, "max_memory_gb": 23.0, "pair": 0.5073304455727339, "sign": 0.6383837461471558, "keep": 0.6965810777619481, "leak": 0.9335858151316643, "replacement": 1.408209279179573}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 3.8129944801330566, "loss_scale": 4.0, "retries": 0, "seconds": 392.5, "max_memory_gb": 23.0, "pair": 0.4491156227886677, "sign": 0.4751017428934574, "keep": 0.02891445381101221, "leak": 0.20530754113497096, "replacement": 0.7524505779147148}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 3.76444673538208, "loss_scale": 4.0, "retries": 0, "seconds": 393.7, "max_memory_gb": 23.0, "pair": 0.1586800292134285, "sign": 0.12444955669343472, "keep": 0.01842613483313471, "leak": 0.9467063397169113, "replacement": 0.4667944610118866}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 4.553337097167969, "loss_scale": 4.0, "retries": 0, "seconds": 394.8, "max_memory_gb": 23.0, "pair": 0.27139599807560444, "sign": 0.3651107903569937, "keep": 0.2471697241999209, "leak": 0.7204891364090145, "replacement": 0.757658526301384}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 11.51271915435791, "loss_scale": 4.0, "retries": 0, "seconds": 395.9, "max_memory_gb": 23.0, "pair": 0.4506417028605938, "sign": 0.5675280205905437, "keep": 0.6210382017306983, "leak": 1.1382394209504128, "replacement": 1.3294847160577774}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 2.9421133995056152, "loss_scale": 4.0, "retries": 0, "seconds": 397.1, "max_memory_gb": 23.0, "pair": 0.40086781792342663, "sign": 0.37179330363869667, "keep": 0.2587196449749172, "leak": 0.5317501232493669, "replacement": 0.8490618318319321}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 2.6986351013183594, "loss_scale": 4.0, "retries": 0, "seconds": 398.2, "max_memory_gb": 23.0, "pair": 0.22463781293481588, "sign": 0.11668086238205433, "keep": 0.017896165954880416, "leak": 1.9320239491062239, "replacement": 0.7749323472380638}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 3.1060991287231445, "loss_scale": 4.0, "retries": 0, "seconds": 399.4, "max_memory_gb": 23.0, "pair": 0.30179100297391415, "sign": 0.3311692252755165, "keep": 0.11925298132700846, "leak": 1.3033531364053488, "replacement": 0.8528404161334038}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 4.943804740905762, "loss_scale": 4.0, "retries": 0, "seconds": 400.5, "max_memory_gb": 23.0, "pair": 0.3174084909260273, "sign": 0.587540976703167, "keep": 0.10440882109105587, "leak": 0.5443868972361088, "replacement": 0.7994801104068756}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 2.579726219177246, "loss_scale": 4.0, "retries": 0, "seconds": 401.7, "max_memory_gb": 23.0, "pair": 0.19093332905322313, "sign": 0.1162048913538456, "keep": 0.07438205322250724, "leak": 0.5351399630308151, "replacement": 0.4200117979198694}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 10.829132080078125, "loss_scale": 4.0, "retries": 0, "seconds": 402.8, "max_memory_gb": 23.0, "pair": 0.3888346031308174, "sign": 0.6078325561247766, "keep": 0.4786092173308134, "leak": 0.9767677709460258, "replacement": 1.1762474179267883}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 7.254215240478516, "loss_scale": 4.0, "retries": 0, "seconds": 403.9, "max_memory_gb": 23.0, "pair": 0.3650926351547241, "sign": 0.47824189253151417, "keep": 0.5982615663670003, "leak": 0.5049754735082388, "replacement": 1.0295882523059845}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 3.716956615447998, "loss_scale": 4.0, "retries": 0, "seconds": 405.0, "max_memory_gb": 23.0, "pair": 0.42071252688765526, "sign": 0.42910314351320267, "keep": 0.2855052144732326, "leak": 0.07741544891541707, "replacement": 0.7973705604672432}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 2.5966639518737793, "loss_scale": 4.0, "retries": 0, "seconds": 406.2, "max_memory_gb": 23.0, "pair": 0.5278749093413353, "sign": 0.4524799957871437, "keep": 0.18910578498616815, "leak": 1.172769809141755, "replacement": 1.1418602764606476}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 5.24293327331543, "loss_scale": 4.0, "retries": 0, "seconds": 407.4, "max_memory_gb": 23.0, "pair": 0.14320328272879124, "sign": 0.15050962008535862, "keep": 0.027753202011808753, "leak": 2.293235182762146, "replacement": 0.8056434839963913}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 6.475088119506836, "loss_scale": 4.0, "retries": 0, "seconds": 408.5, "max_memory_gb": 23.0, "pair": 0.18102547340095043, "sign": 0.12093369220383465, "keep": 0.14288580417633057, "leak": 2.306131578050554, "replacement": 0.8894680887460709}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 5.0432868003845215, "loss_scale": 4.0, "retries": 0, "seconds": 409.7, "max_memory_gb": 23.0, "pair": 0.4902876019477844, "sign": 0.6363932341337204, "keep": 0.5068803275935352, "leak": 0.2915552780032158, "replacement": 1.1348132640123367}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 6.589328765869141, "loss_scale": 4.0, "retries": 0, "seconds": 410.8, "max_memory_gb": 23.0, "pair": 0.3729701582342386, "sign": 0.44833777472376823, "keep": 0.1110645072767511, "leak": 1.0629930272698402, "replacement": 0.9184195622801781}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 4.418521404266357, "loss_scale": 4.0, "retries": 0, "seconds": 412.0, "max_memory_gb": 23.0, "pair": 0.19484948739409447, "sign": 0.18548202328383923, "keep": 0.031731383642181754, "leak": 1.6926534697413445, "replacement": 0.72661954164505}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 7.577271938323975, "loss_scale": 4.0, "retries": 0, "seconds": 413.1, "max_memory_gb": 23.0, "pair": 0.16389738209545612, "sign": 0.3145665396004915, "keep": 0.034459339338354766, "leak": 1.1311267465353012, "replacement": 0.6211920008063316}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 5.132988452911377, "loss_scale": 4.0, "retries": 0, "seconds": 414.3, "max_memory_gb": 23.0, "pair": 0.39945048838853836, "sign": 0.4161628456786275, "keep": 1.0469779563136399, "leak": 0.22803398940595798, "replacement": 1.1880293861031532}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 9.194564819335938, "loss_scale": 4.0, "retries": 0, "seconds": 415.4, "max_memory_gb": 23.0, "pair": 0.39497984014451504, "sign": 0.8138383328914642, "keep": 0.028068334329873323, "leak": 0.9666075790300965, "replacement": 1.0575850903987885}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 2.6086699962615967, "loss_scale": 4.0, "retries": 0, "seconds": 416.5, "max_memory_gb": 23.0, "pair": 0.1674228496849537, "sign": 0.0471041826531291, "keep": 0.07472789310850203, "leak": 1.3123141434043646, "replacement": 0.5564174205064774}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 3.236433267593384, "loss_scale": 4.0, "retries": 0, "seconds": 417.7, "max_memory_gb": 23.0, "pair": 0.30302223190665245, "sign": 0.35070250928401947, "keep": 0.13972413376905024, "leak": 0.29846497252583504, "replacement": 0.6228518113493919}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 5.630066394805908, "loss_scale": 4.0, "retries": 0, "seconds": 418.9, "max_memory_gb": 23.0, "pair": 0.24386295676231384, "sign": 0.2923405608162284, "keep": 0.04281865246593952, "leak": 2.084527552127838, "replacement": 0.9325744658708572}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 3.9257636070251465, "loss_scale": 4.0, "retries": 0, "seconds": 420.0, "max_memory_gb": 23.0, "pair": 0.36775706708431244, "sign": 0.4048525020480156, "keep": 0.3747113720746711, "leak": 0.7425489313900471, "replacement": 0.9431762173771858}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 2.2020816802978516, "loss_scale": 4.0, "retries": 0, "seconds": 421.1, "max_memory_gb": 23.0, "pair": 0.14608916640281677, "sign": 0.10300534451380372, "keep": 0.06694035674445331, "leak": 0.4394093407317996, "replacement": 0.3409143425524235}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 2.8598473072052, "loss_scale": 4.0, "retries": 0, "seconds": 422.3, "max_memory_gb": 23.0, "pair": 0.4396106153726578, "sign": 0.37009946536272764, "keep": 0.04958758380962536, "leak": 0.5334889627993107, "replacement": 0.7828263640403748}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 10.750975608825684, "loss_scale": 4.0, "retries": 0, "seconds": 423.4, "max_memory_gb": 23.0, "pair": 0.22907835617661476, "sign": 0.27915869653224945, "keep": 1.2234287820756435, "leak": 0.6302302423864603, "replacement": 1.1379296332597733}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.178379535675049, "loss_scale": 4.0, "retries": 0, "seconds": 424.5, "max_memory_gb": 23.0, "pair": 0.16448555327951908, "sign": 0.064686075784266, "keep": 0.186142492108047, "leak": 1.2811365574598312, "replacement": 0.6101839989423752}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 4.254329204559326, "loss_scale": 4.0, "retries": 0, "seconds": 425.7, "max_memory_gb": 23.0, "pair": 0.1897220928221941, "sign": 0.20563737815245986, "keep": 0.04088841157499701, "leak": 0.9717912897467613, "replacement": 0.5559327974915504}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 4.3125786781311035, "loss_scale": 4.0, "retries": 0, "seconds": 426.8, "max_memory_gb": 23.0, "pair": 0.4704523868858814, "sign": 0.7027936987578869, "keep": 0.056181394087616354, "leak": 0.43532449658960104, "replacement": 0.9587710350751877}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 5.498935222625732, "loss_scale": 4.0, "retries": 0, "seconds": 428.0, "max_memory_gb": 23.0, "pair": 0.30893315374851227, "sign": 0.29532624687999487, "keep": 0.4811918744817376, "leak": 0.43822248093783855, "replacement": 0.8067478388547897}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 2.521397352218628, "loss_scale": 4.0, "retries": 0, "seconds": 429.1, "max_memory_gb": 23.0, "pair": 0.44047212414443493, "sign": 0.4777493365108967, "keep": 0.30181130702840164, "leak": 0.6090877330861986, "replacement": 0.9825244247913361}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.6398096084594727, "loss_scale": 4.0, "retries": 0, "seconds": 430.2, "max_memory_gb": 23.0, "pair": 0.23739857226610184, "sign": 0.21129161585122347, "keep": 0.12035586021374911, "leak": 0.09254831721773371, "replacement": 0.4263594038784504}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 3.6877896785736084, "loss_scale": 4.0, "retries": 0, "seconds": 431.3, "max_memory_gb": 23.0, "pair": 0.2630681451410055, "sign": 0.3145934771746397, "keep": 0.03404847113415599, "leak": 0.5018828362226486, "replacement": 0.5628598183393478}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 4.105978012084961, "loss_scale": 4.0, "retries": 0, "seconds": 432.5, "max_memory_gb": 23.0, "pair": 0.2227010242640972, "sign": 0.15649909153580666, "keep": 0.10116155911237001, "leak": 0.8382204361259937, "replacement": 0.5610864460468292}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 13.814407348632812, "loss_scale": 4.0, "retries": 0, "seconds": 433.6, "max_memory_gb": 23.0, "pair": 0.2175271501764655, "sign": 0.3633436653763056, "keep": 0.5402940332423896, "leak": 0.598111230880022, "replacement": 0.8188738077878952}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 8.196992874145508, "loss_scale": 4.0, "retries": 0, "seconds": 434.8, "max_memory_gb": 23.0, "pair": 0.2821372523903847, "sign": 0.3710051095113158, "keep": 0.102189801633358, "leak": 1.8132584691047668, "replacement": 0.9720493257045746}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 3.0862345695495605, "loss_scale": 4.0, "retries": 0, "seconds": 435.9, "max_memory_gb": 23.0, "pair": 0.5266905054450035, "sign": 0.739108495414257, "keep": 0.3982121543958783, "leak": 0.3832260826602578, "replacement": 1.1911573708057404}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 5.613177299499512, "loss_scale": 4.0, "retries": 0, "seconds": 437.1, "max_memory_gb": 23.0, "pair": 0.4511660076677799, "sign": 0.5592398038133979, "keep": 1.0428623436018825, "leak": 0.49165831319987774, "replacement": 1.3751316294074059}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 3.123430013656616, "loss_scale": 4.0, "retries": 0, "seconds": 438.2, "max_memory_gb": 23.0, "pair": 0.3283987231552601, "sign": 0.30652436055243015, "keep": 0.15245556738227606, "leak": 0.6429632387589663, "replacement": 0.7186294868588448}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 4.614628791809082, "loss_scale": 4.0, "retries": 0, "seconds": 439.3, "max_memory_gb": 23.0, "pair": 0.09498613188043237, "sign": 0.05867821676656604, "keep": 0.02368266019038856, "leak": 0.9607747148256749, "replacement": 0.3763602590188384}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 9.336750030517578, "loss_scale": 4.0, "retries": 0, "seconds": 440.5, "max_memory_gb": 23.0, "pair": 0.39134993962943554, "sign": 0.5023520905524492, "keep": 0.3297225672286004, "leak": 0.4619457861408591, "replacement": 0.9228737279772758}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 7.7258429527282715, "loss_scale": 4.0, "retries": 0, "seconds": 441.6, "max_memory_gb": 23.0, "pair": 0.2857392579317093, "sign": 0.14674018742516637, "keep": 0.040842748479917645, "leak": 0.2825116589665413, "replacement": 0.45015863329172134}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 4.895473003387451, "loss_scale": 4.0, "retries": 0, "seconds": 442.8, "max_memory_gb": 23.0, "pair": 0.22550081089138985, "sign": 0.42749738320708275, "keep": 0.07580666232388467, "leak": 0.7692398019134998, "replacement": 0.6694627776741982}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 8.225337028503418, "loss_scale": 4.0, "retries": 0, "seconds": 443.9, "max_memory_gb": 23.0, "pair": 0.22450262866914272, "sign": 0.24016115721315145, "keep": 0.08793976297602057, "leak": 0.5479425936937332, "replacement": 0.5255387499928474}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 3.4583611488342285, "loss_scale": 4.0, "retries": 0, "seconds": 445.1, "max_memory_gb": 23.0, "pair": 0.2146726055070758, "sign": 0.098756049759686, "keep": 0.16220257617533207, "leak": 1.374855488538742, "replacement": 0.6888657957315445}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 9.085342407226562, "loss_scale": 4.0, "retries": 0, "seconds": 446.3, "max_memory_gb": 23.0, "pair": 0.24810330756008625, "sign": 0.38271453976631165, "keep": 0.19434660382103175, "leak": 1.092954675666988, "replacement": 0.8098725378513336}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 3.7648255825042725, "loss_scale": 4.0, "retries": 0, "seconds": 447.4, "max_memory_gb": 23.0, "pair": 0.49519314896315336, "sign": 0.7576395161449909, "keep": 0.5037503344938159, "leak": 0.5729776145890355, "replacement": 1.2691324800252914}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 3.029369354248047, "loss_scale": 4.0, "retries": 0, "seconds": 448.6, "max_memory_gb": 23.0, "pair": 0.14809469133615494, "sign": 0.08841561106964946, "keep": 0.05341967771528289, "leak": 0.4468662925064564, "replacement": 0.3307288959622383}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 3.5060079097747803, "loss_scale": 4.0, "retries": 0, "seconds": 449.7, "max_memory_gb": 23.0, "pair": 0.14469978166744113, "sign": 0.19490573555231094, "keep": 0.09455795539543033, "leak": 0.49087241291999817, "replacement": 0.41214973479509354}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 5.835056781768799, "loss_scale": 4.0, "retries": 0, "seconds": 450.9, "max_memory_gb": 23.0, "pair": 0.336426205933094, "sign": 0.4237653939053416, "keep": 0.14932980062440038, "leak": 0.5657412372529507, "replacement": 0.7644091136753559}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 5.381505489349365, "loss_scale": 4.0, "retries": 0, "seconds": 452.0, "max_memory_gb": 23.0, "pair": 0.10393089987337589, "sign": 0.13171685114502907, "keep": 0.05992140539456159, "leak": 1.579316034913063, "replacement": 0.5945790261030197}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 2.799009323120117, "loss_scale": 4.0, "retries": 0, "seconds": 453.1, "max_memory_gb": 23.0, "pair": 0.1433060634881258, "sign": 0.055350637063384056, "keep": 0.06603500433266163, "leak": 0.5192845612764359, "replacement": 0.33382003009319305}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 7.060459136962891, "loss_scale": 4.0, "retries": 0, "seconds": 454.2, "max_memory_gb": 23.0, "pair": 0.27105251140892506, "sign": 0.41320979967713356, "keep": 0.22693677758798003, "leak": 0.7829808257520199, "replacement": 0.7868709936738014}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 6.380291938781738, "loss_scale": 4.0, "retries": 0, "seconds": 455.4, "max_memory_gb": 23.0, "pair": 0.3945602551102638, "sign": 0.407135222107172, "keep": 0.21205040183849633, "leak": 0.5384783543122467, "replacement": 0.8387726321816444}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 13.440308570861816, "loss_scale": 4.0, "retries": 0, "seconds": 456.5, "max_memory_gb": 23.0, "pair": 0.42964571714401245, "sign": 0.7711758548393846, "keep": 1.0772898746654391, "leak": 1.038324104505591, "replacement": 1.6134595796465874}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 6.333609580993652, "loss_scale": 4.0, "retries": 0, "seconds": 457.7, "max_memory_gb": 23.0, "pair": 0.40293387323617935, "sign": 0.28471927158534527, "keep": 0.13905958458781242, "leak": 1.1364778876304626, "replacement": 0.8989427834749222}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 18.191816329956055, "loss_scale": 4.0, "retries": 0, "seconds": 458.9, "max_memory_gb": 23.0, "pair": 0.41761556454002857, "sign": 0.5024306932464242, "keep": 0.4336711748037487, "leak": 1.3056727645453066, "replacement": 1.2120846807956696}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 18.605073928833008, "loss_scale": 4.0, "retries": 0, "seconds": 460.1, "max_memory_gb": 23.0, "pair": 0.833061097189784, "sign": 0.9443242410197854, "keep": 0.477199031971395, "leak": 0.6355679849802982, "replacement": 1.702714741230011}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 3.8525428771972656, "loss_scale": 4.0, "retries": 0, "seconds": 461.2, "max_memory_gb": 23.0, "pair": 0.14913460053503513, "sign": 0.09676086297258735, "keep": 0.01724648056551814, "leak": 1.5309535451233387, "replacement": 0.5888766571879387}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 14.385466575622559, "loss_scale": 4.0, "retries": 0, "seconds": 462.4, "max_memory_gb": 23.0, "pair": 0.304631020873785, "sign": 0.5048754485324025, "keep": 0.5093047954142094, "leak": 0.7556259278208017, "replacement": 1.0006276220083237}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 2.9400126934051514, "loss_scale": 4.0, "retries": 0, "seconds": 463.5, "max_memory_gb": 23.0, "pair": 0.22048155590891838, "sign": 0.19509516889229417, "keep": 0.02885630534728989, "leak": 0.5726538933813572, "replacement": 0.4756207540631294}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 5.571751117706299, "loss_scale": 4.0, "retries": 0, "seconds": 464.7, "max_memory_gb": 23.0, "pair": 0.6470455303788185, "sign": 0.7127349432557821, "keep": 0.16190234385430813, "leak": 0.5460028788365889, "replacement": 1.220864862203598}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 10.419891357421875, "loss_scale": 4.0, "retries": 0, "seconds": 465.8, "max_memory_gb": 23.0, "pair": 0.5842901133000851, "sign": 0.9365312857553363, "keep": 2.3357061464339495, "leak": 0.735774502158165, "replacement": 2.4043524861335754}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 8.123747825622559, "loss_scale": 4.0, "retries": 0, "seconds": 467.0, "max_memory_gb": 23.0, "pair": 0.33123316802084446, "sign": 0.47390816756524146, "keep": 0.48462002840824425, "leak": 0.7552827685140073, "replacement": 0.999317929148674}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 5.586654186248779, "loss_scale": 4.0, "retries": 0, "seconds": 468.1, "max_memory_gb": 23.0, "pair": 0.3809269741177559, "sign": 0.4203871600329876, "keep": 0.31749128247611225, "leak": 1.5345536344984794, "replacement": 1.1335045546293259}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 8.283267974853516, "loss_scale": 4.0, "retries": 0, "seconds": 469.3, "max_memory_gb": 23.0, "pair": 0.18239978328347206, "sign": 0.11766491457819939, "keep": 0.614903113222681, "leak": 0.9681155234575272, "replacement": 0.7907126694917679}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 5.6709980964660645, "loss_scale": 4.0, "retries": 0, "seconds": 470.5, "max_memory_gb": 23.0, "pair": 0.2290289094671607, "sign": 0.5922061838209629, "keep": 0.9953726616804488, "leak": 0.9079784899950027, "replacement": 1.2498129457235336}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 2.9553914070129395, "loss_scale": 4.0, "retries": 0, "seconds": 471.6, "max_memory_gb": 23.0, "pair": 0.45609454810619354, "sign": 0.5939013175666332, "keep": 0.22536992002278566, "leak": 0.28029134473763406, "replacement": 0.935803011059761}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 3.9212565422058105, "loss_scale": 4.0, "retries": 0, "seconds": 472.7, "max_memory_gb": 23.0, "pair": 0.14675286319106817, "sign": 0.12235166598111391, "keep": 0.03909699397627264, "leak": 1.076175184454769, "replacement": 0.49652099609375}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 11.549866676330566, "loss_scale": 4.0, "retries": 0, "seconds": 473.9, "max_memory_gb": 23.0, "pair": 0.3188657071441412, "sign": 0.4092135413084179, "keep": 0.8956373380497098, "leak": 0.5698789828456938, "replacement": 1.1137609407305717}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 10.798473358154297, "loss_scale": 4.0, "retries": 0, "seconds": 474.9, "max_memory_gb": 23.0, "pair": 0.44985535368323326, "sign": 0.40477171912789345, "keep": 1.73144849576056, "leak": 0.2243245455028955, "replacement": 1.5740466713905334}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 2.9794654846191406, "loss_scale": 4.0, "retries": 0, "seconds": 476.0, "max_memory_gb": 23.0, "pair": 0.16829273663461208, "sign": 0.11879984475672245, "keep": 0.06245724903419614, "leak": 0.6138970344909467, "replacement": 0.41239554435014725}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 12.503573417663574, "loss_scale": 4.0, "retries": 0, "seconds": 477.2, "max_memory_gb": 23.0, "pair": 0.7510098516941071, "sign": 1.1513695865869522, "keep": 1.3757954509928823, "leak": 0.4943765737116337, "replacement": 2.1381865441799164}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 2.960509777069092, "loss_scale": 4.0, "retries": 0, "seconds": 478.3, "max_memory_gb": 23.0, "pair": 0.09681173507124186, "sign": 0.14318045880645514, "keep": 0.03722254605963826, "leak": 0.6407718425616622, "replacement": 0.3472061939537525}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 7.14133882522583, "loss_scale": 4.0, "retries": 0, "seconds": 479.4, "max_memory_gb": 23.0, "pair": 0.2876040078699589, "sign": 0.5062661692500114, "keep": 0.05526098271366209, "leak": 1.5337590985000134, "replacement": 0.9518073573708534}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 10.627147674560547, "loss_scale": 4.0, "retries": 0, "seconds": 480.5, "max_memory_gb": 23.0, "pair": 0.18532629776746035, "sign": 0.344690166413784, "keep": 0.487683063140139, "leak": 2.574580617249012, "replacement": 1.2451580315828323}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 6.212597846984863, "loss_scale": 4.0, "retries": 0, "seconds": 481.7, "max_memory_gb": 23.0, "pair": 0.14157235343009233, "sign": 0.1987605942413211, "keep": 0.06199247553013265, "leak": 2.0980189591646194, "replacement": 0.7964536473155022}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 5.803227424621582, "loss_scale": 4.0, "retries": 0, "seconds": 482.8, "max_memory_gb": 23.0, "pair": 0.5543537847697735, "sign": 0.5775793613865972, "keep": 0.6031910665333271, "leak": 0.25859270012006164, "replacement": 1.2093871980905533}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 12.559605598449707, "loss_scale": 4.0, "retries": 0, "seconds": 483.9, "max_memory_gb": 23.0, "pair": 0.5832556709647179, "sign": 0.8700540512800217, "keep": 1.1288358494639397, "leak": 0.433397114276886, "replacement": 1.6910498440265656}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 2.1992242336273193, "loss_scale": 4.0, "retries": 0, "seconds": 485.1, "max_memory_gb": 23.0, "pair": 0.1548311971127987, "sign": 0.07713007181882858, "keep": 0.08422225434333086, "leak": 0.5832681953907013, "replacement": 0.3813244067132473}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 4.873504638671875, "loss_scale": 4.0, "retries": 0, "seconds": 486.2, "max_memory_gb": 23.0, "pair": 0.26656409725546837, "sign": 0.17466033063828945, "keep": 0.05512036569416523, "leak": 0.45645122416317463, "replacement": 0.49556726962327957}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 5.880033493041992, "loss_scale": 4.0, "retries": 0, "seconds": 487.4, "max_memory_gb": 23.0, "pair": 0.5780356153845787, "sign": 0.6236945549026132, "keep": 0.4475348126143217, "leak": 0.3005615063011646, "replacement": 1.1887906789779663}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 3.966071605682373, "loss_scale": 4.0, "retries": 0, "seconds": 488.6, "max_memory_gb": 23.0, "pair": 0.25291821733117104, "sign": 0.26135626807808876, "keep": 0.07964603276923299, "leak": 1.903472200036049, "replacement": 0.8992874026298523}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 3.278010606765747, "loss_scale": 4.0, "retries": 0, "seconds": 489.7, "max_memory_gb": 23.0, "pair": 0.1785525158047676, "sign": 0.09555489849299192, "keep": 0.027207852341234684, "leak": 0.3320884411223233, "replacement": 0.32295599579811096}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 8.805530548095703, "loss_scale": 4.0, "retries": 0, "seconds": 490.9, "max_memory_gb": 23.0, "pair": 0.34485501423478127, "sign": 0.09241153346374631, "keep": 0.04426126356702298, "leak": 0.17873485223390162, "replacement": 0.4578751102089882}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 3.3729727268218994, "loss_scale": 4.0, "retries": 0, "seconds": 492.0, "max_memory_gb": 23.0, "pair": 0.23337448574602604, "sign": 0.21973131457343698, "keep": 0.20648721687030047, "leak": 0.4037874788045883, "replacement": 0.5474306270480156}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.621545314788818, "loss_scale": 4.0, "retries": 0, "seconds": 493.1, "max_memory_gb": 23.0, "pair": 0.12969577312469482, "sign": 0.06399055756628513, "keep": 0.08333668275736272, "leak": 1.828597920863345, "replacement": 0.6605088859796524}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 2.182858943939209, "loss_scale": 4.0, "retries": 0, "seconds": 494.3, "max_memory_gb": 23.0, "pair": 0.13648405112326145, "sign": 0.16348666790872812, "keep": 0.09136349405162036, "leak": 0.4006494105924503, "replacement": 0.36407148465514183}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 2.70389723777771, "loss_scale": 4.0, "retries": 0, "seconds": 495.5, "max_memory_gb": 23.0, "pair": 0.15474515408277512, "sign": 0.2271569138392806, "keep": 0.06553019536659122, "leak": 0.7789452336728573, "replacement": 0.49582501500844955}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 3.4445269107818604, "loss_scale": 4.0, "retries": 0, "seconds": 496.6, "max_memory_gb": 23.0, "pair": 0.16349654272198677, "sign": 0.13205436617136002, "keep": 0.028005598462186754, "leak": 0.10984756331890821, "replacement": 0.27098842337727547}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 13.690008163452148, "loss_scale": 4.0, "retries": 0, "seconds": 497.8, "max_memory_gb": 23.0, "pair": 0.5404115195851773, "sign": 0.47455481393262744, "keep": 0.5690463100909255, "leak": 0.4384547099471092, "replacement": 1.1718257814645767}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 7.203232765197754, "loss_scale": 4.0, "retries": 0, "seconds": 498.9, "max_memory_gb": 23.0, "pair": 0.24707417748868465, "sign": 0.2807742580771446, "keep": 0.06627587066031992, "leak": 1.6635684445500374, "replacement": 0.8364913463592529}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 11.407035827636719, "loss_scale": 4.0, "retries": 0, "seconds": 500.0, "max_memory_gb": 23.0, "pair": 0.2653847988694906, "sign": 0.48758023232221603, "keep": 0.64143724180758, "leak": 1.5042960867285728, "replacement": 1.2059675753116608}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.712006688117981, "loss_scale": 4.0, "retries": 0, "seconds": 501.2, "max_memory_gb": 23.0, "pair": 0.12001236202195287, "sign": 0.08014771062880754, "keep": 0.10626725008478388, "leak": 0.4108802159316838, "replacement": 0.3159398939460516}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 4.706146240234375, "loss_scale": 4.0, "retries": 0, "seconds": 502.3, "max_memory_gb": 23.0, "pair": 0.36613578349351883, "sign": 0.42949042841792107, "keep": 0.5172627065330744, "leak": 0.3008698089529389, "replacement": 0.9147298038005829}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 3.6631920337677, "loss_scale": 4.0, "retries": 0, "seconds": 503.5, "max_memory_gb": 23.0, "pair": 0.16266502626240253, "sign": 0.2620982937514782, "keep": 0.015311597264371812, "leak": 0.5039327815175056, "replacement": 0.4273531585931778}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 10.619665145874023, "loss_scale": 4.0, "retries": 0, "seconds": 504.6, "max_memory_gb": 23.0, "pair": 0.6325111500918865, "sign": 0.7608100865036249, "keep": 0.9998863281216472, "leak": 0.9423693418502808, "replacement": 1.7484516836702824}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 3.1835920810699463, "loss_scale": 4.0, "retries": 0, "seconds": 505.7, "max_memory_gb": 23.0, "pair": 0.3906682878732681, "sign": 0.4090475868433714, "keep": 0.07433817198034376, "leak": 1.4345308542251587, "replacement": 0.9909938797354698}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 2.90360164642334, "loss_scale": 4.0, "retries": 0, "seconds": 506.9, "max_memory_gb": 23.0, "pair": 0.17060887441039085, "sign": 0.2710978761315346, "keep": 0.030749991536140442, "leak": 0.6043059714138508, "replacement": 0.4726093113422394}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 7.872247219085693, "loss_scale": 4.0, "retries": 0, "seconds": 508.0, "max_memory_gb": 23.0, "pair": 0.21607713215053082, "sign": 0.39419225277379155, "keep": 0.20875419629737735, "leak": 0.34305328130722046, "replacement": 0.6033136695623398}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 8.304348945617676, "loss_scale": 4.0, "retries": 0, "seconds": 509.1, "max_memory_gb": 23.0, "pair": 0.6480047963559628, "sign": 0.8851305255666375, "keep": 1.1497507456224412, "leak": 0.9627608431037515, "replacement": 1.90613554418087}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 3.4101779460906982, "loss_scale": 4.0, "retries": 0, "seconds": 510.3, "max_memory_gb": 23.0, "pair": 0.12475443724542856, "sign": 0.17179092671722174, "keep": 0.1994736100314185, "leak": 0.6394827714830171, "replacement": 0.47025740146636963}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 5.256328105926514, "loss_scale": 4.0, "retries": 0, "seconds": 511.4, "max_memory_gb": 23.0, "pair": 0.4023434240370989, "sign": 0.548623813316226, "keep": 0.340620968490839, "leak": 1.3940883381292224, "replacement": 1.1954879015684128}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 4.300124168395996, "loss_scale": 4.0, "retries": 0, "seconds": 512.5, "max_memory_gb": 23.0, "pair": 0.1669636331498623, "sign": 0.2111430149525404, "keep": 0.1649183895206079, "leak": 0.2709325213218108, "replacement": 0.42272747308015823}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 4.354294776916504, "loss_scale": 4.0, "retries": 0, "seconds": 513.7, "max_memory_gb": 23.0, "pair": 0.1519322358071804, "sign": 0.1817898340523243, "keep": 0.023951829643920064, "leak": 0.8102917705546133, "replacement": 0.45737600326538086}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 4.419205665588379, "loss_scale": 4.0, "retries": 0, "seconds": 514.8, "max_memory_gb": 23.0, "pair": 0.2816582974046469, "sign": 0.1659792261198163, "keep": 0.24262883141636848, "leak": 0.5841592524666339, "replacement": 0.6320021525025368}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 5.132648468017578, "loss_scale": 4.0, "retries": 0, "seconds": 515.9, "max_memory_gb": 23.0, "pair": 0.10591317154467106, "sign": 0.11108271777629852, "keep": 0.02405240456573665, "leak": 1.429479657439515, "replacement": 0.5308506414294243}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 5.319559574127197, "loss_scale": 4.0, "retries": 0, "seconds": 517.1, "max_memory_gb": 23.0, "pair": 0.2874475698918104, "sign": 0.3260022681206465, "keep": 0.48309229128062725, "leak": 0.9902588231489062, "replacement": 0.9395595565438271}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 13.722280502319336, "loss_scale": 4.0, "retries": 0, "seconds": 518.2, "max_memory_gb": 23.0, "pair": 0.23580910079181194, "sign": 0.29422884061932564, "keep": 0.08810622501187027, "leak": 1.0537138571962714, "replacement": 0.6904051192104816}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 6.237936019897461, "loss_scale": 4.0, "retries": 0, "seconds": 519.4, "max_memory_gb": 23.0, "pair": 0.2931962311267853, "sign": 0.20944202411919832, "keep": 0.2927854172885418, "leak": 0.7452191337943077, "replacement": 0.7306147292256355}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 4.1858344078063965, "loss_scale": 4.0, "retries": 0, "seconds": 520.5, "max_memory_gb": 23.0, "pair": 0.2070546466857195, "sign": 0.2066307831555605, "keep": 0.27690371917560697, "leak": 1.4802383659407496, "replacement": 0.8188815116882324}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 6.179778099060059, "loss_scale": 4.0, "retries": 0, "seconds": 521.6, "max_memory_gb": 23.0, "pair": 0.2914116643369198, "sign": 0.7636331161484122, "keep": 0.04690151009708643, "leak": 0.7745515434799017, "replacement": 0.8903168514370918}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 1.9346065521240234, "loss_scale": 4.0, "retries": 0, "seconds": 522.8, "max_memory_gb": 23.0, "pair": 0.09683968871831894, "sign": 0.16445532999932766, "keep": 0.013002786436118186, "leak": 0.25019481871277094, "replacement": 0.24811745062470436}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 16.99517059326172, "loss_scale": 4.0, "retries": 0, "seconds": 523.9, "max_memory_gb": 23.0, "pair": 0.35716698318719864, "sign": 0.5868253521621227, "keep": 0.5249011758714914, "leak": 0.8505264092236757, "replacement": 1.1256618797779083}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 4.499906063079834, "loss_scale": 4.0, "retries": 0, "seconds": 525.0, "max_memory_gb": 23.0, "pair": 0.2830696124583483, "sign": 0.3016570834442973, "keep": 0.0747171810362488, "leak": 0.216008760035038, "replacement": 0.5252589397132397}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 2.526196002960205, "loss_scale": 4.0, "retries": 0, "seconds": 526.2, "max_memory_gb": 23.0, "pair": 0.1875309757888317, "sign": 0.20145141147077084, "keep": 0.021524077863432467, "leak": 0.31810943596065044, "replacement": 0.3785460628569126}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 14.267602920532227, "loss_scale": 4.0, "retries": 0, "seconds": 527.4, "max_memory_gb": 23.0, "pair": 0.29119530133903027, "sign": 0.6972112953662872, "keep": 0.9201132035814226, "leak": 1.3928201869130135, "replacement": 1.4480626061558723}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 2.308739185333252, "loss_scale": 4.0, "retries": 0, "seconds": 528.5, "max_memory_gb": 23.0, "pair": 0.15612071938812733, "sign": 0.20465119369328022, "keep": 0.022862013924168423, "leak": 0.5024204598739743, "replacement": 0.3954824358224869}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 15.34730339050293, "loss_scale": 4.0, "retries": 0, "seconds": 529.6, "max_memory_gb": 23.0, "pair": 0.2292789239436388, "sign": 0.21383893117308617, "keep": 0.022683855379000306, "leak": 0.37612101424019784, "replacement": 0.44157057255506516}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 11.39493465423584, "loss_scale": 4.0, "retries": 0, "seconds": 530.8, "max_memory_gb": 23.0, "pair": 0.18235394544899464, "sign": 0.1616987599991262, "keep": 0.014666465460322797, "leak": 1.5822715982794762, "replacement": 0.6661044433712959}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 4.550392150878906, "loss_scale": 4.0, "retries": 0, "seconds": 532.0, "max_memory_gb": 23.0, "pair": 0.15826194360852242, "sign": 0.17272529285401106, "keep": 0.1920866334112361, "leak": 0.9158277548849583, "replacement": 0.5696248710155487}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 7.298285961151123, "loss_scale": 4.0, "retries": 0, "seconds": 533.1, "max_memory_gb": 23.0, "pair": 0.1356164226308465, "sign": 0.1944123045541346, "keep": 0.055010351934470236, "leak": 1.4152919370681047, "replacement": 0.6141507402062416}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 4.321837902069092, "loss_scale": 4.0, "retries": 0, "seconds": 534.2, "max_memory_gb": 23.0, "pair": 0.14826205372810364, "sign": 0.12495957687497139, "keep": 0.03404615097679198, "leak": 1.2621261812746525, "replacement": 0.5432964563369751}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 2.8910164833068848, "loss_scale": 4.0, "retries": 0, "seconds": 535.4, "max_memory_gb": 23.0, "pair": 0.13763449154794216, "sign": 0.1222767410799861, "keep": 0.012696495046839118, "leak": 0.2592089157551527, "replacement": 0.2699233368039131}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 4.809567451477051, "loss_scale": 4.0, "retries": 0, "seconds": 536.5, "max_memory_gb": 23.0, "pair": 0.37741027027368546, "sign": 0.1813191045075655, "keep": 0.2869696521665901, "leak": 1.0195211879909039, "replacement": 0.8664349317550659}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 3.0004751682281494, "loss_scale": 4.0, "retries": 0, "seconds": 537.6, "max_memory_gb": 23.0, "pair": 0.1503158826380968, "sign": 0.07159747276455164, "keep": 0.02066847390960902, "leak": 0.3115724688395858, "replacement": 0.27434196323156357}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 14.238829612731934, "loss_scale": 4.0, "retries": 0, "seconds": 538.8, "max_memory_gb": 23.0, "pair": 0.382422374561429, "sign": 0.5482746213674545, "keep": 0.566926003433764, "leak": 0.5125942691229284, "replacement": 1.0681712999939919}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 9.216059684753418, "loss_scale": 4.0, "retries": 0, "seconds": 539.9, "max_memory_gb": 23.0, "pair": 0.3573004687204957, "sign": 0.4188614785671234, "keep": 0.08262705337256193, "leak": 0.84532181173563, "replacement": 0.8193751722574234}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 11.956435203552246, "loss_scale": 4.0, "retries": 0, "seconds": 541.0, "max_memory_gb": 23.0, "pair": 0.18695298209786415, "sign": 0.12078734394162893, "keep": 0.5349955979036167, "leak": 0.3264910327270627, "replacement": 0.5964672230184078}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 7.888082504272461, "loss_scale": 4.0, "retries": 0, "seconds": 542.2, "max_memory_gb": 23.0, "pair": 0.21729260869324207, "sign": 0.38301235251128674, "keep": 0.3988337749033235, "leak": 0.4244907721877098, "replacement": 0.7143384031951427}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 15.619704246520996, "loss_scale": 4.0, "retries": 0, "seconds": 543.3, "max_memory_gb": 23.0, "pair": 0.2952731288969517, "sign": 0.4331237096339464, "keep": 0.47267973906127736, "leak": 0.46367829013615847, "replacement": 0.864094402641058}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 3.228127956390381, "loss_scale": 4.0, "retries": 0, "seconds": 544.5, "max_memory_gb": 23.0, "pair": 0.1819655280560255, "sign": 0.17671320587396622, "keep": 0.053215728141367435, "leak": 1.9498650133609772, "replacement": 0.7843962423503399}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 4.802563190460205, "loss_scale": 4.0, "retries": 0, "seconds": 545.6, "max_memory_gb": 23.0, "pair": 0.2999641550704837, "sign": 0.38287599803879857, "keep": 0.02891420974629, "leak": 0.4725881586782634, "replacement": 0.6240062713623047}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 5.8735575675964355, "loss_scale": 4.0, "retries": 0, "seconds": 546.7, "max_memory_gb": 23.0, "pair": 0.24980241060256958, "sign": 0.22811949346214533, "keep": 0.018604950397275388, "leak": 0.3745794426649809, "replacement": 0.4668094962835312}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 7.472784042358398, "loss_scale": 4.0, "retries": 0, "seconds": 547.9, "max_memory_gb": 23.0, "pair": 0.5380663182586432, "sign": 0.4836301989853382, "keep": 0.4075475868303329, "leak": 0.5566977527923882, "replacement": 1.1228296607732773}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 4.853240013122559, "loss_scale": 4.0, "retries": 0, "seconds": 549.0, "max_memory_gb": 23.0, "pair": 0.3425713600590825, "sign": 0.4864570517092943, "keep": 0.3192128352820873, "leak": 0.6893755747005343, "replacement": 0.9177502244710922}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 4.64290714263916, "loss_scale": 4.0, "retries": 0, "seconds": 550.1, "max_memory_gb": 23.0, "pair": 0.1766966488212347, "sign": 0.12466795602813363, "keep": 0.11531310516875237, "leak": 0.41273110802285373, "replacement": 0.39986997470259666}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 10.251178741455078, "loss_scale": 4.0, "retries": 0, "seconds": 551.3, "max_memory_gb": 23.0, "pair": 0.18004696816205978, "sign": 0.405706238001585, "keep": 0.3322466191602871, "leak": 0.6420349106192589, "replacement": 0.7095321416854858}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 16.465579986572266, "loss_scale": 4.0, "retries": 0, "seconds": 552.4, "max_memory_gb": 23.0, "pair": 0.40317995846271515, "sign": 0.7516750693321228, "keep": 0.54077791236341, "leak": 0.30380372892250307, "replacement": 1.125357374548912}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 20.498674392700195, "loss_scale": 4.0, "retries": 0, "seconds": 553.5, "max_memory_gb": 23.0, "pair": 0.6000288277864456, "sign": 0.9551293104887009, "keep": 1.4662990681827068, "leak": 0.4822354782372713, "replacement": 1.931301899254322}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 12.069124221801758, "loss_scale": 4.0, "retries": 0, "seconds": 554.7, "max_memory_gb": 23.0, "pair": 0.25048605166375637, "sign": 0.47898315265774727, "keep": 0.018868939601816237, "leak": 0.8483514003455639, "replacement": 0.7114999331533909}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 13.095220565795898, "loss_scale": 4.0, "retries": 0, "seconds": 555.8, "max_memory_gb": 23.0, "pair": 0.39023522567003965, "sign": 0.6983766695484519, "keep": 0.9898594813421369, "leak": 0.5806360170245171, "replacement": 1.379512295126915}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 3.218076229095459, "loss_scale": 4.0, "retries": 0, "seconds": 556.9, "max_memory_gb": 23.0, "pair": 0.19400823675096035, "sign": 0.15445064846426249, "keep": 0.08450078725581989, "leak": 0.305062094819732, "replacement": 0.3897494897246361}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 3.906247138977051, "loss_scale": 4.0, "retries": 0, "seconds": 558.0, "max_memory_gb": 23.0, "pair": 0.2640668135136366, "sign": 0.3013142617419362, "keep": 0.10011699236929417, "leak": 1.2893299236893654, "replacement": 0.7871149182319641}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 6.490664958953857, "loss_scale": 4.0, "retries": 0, "seconds": 559.2, "max_memory_gb": 23.0, "pair": 0.11858088057488203, "sign": 0.22263389267027378, "keep": 0.5155835807672702, "leak": 1.4721725306590088, "replacement": 0.8557327464222908}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 6.626985549926758, "loss_scale": 4.0, "retries": 0, "seconds": 560.3, "max_memory_gb": 23.0, "pair": 0.1637536846101284, "sign": 0.21110396925359964, "keep": 0.015927694970741868, "leak": 0.7757510719820857, "replacement": 0.47120729088783264}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 11.273354530334473, "loss_scale": 4.0, "retries": 0, "seconds": 561.4, "max_memory_gb": 23.0, "pair": 0.15770392213016748, "sign": 0.32511821761727333, "keep": 0.18314320559147745, "leak": 1.6018842570483685, "replacement": 0.8123056776821613}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 3.93888783454895, "loss_scale": 4.0, "retries": 0, "seconds": 562.5, "max_memory_gb": 23.0, "pair": 0.14972797594964504, "sign": 0.11688959784805775, "keep": 0.013872249866835773, "leak": 1.069663468748331, "replacement": 0.482524748891592}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 13.397228240966797, "loss_scale": 4.0, "retries": 0, "seconds": 563.7, "max_memory_gb": 23.0, "pair": 0.5055512813851237, "sign": 0.8536021858453751, "keep": 0.6059646662324667, "leak": 1.2769692987203598, "replacement": 1.554577048867941}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 6.11439323425293, "loss_scale": 4.0, "retries": 0, "seconds": 564.9, "max_memory_gb": 23.0, "pair": 0.05347114196047187, "sign": 0.15526121738366783, "keep": 0.013306168373674154, "leak": 1.301843322813511, "replacement": 0.463215671479702}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 5.276312828063965, "loss_scale": 4.0, "retries": 0, "seconds": 566.0, "max_memory_gb": 23.0, "pair": 0.17000766657292843, "sign": 0.23591319378465414, "keep": 0.025909476447850466, "leak": 0.6824948647990823, "replacement": 0.47154271602630615}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 2.12408185005188, "loss_scale": 4.0, "retries": 0, "seconds": 567.1, "max_memory_gb": 23.0, "pair": 0.1545578520745039, "sign": 0.06743587926030159, "keep": 0.08322497992776334, "leak": 0.7634538356214762, "replacement": 0.42075175046920776}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 19.020971298217773, "loss_scale": 4.0, "retries": 0, "seconds": 568.2, "max_memory_gb": 23.0, "pair": 0.3789949957281351, "sign": 0.6591484490782022, "keep": 1.103844658297021, "leak": 1.1597805290366523, "replacement": 1.5504366494715214}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 6.069380283355713, "loss_scale": 4.0, "retries": 0, "seconds": 569.3, "max_memory_gb": 23.0, "pair": 0.43576787784695625, "sign": 0.3862926624715328, "keep": 0.25004867336247116, "leak": 0.5447945520281792, "replacement": 0.8901371881365776}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 19.57181167602539, "loss_scale": 4.0, "retries": 0, "seconds": 570.5, "max_memory_gb": 23.0, "pair": 0.41861368622630835, "sign": 0.4193338621407747, "keep": 0.5804609679616988, "leak": 0.8380692172795534, "replacement": 1.1280284561216831}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 19.63399887084961, "loss_scale": 4.0, "retries": 0, "seconds": 571.6, "max_memory_gb": 23.0, "pair": 0.37341526709496975, "sign": 0.5925013110972941, "keep": 0.9676128412829712, "leak": 0.9992489842697978, "replacement": 1.403284527361393}
diff --git a/runs/single_llava/replacement/L18_s13/adapter/README.md b/runs/single_llava/replacement/L18_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L18_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L18_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L18_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..aa3e0dba81265d79f8d6f7188481483cc40912bc
--- /dev/null
+++ b/runs/single_llava/replacement/L18_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.18.mlp.down_proj",
+    "model.language_model.layers.18.self_attn.o_proj",
+    "model.language_model.layers.18.self_attn.k_proj",
+    "model.language_model.layers.18.mlp.up_proj",
+    "model.language_model.layers.18.self_attn.q_proj",
+    "model.language_model.layers.18.self_attn.v_proj",
+    "model.language_model.layers.18.mlp.gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L18_s13/config.json b/runs/single_llava/replacement/L18_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d65e73295009efc35f2d2a6ee54bbbbb4e35acc3
--- /dev/null
+++ b/runs/single_llava/replacement/L18_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 18,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L18_s13/steps.jsonl b/runs/single_llava/replacement/L18_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ac1cdd5c8ea04fdca69005ff6a9a49b991ab9569
--- /dev/null
+++ b/runs/single_llava/replacement/L18_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.36591821908950806, "loss_scale": 4.0, "retries": 0, "seconds": 1.3, "max_memory_gb": 21.21, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.5570325255393982, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 21.21, "pair": 1.0973104536533356, "sign": 0.5360876135528088, "keep": 5.373880201717839, "leak": 2.6517056932462424e-06, "replacement": 4.052294999361038}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.0046736001968384, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 21.26, "pair": 1.0999770164489746, "sign": 0.5425914963707328, "keep": 10.804240345954895, "leak": 3.319303095850046e-05, "replacement": 6.773401260375977}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.743478536605835, "loss_scale": 4.0, "retries": 0, "seconds": 4.8, "max_memory_gb": 21.27, "pair": 1.0952350199222565, "sign": 0.47192843817174435, "keep": 7.669976407662034, "leak": 1.9890569433300698e-05, "replacement": 5.166192442178726}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.34735623002052307, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 21.27, "pair": 1.0976077616214752, "sign": 0.7765554785728455, "keep": 0.15918988827615976, "leak": 6.785516966312244e-06, "replacement": 1.5654821395874023}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.46116209030151367, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 21.27, "pair": 1.097279131412506, "sign": 0.7375741004943848, "keep": 2.932844676077366, "leak": 2.9402655041366188e-05, "replacement": 2.9324958622455597}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.7778660655021667, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 21.27, "pair": 1.086542010307312, "sign": 0.5459886370226741, "keep": 6.8791568940505385, "leak": 0.00010694470893213293, "replacement": 4.799141466617584}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.7848952412605286, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 21.27, "pair": 1.0916440784931183, "sign": 0.5884310454130173, "keep": 7.384826809167862, "leak": 0.00011680656687929059, "replacement": 5.078302323818207}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.8146005272865295, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 21.27, "pair": 1.0863405764102936, "sign": 0.6187678896822035, "keep": 5.7085446789860725, "leak": 0.0001981015916427964, "replacement": 4.250046223402023}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.5436443090438843, "loss_scale": 4.0, "retries": 0, "seconds": 11.7, "max_memory_gb": 21.27, "pair": 1.087419718503952, "sign": 0.7808910235762596, "keep": 3.2712946087121964, "leak": 0.0002557582813551562, "replacement": 3.113576352596283}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.8942914009094238, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 21.27, "pair": 1.077370673418045, "sign": 0.45323605090379715, "keep": 5.263956416398287, "leak": 0.0005002097017552387, "replacement": 3.936091959476471}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.6872444748878479, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 21.27, "pair": 1.080754280090332, "sign": 0.6103026680648327, "keep": 3.0561957308091223, "leak": 0.0004953600407588965, "replacement": 2.9141274094581604}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.8184476494789124, "loss_scale": 4.0, "retries": 0, "seconds": 15.1, "max_memory_gb": 21.27, "pair": 1.0841697454452515, "sign": 0.510081522166729, "keep": 5.34625736810267, "leak": 0.0003358336682595109, "replacement": 4.012423008680344}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.2301100492477417, "loss_scale": 4.0, "retries": 0, "seconds": 16.3, "max_memory_gb": 21.27, "pair": 1.0628725588321686, "sign": 0.39637948758900166, "keep": 7.786845497786999, "leak": 0.000702540457041323, "replacement": 5.154660761356354}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.781815767288208, "loss_scale": 4.0, "retries": 0, "seconds": 17.4, "max_memory_gb": 21.27, "pair": 1.0748258531093597, "sign": 0.6893262639641762, "keep": 5.178817437961698, "leak": 0.0007607731586176669, "replacement": 4.009088039398193}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.5959939956665039, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 21.27, "pair": 1.0833075046539307, "sign": 0.782980740070343, "keep": 2.762367984163575, "leak": 0.0002740660693234531, "replacement": 2.856050342321396}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.5865181088447571, "loss_scale": 4.0, "retries": 0, "seconds": 19.8, "max_memory_gb": 21.27, "pair": 1.0819317698478699, "sign": 0.7134667858481407, "keep": 3.1871830821037292, "leak": 0.000496619691602973, "replacement": 3.0323808789253235}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.1173585653305054, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 21.32, "pair": 1.0677302181720734, "sign": 0.5369434170424938, "keep": 7.179353542625904, "leak": 0.0005591593435383402, "replacement": 4.926018416881561}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.9055582284927368, "loss_scale": 4.0, "retries": 0, "seconds": 22.1, "max_memory_gb": 21.32, "pair": 1.0326159596443176, "sign": 0.533455390483141, "keep": 10.67594861984253, "leak": 0.0040384414460277185, "replacement": 6.638327598571777}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.5610705614089966, "loss_scale": 4.0, "retries": 0, "seconds": 23.3, "max_memory_gb": 21.32, "pair": 1.08297398686409, "sign": 0.42946430295705795, "keep": 0.02363916952162981, "leak": 0.00034310363116674125, "replacement": 1.3096115291118622}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.263068437576294, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 21.32, "pair": 1.0438079684972763, "sign": 0.3707630727440119, "keep": 5.524142424110323, "leak": 0.0009407729230588302, "replacement": 3.9914959967136383}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.1554274559020996, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 21.32, "pair": 1.0517860651016235, "sign": 0.5709821358323097, "keep": 5.911328695714474, "leak": 0.004297150648199022, "replacement": 4.294015645980835}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.561429500579834, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 21.32, "pair": 1.0114182233810425, "sign": 0.2476424854248762, "keep": 5.283466280438006, "leak": 0.00751996117469389, "replacement": 3.7788526713848114}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.081239700317383, "loss_scale": 4.0, "retries": 0, "seconds": 27.9, "max_memory_gb": 21.32, "pair": 1.0004873424768448, "sign": 0.2565777897834778, "keep": 10.007391691207886, "leak": 0.003740733220183756, "replacement": 6.133407354354858}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.710402011871338, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 21.32, "pair": 1.0212234109640121, "sign": 0.3907710090279579, "keep": 8.441447407007217, "leak": 0.007592432725641141, "replacement": 5.439230680465698}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.467336654663086, "loss_scale": 4.0, "retries": 0, "seconds": 30.3, "max_memory_gb": 21.32, "pair": 0.968059629201889, "sign": 0.6546407360583544, "keep": 10.46170163154602, "leak": 0.014190090936608613, "replacement": 6.5297781229019165}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.8668678998947144, "loss_scale": 4.0, "retries": 0, "seconds": 31.4, "max_memory_gb": 21.32, "pair": 1.0160257667303085, "sign": 0.5893559232354164, "keep": 8.368286684155464, "leak": 0.008842279567033984, "replacement": 5.49705770611763}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.5336909294128418, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 21.32, "pair": 1.0230382084846497, "sign": 0.7710621617734432, "keep": 5.592647284269333, "leak": 0.009877346537905396, "replacement": 4.207362353801727}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.4409111738204956, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 21.32, "pair": 1.0339285135269165, "sign": 0.8332006819546223, "keep": 5.180980822071433, "leak": 0.008464673322578165, "replacement": 4.04313537478447}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.6876565217971802, "loss_scale": 4.0, "retries": 0, "seconds": 34.9, "max_memory_gb": 21.32, "pair": 0.9651165753602982, "sign": 0.3429438639432192, "keep": 5.492359863594174, "leak": 0.02116702162538786, "replacement": 3.8880601823329926}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.582924485206604, "loss_scale": 4.0, "retries": 0, "seconds": 36.1, "max_memory_gb": 21.32, "pair": 1.004540592432022, "sign": 0.7979183606803417, "keep": 5.652460724115372, "leak": 0.013442890920487116, "replacement": 4.233090877532959}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.7645909786224365, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 21.32, "pair": 0.9613071233034134, "sign": 0.3757696859538555, "keep": 10.644126653671265, "leak": 0.020835565635934472, "replacement": 6.4764643907547}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.7354103326797485, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 21.32, "pair": 0.9796448647975922, "sign": 0.6127785556018353, "keep": 4.850652860477567, "leak": 0.02259964063046027, "replacement": 3.7170103788375854}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.8024810552597046, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 21.32, "pair": 0.9845154136419296, "sign": 0.5190275944769382, "keep": 5.243349361233413, "leak": 0.009420140188012738, "replacement": 3.868058830499649}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.9899952411651611, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 21.32, "pair": 0.9275543987751007, "sign": 0.4248184412717819, "keep": 5.419522721320391, "leak": 0.038829670898849145, "replacement": 3.859432339668274}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.8433749675750732, "loss_scale": 4.0, "retries": 0, "seconds": 41.9, "max_memory_gb": 21.32, "pair": 0.9378526955842972, "sign": 0.42820954136550426, "keep": 5.121634319424629, "leak": 0.029289436737599317, "replacement": 3.720096915960312}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 2.014324903488159, "loss_scale": 4.0, "retries": 0, "seconds": 43.1, "max_memory_gb": 21.32, "pair": 0.9352527111768723, "sign": 0.575155658647418, "keep": 4.889978691935539, "leak": 0.03957791747961892, "replacement": 3.6777142882347107}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.2605652809143066, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 21.32, "pair": 0.8952323645353317, "sign": 0.746930219233036, "keep": 9.342793583869934, "leak": 0.04657333021168597, "replacement": 5.951737761497498}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.3941401243209839, "loss_scale": 4.0, "retries": 0, "seconds": 45.4, "max_memory_gb": 21.32, "pair": 0.9929404407739639, "sign": 0.5509565100073814, "keep": 2.8488959930837154, "leak": 0.04028141006710939, "replacement": 2.702937036752701}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.047220230102539, "loss_scale": 4.0, "retries": 0, "seconds": 46.5, "max_memory_gb": 21.32, "pair": 0.9516309350728989, "sign": 0.48617877438664436, "keep": 5.444241508841515, "leak": 0.020848713349550962, "replacement": 3.922053426504135}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.984968364238739, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 21.32, "pair": 1.0532410442829132, "sign": 1.0156498849391937, "keep": 2.9151522517204285, "leak": 0.011653299598284983, "replacement": 3.02155539393425}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 2.245460033416748, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 21.32, "pair": 0.9300114065408707, "sign": 0.3898953143507242, "keep": 4.61054105963558, "leak": 0.025171837187372148, "replacement": 3.4365225434303284}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.3564953804016113, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 21.32, "pair": 0.9133584201335907, "sign": 0.4962148889899254, "keep": 4.782179156318307, "leak": 0.030220064567402005, "replacement": 3.5601103603839874}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.41097891330719, "loss_scale": 4.0, "retries": 0, "seconds": 51.2, "max_memory_gb": 21.32, "pair": 0.9580659717321396, "sign": 0.563018873333931, "keep": 2.712514874525368, "leak": 0.03702934563625604, "replacement": 2.6050902009010315}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.3163580894470215, "loss_scale": 4.0, "retries": 0, "seconds": 52.3, "max_memory_gb": 21.32, "pair": 1.0052484720945358, "sign": 0.8728579729795456, "keep": 2.766626089811325, "leak": 0.015248239867844404, "replacement": 2.828802615404129}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.4700639247894287, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 21.32, "pair": 0.9025015383958817, "sign": 0.7018936797976494, "keep": 4.434621721506119, "leak": 0.08167793213215191, "replacement": 3.4911787509918213}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.2793896198272705, "loss_scale": 4.0, "retries": 0, "seconds": 54.6, "max_memory_gb": 21.32, "pair": 0.8661472499370575, "sign": 0.6502421870827675, "keep": 4.0201156437397, "leak": 0.060676514556689654, "replacement": 3.2164953649044037}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 4.09048318862915, "loss_scale": 4.0, "retries": 0, "seconds": 55.8, "max_memory_gb": 21.32, "pair": 0.7073421701788902, "sign": 0.18503294605761766, "keep": 7.260385259985924, "leak": 0.20328400086145848, "replacement": 4.480872243642807}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.453568696975708, "loss_scale": 4.0, "retries": 0, "seconds": 56.9, "max_memory_gb": 21.32, "pair": 0.8034007847309113, "sign": 0.595530197955668, "keep": 6.937713786959648, "leak": 0.10390661330893636, "replacement": 4.595999211072922}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 3.7567245960235596, "loss_scale": 4.0, "retries": 0, "seconds": 58.0, "max_memory_gb": 21.32, "pair": 0.868695393204689, "sign": 0.4178761914372444, "keep": 6.985028645023704, "leak": 0.0679553805384785, "replacement": 4.587136626243591}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.5860888957977295, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 21.32, "pair": 0.9120455831289291, "sign": 0.56455341540277, "keep": 2.6275360928848386, "leak": 0.06410788733046502, "replacement": 2.5241172313690186}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 2.7414848804473877, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 21.32, "pair": 0.8153909221291542, "sign": 0.6776952166110277, "keep": 4.555210418999195, "leak": 0.15997721598250791, "replacement": 3.4718379974365234}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 3.9696261882781982, "loss_scale": 4.0, "retries": 0, "seconds": 61.5, "max_memory_gb": 21.32, "pair": 0.8462524563074112, "sign": 0.7102557122707367, "keep": 7.126434832811356, "leak": 0.1252166583071812, "replacement": 4.795902073383331}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 4.282188892364502, "loss_scale": 4.0, "retries": 0, "seconds": 62.6, "max_memory_gb": 21.32, "pair": 0.7644398957490921, "sign": 0.5901565328240395, "keep": 7.422140866518021, "leak": 0.19879815477179363, "replacement": 4.820288062095642}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 3.2164416313171387, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 21.32, "pair": 0.8121031075716019, "sign": 0.4083660636097193, "keep": 4.579173954203725, "leak": 0.1334398086182773, "replacement": 3.3392330706119537}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 6.391759872436523, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 21.32, "pair": 0.5295099392533302, "sign": 0.14382232166826725, "keep": 8.824753046035767, "leak": 0.23257577419281006, "replacement": 5.071941614151001}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.6703318357467651, "loss_scale": 4.0, "retries": 0, "seconds": 66.0, "max_memory_gb": 21.32, "pair": 0.9275243282318115, "sign": 0.7808798775076866, "keep": 2.3314959835261106, "leak": 0.024959793663583696, "replacement": 2.4899522066116333}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 3.4474430084228516, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 21.32, "pair": 0.8158795982599258, "sign": 0.26177311688661575, "keep": 4.410482282284647, "leak": 0.21343329013325274, "replacement": 3.2053657472133636}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 2.3310956954956055, "loss_scale": 4.0, "retries": 0, "seconds": 68.3, "max_memory_gb": 21.32, "pair": 0.932033509016037, "sign": 0.6416916400194168, "keep": 2.64037474244833, "leak": 0.015279500612791708, "replacement": 2.576886534690857}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 3.328432559967041, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 21.32, "pair": 0.748863197863102, "sign": 0.6052176505327225, "keep": 4.415422143414617, "leak": 0.14915044594090432, "replacement": 3.296470820903778}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 5.640497207641602, "loss_scale": 4.0, "retries": 0, "seconds": 70.6, "max_memory_gb": 21.32, "pair": 0.6980657055974007, "sign": 0.6204422786831856, "keep": 6.1099593341350555, "leak": 0.3046013159237191, "replacement": 4.139416873455048}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 2.053359270095825, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 21.32, "pair": 0.8813994377851486, "sign": 0.6936121843755245, "keep": 2.47478093765676, "leak": 0.09378179027407896, "replacement": 2.489041566848755}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 5.169939041137695, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 21.32, "pair": 0.8051373958587646, "sign": 0.5544237345457077, "keep": 6.781832452863455, "leak": 0.06308470806197874, "replacement": 4.489036589860916}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 6.625250816345215, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 21.36, "pair": 0.6109752506017685, "sign": 0.4057486639358103, "keep": 6.096435327082872, "leak": 0.48350386694073677, "replacement": 3.982943296432495}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 6.158146381378174, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 21.36, "pair": 0.7095083594322205, "sign": 0.5152756879106164, "keep": 6.423332661390305, "leak": 0.28596575566916727, "replacement": 4.250304043292999}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.833949089050293, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 21.36, "pair": 0.9442636668682098, "sign": 0.5870162174105644, "keep": 1.6998051764676347, "leak": 0.0172884521598462, "replacement": 2.091996520757675}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 6.76207971572876, "loss_scale": 4.0, "retries": 0, "seconds": 77.6, "max_memory_gb": 21.36, "pair": 0.5501222461462021, "sign": 0.35728594847023487, "keep": 5.608809739351273, "leak": 0.6417371541319881, "replacement": 3.6936044692993164}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 2.6621334552764893, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 21.36, "pair": 0.837156742811203, "sign": 0.4793335944414139, "keep": 2.0128899491392076, "leak": 0.20193490821353066, "replacement": 2.1337522864341736}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 5.220106601715088, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 21.36, "pair": 0.732167087495327, "sign": 0.7175742033869028, "keep": 3.647991945501417, "leak": 0.17713382256260957, "replacement": 2.9592336416244507}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 5.160741806030273, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 21.36, "pair": 0.7826176434755325, "sign": 0.5653740242123604, "keep": 3.584667179733515, "leak": 0.3176873563497793, "replacement": 2.9370600283145905}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 7.886537075042725, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 21.36, "pair": 0.49799108505249023, "sign": 0.14438248425722122, "keep": 5.428203010931611, "leak": 0.41849971236661077, "replacement": 3.388908863067627}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 10.48335075378418, "loss_scale": 4.0, "retries": 0, "seconds": 83.5, "max_memory_gb": 21.36, "pair": 0.4295591413974762, "sign": 0.22607142850756645, "keep": 5.983976125717163, "leak": 0.5246916264295578, "replacement": 3.665755808353424}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 1.1154721975326538, "loss_scale": 4.0, "retries": 0, "seconds": 84.6, "max_memory_gb": 21.36, "pair": 0.9895046502351761, "sign": 0.7897252440452576, "keep": 0.42424519546329975, "leak": 0.012410211264068494, "replacement": 1.5995924472808838}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 6.417019367218018, "loss_scale": 4.0, "retries": 0, "seconds": 85.7, "max_memory_gb": 21.36, "pair": 0.6985232904553413, "sign": 0.4791431911289692, "keep": 3.233498797751963, "leak": 0.2725616168754641, "replacement": 2.622984677553177}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 3.408291816711426, "loss_scale": 4.0, "retries": 0, "seconds": 86.9, "max_memory_gb": 21.36, "pair": 0.8459671884775162, "sign": 0.888239486142993, "keep": 2.6678026914596558, "leak": 0.3732100320958125, "replacement": 2.717290759086609}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 6.300525188446045, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 21.36, "pair": 0.607019443064928, "sign": 0.41075531719252467, "keep": 2.8550694156438112, "leak": 0.5795486117713153, "replacement": 2.3848189413547516}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 8.783438682556152, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 21.36, "pair": 0.5828931480646133, "sign": 0.43969732988625765, "keep": 4.645395398139954, "leak": 1.1906375406542793, "replacement": 3.423098862171173}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 5.57225227355957, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 21.36, "pair": 0.7720228135585785, "sign": 0.6915452256798744, "keep": 2.4980108439922333, "leak": 0.3395680414978415, "replacement": 2.451692759990692}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 3.628242015838623, "loss_scale": 4.0, "retries": 0, "seconds": 91.5, "max_memory_gb": 21.36, "pair": 0.8236952871084213, "sign": 0.559494286775589, "keep": 1.1754099363461137, "leak": 0.16100655638729222, "replacement": 1.7313989698886871}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 6.516407012939453, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 21.36, "pair": 0.6177829932421446, "sign": 0.3550743944942951, "keep": 2.079523410415277, "leak": 1.2057823636569083, "replacement": 2.136527508497238}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 10.040830612182617, "loss_scale": 4.0, "retries": 0, "seconds": 93.7, "max_memory_gb": 21.36, "pair": 0.4863131381571293, "sign": 0.4228086266666651, "keep": 3.1618825867772102, "leak": 1.1771557736356044, "replacement": 2.5729476511478424}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 11.431991577148438, "loss_scale": 4.0, "retries": 0, "seconds": 94.8, "max_memory_gb": 21.36, "pair": 0.4361913166940212, "sign": 0.3351651565171778, "keep": 2.8944908380508423, "leak": 0.95829187866957, "replacement": 2.290592283010483}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 9.719754219055176, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 21.36, "pair": 0.5385950654745102, "sign": 0.3907702974975109, "keep": 2.6858194917440414, "leak": 0.6713415943086147, "replacement": 2.244725316762924}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 4.005076885223389, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 21.36, "pair": 0.7989648729562759, "sign": 0.7010587379336357, "keep": 1.341232286999002, "leak": 0.3593564926413819, "replacement": 1.9099496006965637}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.3445467948913574, "loss_scale": 4.0, "retries": 0, "seconds": 98.4, "max_memory_gb": 21.36, "pair": 0.9548977613449097, "sign": 0.5684325024485588, "keep": 0.13598715630359948, "leak": 0.03498648607637733, "replacement": 1.315854161977768}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 11.511700630187988, "loss_scale": 4.0, "retries": 0, "seconds": 99.5, "max_memory_gb": 21.36, "pair": 0.5422037914395332, "sign": 0.6614760980010033, "keep": 2.57285438477993, "leak": 0.7170341598830419, "replacement": 2.3386275470256805}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 9.706843376159668, "loss_scale": 4.0, "retries": 0, "seconds": 100.6, "max_memory_gb": 21.36, "pair": 0.49482154101133347, "sign": 0.33942549861967564, "keep": 1.9708825126290321, "leak": 2.018122513865819, "replacement": 2.1545061469078064}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 6.705821514129639, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 21.36, "pair": 0.6632489021867514, "sign": 0.6170543706975877, "keep": 1.1363594608847052, "leak": 0.7241874673254642, "replacement": 1.7210026979446411}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 1.2248501777648926, "loss_scale": 4.0, "retries": 0, "seconds": 103.0, "max_memory_gb": 21.36, "pair": 0.9542926251888275, "sign": 0.6912580356001854, "keep": 0.2092338033253327, "leak": 0.04456681138080398, "replacement": 1.4156802594661713}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 6.415653228759766, "loss_scale": 4.0, "retries": 0, "seconds": 104.1, "max_memory_gb": 21.36, "pair": 0.6273121014237404, "sign": 0.5124711133539677, "keep": 1.0422475580126047, "leak": 0.600090823485516, "replacement": 1.55469411611557}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 1.1129978895187378, "loss_scale": 4.0, "retries": 0, "seconds": 105.3, "max_memory_gb": 21.36, "pair": 1.003085434436798, "sign": 0.9976415485143661, "keep": 0.6132427752017975, "leak": 0.01856494846026635, "replacement": 1.8131688833236694}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.1310224533081055, "loss_scale": 4.0, "retries": 0, "seconds": 106.5, "max_memory_gb": 21.36, "pair": 0.8265776932239532, "sign": 0.5203371942043304, "keep": 0.4441849134163931, "leak": 1.2993116360157728, "replacement": 1.6336666941642761}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 6.201544284820557, "loss_scale": 4.0, "retries": 0, "seconds": 107.6, "max_memory_gb": 21.36, "pair": 0.524437639862299, "sign": 0.27425264939665794, "keep": 0.8122029221849516, "leak": 0.6232351437211037, "replacement": 1.2234741747379303}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 10.746099472045898, "loss_scale": 4.0, "retries": 0, "seconds": 108.7, "max_memory_gb": 21.36, "pair": 0.3939102366566658, "sign": 0.33369186241179705, "keep": 1.250516414642334, "leak": 1.3630838539047545, "replacement": 1.5267853140830994}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 6.110610485076904, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 21.36, "pair": 0.5455380976200104, "sign": 0.44290607143193483, "keep": 0.8575320402160287, "leak": 0.7549196600048163, "replacement": 1.3844870924949646}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 5.427114963531494, "loss_scale": 4.0, "retries": 0, "seconds": 111.0, "max_memory_gb": 21.36, "pair": 0.39341922476887703, "sign": 0.3760747262276709, "keep": 0.7294159308075905, "leak": 3.2695175454391574, "replacement": 1.763543963432312}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.8463189601898193, "loss_scale": 4.0, "retries": 0, "seconds": 112.1, "max_memory_gb": 21.36, "pair": 0.7003207094967365, "sign": 0.4813183583319187, "keep": 0.5013934634625912, "leak": 0.9386556814424694, "replacement": 1.4263405203819275}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 3.7720978260040283, "loss_scale": 4.0, "retries": 0, "seconds": 113.2, "max_memory_gb": 21.36, "pair": 0.6148672960698605, "sign": 0.5035024145618081, "keep": 0.2999073900282383, "leak": 1.2547890283167362, "replacement": 1.330269455909729}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 1.2034598588943481, "loss_scale": 4.0, "retries": 0, "seconds": 114.4, "max_memory_gb": 21.36, "pair": 0.9316947907209396, "sign": 0.8582072034478188, "keep": 0.3917114398209378, "leak": 0.03428321797400713, "replacement": 1.5652248710393906}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 5.5883049964904785, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 21.36, "pair": 0.4821889102458954, "sign": 0.5542167178355157, "keep": 0.6880734339356422, "leak": 1.9716158988885581, "replacement": 1.596237912774086}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 5.674121379852295, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 21.36, "pair": 0.2754586338996887, "sign": 0.1420462664682418, "keep": 0.4875806048512459, "leak": 1.1945643108338118, "replacement": 0.8889131247997284}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 4.880837917327881, "loss_scale": 4.0, "retries": 0, "seconds": 117.9, "max_memory_gb": 21.36, "pair": 0.40920254588127136, "sign": 0.29228367283940315, "keep": 0.8078105747699738, "leak": 1.4882218316197395, "replacement": 1.3313051164150238}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 2.371293067932129, "loss_scale": 4.0, "retries": 0, "seconds": 119.0, "max_memory_gb": 21.36, "pair": 0.51258809491992, "sign": 0.30719093978405, "keep": 0.1310413251630962, "leak": 1.5947116129100323, "replacement": 1.1303821504116058}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 3.5031673908233643, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 21.36, "pair": 0.8036036491394043, "sign": 0.9151706695556641, "keep": 0.11041153792757541, "leak": 0.30828768527135253, "replacement": 1.3934666067361832}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 3.6788113117218018, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 21.36, "pair": 0.42419271171092987, "sign": 0.33298152312636375, "keep": 0.3301308713853359, "leak": 2.630381916416809, "replacement": 1.4133443981409073}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 2.874562978744507, "loss_scale": 4.0, "retries": 0, "seconds": 122.4, "max_memory_gb": 21.36, "pair": 0.4126745592802763, "sign": 0.414200640283525, "keep": 0.23868327215313911, "leak": 2.8042306653951528, "replacement": 1.4401741921901703}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.3901803493499756, "loss_scale": 4.0, "retries": 0, "seconds": 123.6, "max_memory_gb": 21.36, "pair": 0.7631209194660187, "sign": 0.7437601685523987, "keep": 0.24953171983361244, "leak": 0.30655226577073336, "replacement": 1.3364049047231674}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.2348134517669678, "loss_scale": 4.0, "retries": 0, "seconds": 124.7, "max_memory_gb": 21.36, "pair": 0.7198366560041904, "sign": 0.6652749571949244, "keep": 0.221906672231853, "leak": 2.0659336952958256, "replacement": 1.6799108386039734}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 4.884514331817627, "loss_scale": 4.0, "retries": 0, "seconds": 125.9, "max_memory_gb": 21.36, "pair": 0.3350957464426756, "sign": 0.38262589229270816, "keep": 0.3677759952843189, "leak": 3.3174979591567535, "replacement": 1.5396711975336075}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 2.717536687850952, "loss_scale": 4.0, "retries": 0, "seconds": 127.1, "max_memory_gb": 21.36, "pair": 0.5444591771811247, "sign": 0.4928469080477953, "keep": 0.3493375573307276, "leak": 0.9107606443722034, "replacement": 1.1932415962219238}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 1.5260854959487915, "loss_scale": 4.0, "retries": 0, "seconds": 128.2, "max_memory_gb": 21.36, "pair": 0.4573642499744892, "sign": 0.40985720558092, "keep": 0.0951304230839014, "leak": 0.3755607232451439, "replacement": 0.8037482351064682}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 2.0554630756378174, "loss_scale": 4.0, "retries": 0, "seconds": 129.4, "max_memory_gb": 21.36, "pair": 0.5549329854547977, "sign": 0.4063694179058075, "keep": 0.2058731196448207, "leak": 0.5256842086091638, "replacement": 0.992475301027298}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 2.039790153503418, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 21.36, "pair": 0.5553443990647793, "sign": 0.49418761394917965, "keep": 0.18851512670516968, "leak": 0.7381155012217278, "replacement": 1.0812246650457382}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 2.8758020401000977, "loss_scale": 4.0, "retries": 0, "seconds": 131.8, "max_memory_gb": 21.36, "pair": 0.4361247606575489, "sign": 0.4562429264187813, "keep": 0.5033976901322603, "leak": 1.8083468016702682, "replacement": 1.3680317401885986}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 2.039602279663086, "loss_scale": 4.0, "retries": 0, "seconds": 133.0, "max_memory_gb": 21.36, "pair": 0.5984187684953213, "sign": 0.549738984555006, "keep": 0.3658112185075879, "leak": 1.6858844187809154, "replacement": 1.4776650071144104}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 1.4646176099777222, "loss_scale": 4.0, "retries": 0, "seconds": 134.1, "max_memory_gb": 21.36, "pair": 0.8157469779253006, "sign": 0.6670781262218952, "keep": 0.21008910960517824, "leak": 0.12985751205997076, "replacement": 1.2867950052022934}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 1.999405860900879, "loss_scale": 4.0, "retries": 0, "seconds": 135.3, "max_memory_gb": 21.36, "pair": 0.5913870185613632, "sign": 0.42800672445446253, "keep": 0.5537608295562677, "leak": 1.1059078676626086, "replacement": 1.3587477952241898}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 1.4085042476654053, "loss_scale": 4.0, "retries": 0, "seconds": 136.5, "max_memory_gb": 21.36, "pair": 0.6312254220247269, "sign": 0.5090609260369092, "keep": 0.2680615510325879, "leak": 0.6097070974319649, "replacement": 1.1722134202718735}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 1.791616678237915, "loss_scale": 4.0, "retries": 0, "seconds": 137.6, "max_memory_gb": 21.36, "pair": 0.7654545679688454, "sign": 0.7756771743297577, "keep": 0.14124878495931625, "leak": 0.39144053324525885, "replacement": 1.3217777013778687}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 1.9671679735183716, "loss_scale": 4.0, "retries": 0, "seconds": 138.8, "max_memory_gb": 21.36, "pair": 0.5486513935029507, "sign": 0.47639007633551955, "keep": 0.49565850012004375, "leak": 0.7477430100552738, "replacement": 1.2216114699840546}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.40628981590271, "loss_scale": 4.0, "retries": 0, "seconds": 140.0, "max_memory_gb": 21.36, "pair": 0.7290479987859726, "sign": 0.6715122237801552, "keep": 0.5471500330604613, "leak": 2.2777856981847435, "replacement": 1.9078255593776703}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 3.1344621181488037, "loss_scale": 4.0, "retries": 0, "seconds": 141.2, "max_memory_gb": 21.36, "pair": 0.6210150681436062, "sign": 0.683705061674118, "keep": 0.2174840560182929, "leak": 0.747453969437629, "replacement": 1.2584731578826904}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 1.6593728065490723, "loss_scale": 4.0, "retries": 0, "seconds": 142.3, "max_memory_gb": 21.36, "pair": 0.752386674284935, "sign": 0.7733368650078773, "keep": 0.9243456795811653, "leak": 0.45186191523680463, "replacement": 1.7141934037208557}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 1.8259906768798828, "loss_scale": 4.0, "retries": 0, "seconds": 143.5, "max_memory_gb": 21.36, "pair": 0.8848616778850555, "sign": 1.0356220752000809, "keep": 0.4352005599066615, "leak": 0.07984147797105834, "replacement": 1.6402333974838257}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 1.671856164932251, "loss_scale": 4.0, "retries": 0, "seconds": 144.6, "max_memory_gb": 21.36, "pair": 0.7229780778288841, "sign": 0.7522315233945847, "keep": 0.21927518816664815, "leak": 0.3522533674258739, "replacement": 1.2967947870492935}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 1.537618637084961, "loss_scale": 4.0, "retries": 0, "seconds": 145.7, "max_memory_gb": 21.36, "pair": 0.5931283291429281, "sign": 0.601870451355353, "keep": 0.5355426277965307, "leak": 0.3897129253309686, "replacement": 1.2592630684375763}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 2.6675658226013184, "loss_scale": 4.0, "retries": 0, "seconds": 146.9, "max_memory_gb": 21.36, "pair": 0.5100920032709837, "sign": 0.3243650058284402, "keep": 0.0502651771530509, "leak": 0.757570892572403, "replacement": 0.886799767613411}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 4.248131275177002, "loss_scale": 4.0, "retries": 0, "seconds": 148.0, "max_memory_gb": 21.36, "pair": 0.4087175875902176, "sign": 0.2909521006513387, "keep": 0.19374451390467584, "leak": 1.3950846549123526, "replacement": 0.9998370409011841}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 2.466148853302002, "loss_scale": 4.0, "retries": 0, "seconds": 149.2, "max_memory_gb": 21.36, "pair": 0.2969959285110235, "sign": 0.17793998727574944, "keep": 0.06474557612091303, "leak": 1.6913252403610386, "replacement": 0.841170035302639}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 2.86344575881958, "loss_scale": 4.0, "retries": 0, "seconds": 150.3, "max_memory_gb": 21.36, "pair": 0.6408263705670834, "sign": 0.8775191642343998, "keep": 0.5890027964487672, "leak": 0.8039545798674226, "replacement": 1.5750759989023209}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 1.650554895401001, "loss_scale": 4.0, "retries": 0, "seconds": 151.5, "max_memory_gb": 21.36, "pair": 0.6428182981908321, "sign": 0.645260963588953, "keep": 0.40159691870212555, "leak": 0.26047603704500943, "replacement": 1.2313662767410278}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 2.3257246017456055, "loss_scale": 4.0, "retries": 0, "seconds": 152.6, "max_memory_gb": 21.36, "pair": 0.3381049484014511, "sign": 0.11453942488878965, "keep": 0.7732311154250056, "leak": 0.7439180314540863, "replacement": 0.9679697602987289}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 2.739452600479126, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 21.36, "pair": 0.4526584455743432, "sign": 0.3836177531629801, "keep": 0.26655290368944407, "leak": 2.2192181358568632, "replacement": 1.332548275589943}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 1.6805551052093506, "loss_scale": 4.0, "retries": 0, "seconds": 155.0, "max_memory_gb": 21.36, "pair": 0.7713108211755753, "sign": 0.7020052820444107, "keep": 0.3063171594403684, "leak": 0.2971120891161263, "replacement": 1.3497500270605087}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 2.9914138317108154, "loss_scale": 4.0, "retries": 0, "seconds": 156.1, "max_memory_gb": 21.36, "pair": 0.39155235327780247, "sign": 0.3994877561926842, "keep": 0.18359827809035778, "leak": 2.5993770539498655, "replacement": 1.332939624786377}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 3.387136936187744, "loss_scale": 4.0, "retries": 0, "seconds": 157.2, "max_memory_gb": 21.36, "pair": 0.5769056156277657, "sign": 0.4522250406444073, "keep": 0.301123658195138, "leak": 0.19653093628585339, "replacement": 1.0027127042412758}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 3.106766939163208, "loss_scale": 4.0, "retries": 0, "seconds": 158.5, "max_memory_gb": 21.36, "pair": 0.5020799860358238, "sign": 0.41435485053807497, "keep": 0.16619186894968152, "leak": 2.3389591073500924, "replacement": 1.3770931512117386}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 2.037017345428467, "loss_scale": 4.0, "retries": 0, "seconds": 159.6, "max_memory_gb": 21.36, "pair": 0.4252544306218624, "sign": 0.25293540488928556, "keep": 0.18825285509228706, "leak": 0.26444496493786573, "replacement": 0.7119598314166069}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 6.958309173583984, "loss_scale": 4.0, "retries": 0, "seconds": 160.8, "max_memory_gb": 21.36, "pair": 0.4930536262691021, "sign": 0.6568706631660461, "keep": 0.5109166465699673, "leak": 1.108172919601202, "replacement": 1.3539904803037643}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 3.4182605743408203, "loss_scale": 4.0, "retries": 0, "seconds": 162.0, "max_memory_gb": 21.36, "pair": 0.4328917786478996, "sign": 0.4523076554760337, "keep": 0.12256801594048738, "leak": 1.0855026170611382, "replacement": 0.9917052537202835}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 2.4020745754241943, "loss_scale": 4.0, "retries": 0, "seconds": 163.2, "max_memory_gb": 21.36, "pair": 0.5007730089128017, "sign": 0.5078456616029143, "keep": 0.3120567835867405, "leak": 1.3455596389248967, "replacement": 1.2471141070127487}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.1627097129821777, "loss_scale": 4.0, "retries": 0, "seconds": 164.4, "max_memory_gb": 21.36, "pair": 0.47294044494628906, "sign": 0.3784524956718087, "keep": 0.21415013109799474, "leak": 0.5984963067166973, "replacement": 0.9188658595085144}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 1.7263277769088745, "loss_scale": 4.0, "retries": 0, "seconds": 165.5, "max_memory_gb": 21.36, "pair": 0.834258571267128, "sign": 1.0522429198026657, "keep": 0.21324980148347095, "leak": 0.28533112051081844, "replacement": 1.5383377075195312}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 2.6113109588623047, "loss_scale": 4.0, "retries": 0, "seconds": 166.7, "max_memory_gb": 21.36, "pair": 0.4546755291521549, "sign": 0.3595442771911621, "keep": 0.3210917832329869, "leak": 1.3851086222566664, "replacement": 1.1412707418203354}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 3.1761865615844727, "loss_scale": 4.0, "retries": 0, "seconds": 167.9, "max_memory_gb": 21.36, "pair": 0.4048695508390665, "sign": 0.28037880454212427, "keep": 0.02675901260226965, "leak": 1.7007497400045395, "replacement": 0.9836259037256241}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.5069315433502197, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 21.36, "pair": 0.4957210049033165, "sign": 0.5249197911471128, "keep": 0.19256800319999456, "leak": 1.1660341036040336, "replacement": 1.1459734439849854}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 3.959686040878296, "loss_scale": 4.0, "retries": 0, "seconds": 170.2, "max_memory_gb": 21.36, "pair": 0.4718533046543598, "sign": 0.4783391244709492, "keep": 0.43002899980638176, "leak": 1.2766591543331742, "replacement": 1.2452021092176437}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.6392431259155273, "loss_scale": 4.0, "retries": 0, "seconds": 171.3, "max_memory_gb": 21.36, "pair": 0.46832239627838135, "sign": 0.2707038000226021, "keep": 0.02654405834618956, "leak": 0.425728265196085, "replacement": 0.7233784198760986}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 2.0263874530792236, "loss_scale": 4.0, "retries": 0, "seconds": 172.5, "max_memory_gb": 21.36, "pair": 0.5487754121422768, "sign": 0.4588413070887327, "keep": 0.153665347606875, "leak": 0.29412081465125084, "replacement": 0.9285589084029198}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 2.9712414741516113, "loss_scale": 4.0, "retries": 0, "seconds": 173.7, "max_memory_gb": 21.36, "pair": 0.44377125054597855, "sign": 0.3371661305427551, "keep": 0.3857170045375824, "leak": 2.21319483499974, "replacement": 1.3585115522146225}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 2.4665329456329346, "loss_scale": 4.0, "retries": 0, "seconds": 174.9, "max_memory_gb": 21.36, "pair": 0.3840800076723099, "sign": 0.30304037779569626, "keep": 0.2679775300202891, "leak": 0.3422272731550038, "replacement": 0.7551457732915878}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 3.0254158973693848, "loss_scale": 4.0, "retries": 0, "seconds": 176.0, "max_memory_gb": 21.36, "pair": 0.282469917088747, "sign": 0.1899471255019307, "keep": 0.11108648683875799, "leak": 2.2051807641983032, "replacement": 0.984281912446022}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 2.625014066696167, "loss_scale": 4.0, "retries": 0, "seconds": 177.2, "max_memory_gb": 21.36, "pair": 0.34359193220734596, "sign": 0.22862280905246735, "keep": 0.04410308203659952, "leak": 1.45400969684124, "replacement": 0.8434573262929916}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 3.0087032318115234, "loss_scale": 4.0, "retries": 0, "seconds": 178.4, "max_memory_gb": 21.36, "pair": 0.3726573660969734, "sign": 0.2201843960210681, "keep": 0.7065996406599879, "leak": 0.615714531391859, "replacement": 0.9899780377745628}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 3.1168816089630127, "loss_scale": 4.0, "retries": 0, "seconds": 179.6, "max_memory_gb": 21.36, "pair": 0.4493366554379463, "sign": 0.46434512734413147, "keep": 1.3230073904851452, "leak": 1.1997791193425655, "replacement": 1.642957717180252}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.8683385848999023, "loss_scale": 4.0, "retries": 0, "seconds": 180.7, "max_memory_gb": 21.36, "pair": 0.5346266403794289, "sign": 0.5025074789300561, "keep": 0.2157455492997542, "leak": 0.6667821069186175, "replacement": 1.0604487359523773}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 2.596564531326294, "loss_scale": 4.0, "retries": 0, "seconds": 181.8, "max_memory_gb": 21.36, "pair": 0.7447942271828651, "sign": 1.1412125825881958, "keep": 0.24014528386760503, "leak": 0.26444139052182436, "replacement": 1.501583531498909}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.8166985511779785, "loss_scale": 4.0, "retries": 0, "seconds": 182.9, "max_memory_gb": 21.36, "pair": 0.583662036806345, "sign": 0.6317789517343044, "keep": 0.41743019968271255, "leak": 0.1514948692638427, "replacement": 1.146140307188034}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 3.26153564453125, "loss_scale": 4.0, "retries": 0, "seconds": 184.1, "max_memory_gb": 21.36, "pair": 0.2740105986595154, "sign": 0.15963076706975698, "keep": 0.05591453250963241, "leak": 1.7030021399259567, "replacement": 0.8075338006019592}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 3.187152862548828, "loss_scale": 4.0, "retries": 0, "seconds": 185.3, "max_memory_gb": 21.36, "pair": 0.4559359923005104, "sign": 0.3366625071503222, "keep": 0.1763151795603335, "leak": 0.8791943013202399, "replacement": 0.932223416864872}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 2.0203216075897217, "loss_scale": 4.0, "retries": 0, "seconds": 186.4, "max_memory_gb": 21.36, "pair": 0.5479617193341255, "sign": 0.5508919861167669, "keep": 0.3340334640815854, "leak": 0.5858052105177194, "replacement": 1.1368757784366608}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 2.219744920730591, "loss_scale": 4.0, "retries": 0, "seconds": 187.5, "max_memory_gb": 21.36, "pair": 0.5169581100344658, "sign": 0.5412260526791215, "keep": 0.07295295130461454, "leak": 0.2376798023469746, "replacement": 0.8834675848484039}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 2.5787582397460938, "loss_scale": 4.0, "retries": 0, "seconds": 188.5, "max_memory_gb": 21.36, "pair": 0.45330196991562843, "sign": 0.32668923027813435, "keep": 0.10568808671087027, "leak": 0.4048305144533515, "replacement": 0.770698219537735}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 3.2323901653289795, "loss_scale": 4.0, "retries": 0, "seconds": 189.7, "max_memory_gb": 21.36, "pair": 0.6469519548118114, "sign": 0.814121613278985, "keep": 0.4478905573487282, "leak": 0.42578394421434496, "replacement": 1.384404018521309}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 3.187204122543335, "loss_scale": 4.0, "retries": 0, "seconds": 190.9, "max_memory_gb": 21.36, "pair": 0.525134664028883, "sign": 0.5651420056819916, "keep": 0.17846947629004717, "leak": 1.1294112540781498, "replacement": 1.179293230175972}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 2.87784743309021, "loss_scale": 4.0, "retries": 0, "seconds": 192.1, "max_memory_gb": 21.36, "pair": 0.5536147281527519, "sign": 0.49985622987151146, "keep": 0.27537379786372185, "leak": 0.8332384370296495, "replacement": 1.1495393365621567}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 4.168179988861084, "loss_scale": 4.0, "retries": 0, "seconds": 193.3, "max_memory_gb": 21.36, "pair": 0.4373744986951351, "sign": 0.3940234323963523, "keep": 0.46998531743884087, "leak": 1.3252646068576723, "replacement": 1.2006950080394745}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 3.014568567276001, "loss_scale": 4.0, "retries": 0, "seconds": 194.4, "max_memory_gb": 21.36, "pair": 0.35825277864933014, "sign": 0.3435008656233549, "keep": 0.1038464477751404, "leak": 1.342751756310463, "replacement": 0.9176143482327461}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 3.7126026153564453, "loss_scale": 4.0, "retries": 0, "seconds": 195.5, "max_memory_gb": 21.36, "pair": 0.3842750731855631, "sign": 0.4391781985759735, "keep": 1.0293838786892593, "leak": 0.8059140471741557, "replacement": 1.3200346305966377}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 4.2505784034729, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 21.36, "pair": 0.1946396753191948, "sign": 0.0872795837931335, "keep": 0.01781346555799246, "leak": 1.7340431064367294, "replacement": 0.6806969791650772}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 2.039456844329834, "loss_scale": 4.0, "retries": 0, "seconds": 197.8, "max_memory_gb": 21.36, "pair": 0.4129308294504881, "sign": 0.36793341767042875, "keep": 0.10598402551840991, "leak": 0.6653999015688896, "replacement": 0.8162395134568214}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 5.633872032165527, "loss_scale": 4.0, "retries": 0, "seconds": 199.1, "max_memory_gb": 21.36, "pair": 0.22511836513876915, "sign": 0.07239759154617786, "keep": 0.13478255947120488, "leak": 3.905690328218043, "replacement": 1.3051310628652573}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 5.667068004608154, "loss_scale": 4.0, "retries": 0, "seconds": 200.1, "max_memory_gb": 21.36, "pair": 0.22578543797135353, "sign": 0.32325062714517117, "keep": 0.3102535344660282, "leak": 1.4207929372787476, "replacement": 0.8977357521653175}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 3.493631362915039, "loss_scale": 4.0, "retries": 0, "seconds": 201.3, "max_memory_gb": 21.36, "pair": 0.35636071115732193, "sign": 0.24351763352751732, "keep": 0.1403374692890793, "leak": 1.3570480607450008, "replacement": 0.8875502794981003}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.4769577980041504, "loss_scale": 4.0, "retries": 0, "seconds": 202.4, "max_memory_gb": 21.36, "pair": 0.4451853781938553, "sign": 0.5153948795050383, "keep": 0.24736038502305746, "leak": 0.8705780790187418, "replacement": 1.04420755058527}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 3.0939977169036865, "loss_scale": 4.0, "retries": 0, "seconds": 203.6, "max_memory_gb": 21.36, "pair": 0.40684986114501953, "sign": 0.17153815180063248, "keep": 0.06740383803844452, "leak": 1.34341830201447, "replacement": 0.8621754348278046}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 7.920099258422852, "loss_scale": 4.0, "retries": 0, "seconds": 204.6, "max_memory_gb": 21.36, "pair": 0.45561546832323074, "sign": 0.33743312768638134, "keep": 0.3816835936740972, "leak": 0.5822195075452328, "replacement": 0.9607287049293518}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.6585264205932617, "loss_scale": 4.0, "retries": 0, "seconds": 205.8, "max_memory_gb": 21.36, "pair": 0.44955698400735855, "sign": 0.47296356223523617, "keep": 0.2047177596250549, "leak": 1.6599734576884657, "replacement": 1.2033910006284714}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 2.676403760910034, "loss_scale": 4.0, "retries": 0, "seconds": 206.9, "max_memory_gb": 21.36, "pair": 0.6645759157836437, "sign": 0.7449943125247955, "keep": 0.5712249749340117, "leak": 0.7112102215178311, "replacement": 1.5004881173372269}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 6.355111598968506, "loss_scale": 4.0, "retries": 0, "seconds": 208.0, "max_memory_gb": 21.36, "pair": 0.6736181974411011, "sign": 0.5843163337558508, "keep": 0.8424391290172935, "leak": 0.4452873654663563, "replacement": 1.4983177036046982}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 7.73330020904541, "loss_scale": 4.0, "retries": 0, "seconds": 209.1, "max_memory_gb": 21.36, "pair": 0.430623859167099, "sign": 0.1905634757131338, "keep": 0.114246123470366, "leak": 1.0846997871994972, "replacement": 0.8542035967111588}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 3.099203586578369, "loss_scale": 4.0, "retries": 0, "seconds": 210.2, "max_memory_gb": 21.36, "pair": 0.540470365434885, "sign": 0.5655544390901923, "keep": 0.3054851154447533, "leak": 0.957111394032836, "replacement": 1.2152680158615112}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 3.4247491359710693, "loss_scale": 4.0, "retries": 0, "seconds": 211.3, "max_memory_gb": 21.36, "pair": 0.25053235329687595, "sign": 0.11602828372269869, "keep": 0.012770973233273253, "leak": 1.3944329768419266, "replacement": 0.663540244102478}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 4.361065864562988, "loss_scale": 4.0, "retries": 0, "seconds": 212.5, "max_memory_gb": 21.36, "pair": 0.3857466820627451, "sign": 0.28743037348613143, "keep": 0.07975786866154522, "leak": 1.6048776740208268, "replacement": 0.9705602079629898}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 2.8940792083740234, "loss_scale": 4.0, "retries": 0, "seconds": 213.6, "max_memory_gb": 21.36, "pair": 0.3639657683670521, "sign": 0.23459919728338718, "keep": 0.05359587771818042, "leak": 1.8238367913290858, "replacement": 0.9640225023031235}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 4.161771774291992, "loss_scale": 4.0, "retries": 0, "seconds": 214.8, "max_memory_gb": 21.36, "pair": 0.3138021379709244, "sign": 0.18308030627667904, "keep": 0.07385133329080418, "leak": 0.31154192984104156, "replacement": 0.5201534405350685}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 5.990965366363525, "loss_scale": 4.0, "retries": 0, "seconds": 216.0, "max_memory_gb": 21.36, "pair": 0.34830960631370544, "sign": 0.08485908899456263, "keep": 0.2877096962183714, "leak": 1.099636247381568, "replacement": 0.8095030710101128}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 2.4970881938934326, "loss_scale": 4.0, "retries": 0, "seconds": 217.1, "max_memory_gb": 21.36, "pair": 0.5120888873934746, "sign": 0.5345046073198318, "keep": 0.2218372246134095, "leak": 0.3855026912642643, "replacement": 0.9866355061531067}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.6946868896484375, "loss_scale": 4.0, "retries": 0, "seconds": 218.2, "max_memory_gb": 21.36, "pair": 0.25126173347234726, "sign": 0.19709009490907192, "keep": 0.07023341278545558, "leak": 0.554096095263958, "replacement": 0.5234475061297417}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 4.385491371154785, "loss_scale": 4.0, "retries": 0, "seconds": 219.4, "max_memory_gb": 21.36, "pair": 0.14339222759008408, "sign": 0.11211624136194587, "keep": 0.037402470130473375, "leak": 1.4895923845469952, "replacement": 0.5905496850609779}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 3.3228633403778076, "loss_scale": 4.0, "retries": 0, "seconds": 220.4, "max_memory_gb": 21.36, "pair": 0.5197247676551342, "sign": 0.58506272546947, "keep": 0.7232918795198202, "leak": 0.416073814034462, "replacement": 1.277920514345169}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 2.482057809829712, "loss_scale": 4.0, "retries": 0, "seconds": 221.6, "max_memory_gb": 21.36, "pair": 0.538527600467205, "sign": 0.4517692066729069, "keep": 0.32872133899945766, "leak": 0.7801478686160408, "replacement": 1.1238098293542862}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 2.29079532623291, "loss_scale": 4.0, "retries": 0, "seconds": 222.7, "max_memory_gb": 21.36, "pair": 0.3166936822235584, "sign": 0.16571850329637527, "keep": 0.06287637306377292, "leak": 0.9231688380241394, "replacement": 0.6617833226919174}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 6.746323585510254, "loss_scale": 4.0, "retries": 0, "seconds": 223.9, "max_memory_gb": 21.36, "pair": 0.5986747592687607, "sign": 0.7945199916139245, "keep": 1.0602080128155649, "leak": 0.9183165989816189, "replacement": 1.755617916584015}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 4.280107498168945, "loss_scale": 4.0, "retries": 0, "seconds": 225.1, "max_memory_gb": 21.36, "pair": 0.5665567368268967, "sign": 0.3677121512591839, "keep": 0.3032711334526539, "leak": 0.809429295361042, "replacement": 1.104405716061592}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 6.551565647125244, "loss_scale": 4.0, "retries": 0, "seconds": 226.3, "max_memory_gb": 21.36, "pair": 0.1911116037517786, "sign": 0.09895223844796419, "keep": 0.04242823785170913, "leak": 2.2066766023635864, "replacement": 0.8134710043668747}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 3.351689100265503, "loss_scale": 4.0, "retries": 0, "seconds": 227.5, "max_memory_gb": 21.36, "pair": 0.43876415118575096, "sign": 0.4876205213367939, "keep": 0.16274523502215743, "leak": 0.819641581736505, "replacement": 0.9688574522733688}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 4.329230308532715, "loss_scale": 4.0, "retries": 0, "seconds": 228.6, "max_memory_gb": 21.36, "pair": 0.6120578348636627, "sign": 0.5965830087661743, "keep": 0.2673628404736519, "leak": 0.4044352554483339, "replacement": 1.1451395601034164}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.871175527572632, "loss_scale": 4.0, "retries": 0, "seconds": 229.8, "max_memory_gb": 21.36, "pair": 0.2617383301258087, "sign": 0.17887595389038324, "keep": 0.16243304638192058, "leak": 1.017646849155426, "replacement": 0.6868045553565025}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 3.9078125953674316, "loss_scale": 4.0, "retries": 0, "seconds": 230.9, "max_memory_gb": 21.36, "pair": 0.579170998185873, "sign": 0.5760465878993273, "keep": 0.3664995856815949, "leak": 0.349186847161036, "replacement": 1.137740820646286}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 3.4584364891052246, "loss_scale": 4.0, "retries": 0, "seconds": 232.1, "max_memory_gb": 21.36, "pair": 0.4807360805571079, "sign": 0.6096724467352033, "keep": 0.2165642622858286, "leak": 0.6488756537437439, "replacement": 1.056073322892189}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 2.8064348697662354, "loss_scale": 4.0, "retries": 0, "seconds": 233.2, "max_memory_gb": 21.36, "pair": 0.3686777949333191, "sign": 0.32927217707037926, "keep": 0.004498615919146687, "leak": 1.0313248038291931, "replacement": 0.7933944165706635}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 2.3918421268463135, "loss_scale": 4.0, "retries": 0, "seconds": 234.4, "max_memory_gb": 21.36, "pair": 0.36333997175097466, "sign": 0.417867798358202, "keep": 0.22185194492340088, "leak": 0.772660979651846, "replacement": 0.8763650692999363}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 6.12645959854126, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 21.36, "pair": 0.5573122948408127, "sign": 0.6329853856004775, "keep": 0.24818493705242872, "leak": 0.45822587748989463, "replacement": 1.1124539375305176}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 3.34521746635437, "loss_scale": 4.0, "retries": 0, "seconds": 236.7, "max_memory_gb": 21.36, "pair": 0.27498818561434746, "sign": 0.34096579253673553, "keep": 0.11469931420288049, "leak": 0.8776382207870483, "replacement": 0.7222303077578545}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 3.746840715408325, "loss_scale": 4.0, "retries": 0, "seconds": 237.9, "max_memory_gb": 21.36, "pair": 0.4297475107014179, "sign": 0.4846711363643408, "keep": 0.16091237214277498, "leak": 1.4488388299942017, "replacement": 1.1147489994764328}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 5.2790937423706055, "loss_scale": 4.0, "retries": 0, "seconds": 239.0, "max_memory_gb": 21.36, "pair": 0.7833981681615114, "sign": 0.9380671605467796, "keep": 1.5472468149964698, "leak": 0.4190808776766062, "replacement": 2.1308253556489944}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 4.065107822418213, "loss_scale": 4.0, "retries": 0, "seconds": 240.2, "max_memory_gb": 21.36, "pair": 0.20059733651578426, "sign": 0.07947059394791722, "keep": 0.025765716447494924, "leak": 1.5322045646607876, "replacement": 0.6362666264176369}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 2.592836618423462, "loss_scale": 4.0, "retries": 0, "seconds": 241.4, "max_memory_gb": 21.36, "pair": 0.4687754176557064, "sign": 0.5445412741973996, "keep": 0.299350188113749, "leak": 0.49316001776605844, "replacement": 1.0140111446380615}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 4.114535331726074, "loss_scale": 4.0, "retries": 0, "seconds": 242.6, "max_memory_gb": 21.36, "pair": 0.38769564777612686, "sign": 0.4769628867506981, "keep": 0.7399617349728942, "leak": 2.218775413930416, "replacement": 1.550851784646511}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 5.314186096191406, "loss_scale": 4.0, "retries": 0, "seconds": 243.7, "max_memory_gb": 21.36, "pair": 0.43156072683632374, "sign": 0.540021747816354, "keep": 0.3186814244836569, "leak": 1.1531460872502066, "replacement": 1.1491988152265549}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 5.078086853027344, "loss_scale": 4.0, "retries": 0, "seconds": 244.9, "max_memory_gb": 21.36, "pair": 0.46258530765771866, "sign": 0.7090642675757408, "keep": 0.752545177936554, "leak": 0.2573692100122571, "replacement": 1.2577323392033577}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 7.440600872039795, "loss_scale": 4.0, "retries": 0, "seconds": 246.1, "max_memory_gb": 21.36, "pair": 0.3405550867319107, "sign": 0.4436595607548952, "keep": 0.8453361236024648, "leak": 2.2803106494247913, "replacement": 1.5551306009292603}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 4.011364459991455, "loss_scale": 4.0, "retries": 0, "seconds": 247.3, "max_memory_gb": 21.36, "pair": 0.5083456262946129, "sign": 0.5779503509402275, "keep": 0.9756194305955432, "leak": 0.24702032582717948, "replacement": 1.3468855917453766}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 8.172988891601562, "loss_scale": 4.0, "retries": 0, "seconds": 248.5, "max_memory_gb": 21.36, "pair": 0.7341534197330475, "sign": 0.9728220403194427, "keep": 1.8902763831429183, "leak": 0.6027612254838459, "replacement": 2.316392809152603}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 2.8148813247680664, "loss_scale": 4.0, "retries": 0, "seconds": 249.7, "max_memory_gb": 21.36, "pair": 0.2588806115090847, "sign": 0.25965845212340355, "keep": 0.3702281005680561, "leak": 0.5920606888830662, "replacement": 0.7218390479683876}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 4.519750118255615, "loss_scale": 4.0, "retries": 0, "seconds": 250.8, "max_memory_gb": 21.36, "pair": 0.20154022797942162, "sign": 0.18373126117512584, "keep": 0.05391238338779658, "leak": 1.323875218629837, "replacement": 0.65133086591959}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 3.8710381984710693, "loss_scale": 4.0, "retries": 0, "seconds": 252.0, "max_memory_gb": 21.36, "pair": 0.6905471980571747, "sign": 0.955570749938488, "keep": 0.31001616042340174, "leak": 0.2628087182965828, "replacement": 1.3890428841114044}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 4.84311580657959, "loss_scale": 4.0, "retries": 0, "seconds": 253.2, "max_memory_gb": 21.36, "pair": 0.4247503727674484, "sign": 0.48739054799079895, "keep": 0.26812897133640945, "leak": 0.25371893937699497, "replacement": 0.8659398555755615}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 5.5665411949157715, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 21.36, "pair": 0.3757810667157173, "sign": 0.5175916738808155, "keep": 0.35882260010112077, "leak": 0.7903231009840965, "replacement": 1.0115689784288406}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 3.093839406967163, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 21.36, "pair": 0.31918114330619574, "sign": 0.24662068579345942, "keep": 0.2647589687258005, "leak": 0.6105593033134937, "replacement": 0.7275108098983765}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 3.798328399658203, "loss_scale": 4.0, "retries": 0, "seconds": 256.6, "max_memory_gb": 21.36, "pair": 0.32596575655043125, "sign": 0.20019745361059904, "keep": 0.29529020143672824, "leak": 0.9452531162824016, "replacement": 0.8100228682160378}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 4.355034351348877, "loss_scale": 4.0, "retries": 0, "seconds": 257.8, "max_memory_gb": 21.36, "pair": 0.5245202518999577, "sign": 0.6495196595788002, "keep": 0.44636296632234007, "leak": 0.36913900543004274, "replacement": 1.1647463142871857}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 2.4135870933532715, "loss_scale": 4.0, "retries": 0, "seconds": 259.0, "max_memory_gb": 21.36, "pair": 0.3664605598896742, "sign": 0.4071841100230813, "keep": 0.11446802114369348, "leak": 0.43746791424928233, "replacement": 0.7366536259651184}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 2.303454637527466, "loss_scale": 4.0, "retries": 0, "seconds": 260.1, "max_memory_gb": 21.36, "pair": 0.48472459614276886, "sign": 0.5298880664631724, "keep": 0.14817925449460745, "leak": 0.47955306712538004, "replacement": 0.9436464626342058}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 3.825605869293213, "loss_scale": 4.0, "retries": 0, "seconds": 261.3, "max_memory_gb": 21.36, "pair": 0.25121339224278927, "sign": 0.28652068227529526, "keep": 0.19185603747610003, "leak": 0.7186695858836174, "replacement": 0.6700691357254982}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.829970121383667, "loss_scale": 4.0, "retries": 0, "seconds": 262.4, "max_memory_gb": 21.36, "pair": 0.2954065054655075, "sign": 0.15538502857089043, "keep": 0.028480054752435535, "leak": 0.8459878638386726, "replacement": 0.5988359898328781}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 3.1447556018829346, "loss_scale": 4.0, "retries": 0, "seconds": 263.5, "max_memory_gb": 21.36, "pair": 0.2007601261138916, "sign": 0.29805780574679375, "keep": 0.03028941573575139, "leak": 0.9262072965502739, "replacement": 0.5964855551719666}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 4.02589225769043, "loss_scale": 4.0, "retries": 0, "seconds": 264.7, "max_memory_gb": 21.36, "pair": 0.4804484639316797, "sign": 0.6821459063794464, "keep": 0.7004626821144484, "leak": 0.8488453701138496, "replacement": 1.383964091539383}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 2.1865811347961426, "loss_scale": 4.0, "retries": 0, "seconds": 265.8, "max_memory_gb": 21.36, "pair": 0.6184660941362381, "sign": 0.7150770165026188, "keep": 0.591464669269044, "leak": 0.2060587856449274, "replacement": 1.3232516646385193}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 4.68886137008667, "loss_scale": 4.0, "retries": 0, "seconds": 266.9, "max_memory_gb": 21.36, "pair": 0.6058790683746338, "sign": 0.7681118622422218, "keep": 0.5945476000197232, "leak": 0.3839711770415306, "replacement": 1.3832015693187714}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 5.653889179229736, "loss_scale": 4.0, "retries": 0, "seconds": 268.1, "max_memory_gb": 21.36, "pair": 0.32966194301843643, "sign": 0.45492173731327057, "keep": 0.03170846728608012, "leak": 0.9721840736456215, "replacement": 0.816023051738739}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 3.995826482772827, "loss_scale": 4.0, "retries": 0, "seconds": 269.2, "max_memory_gb": 21.36, "pair": 0.15828231163322926, "sign": 0.11110506393015385, "keep": 0.24366963258944452, "leak": 1.6059319861233234, "replacement": 0.7371526658535004}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 5.671478271484375, "loss_scale": 4.0, "retries": 0, "seconds": 270.3, "max_memory_gb": 21.36, "pair": 0.36531169805675745, "sign": 0.4239499634131789, "keep": 0.30545431846985593, "leak": 2.002705104649067, "replacement": 1.2306901216506958}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 4.148064613342285, "loss_scale": 4.0, "retries": 0, "seconds": 271.5, "max_memory_gb": 21.36, "pair": 0.46359264105558395, "sign": 0.601146437227726, "keep": 0.7435354676563293, "leak": 0.8211092054843903, "replacement": 1.3412109315395355}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 4.896701812744141, "loss_scale": 4.0, "retries": 0, "seconds": 272.7, "max_memory_gb": 21.36, "pair": 0.5133080407977104, "sign": 0.4943973682820797, "keep": 0.4339100133802276, "leak": 0.4192570894956589, "replacement": 1.0822759568691254}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 2.7466654777526855, "loss_scale": 4.0, "retries": 0, "seconds": 273.8, "max_memory_gb": 21.36, "pair": 0.3959856331348419, "sign": 0.3909968789666891, "keep": 0.3033151074196212, "leak": 0.43604943610262126, "replacement": 0.8521539866924286}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 3.0845439434051514, "loss_scale": 4.0, "retries": 0, "seconds": 274.9, "max_memory_gb": 21.36, "pair": 0.2718368135392666, "sign": 0.2019004845060408, "keep": 0.10432220972143114, "leak": 1.0158179849386215, "replacement": 0.6789026409387589}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 3.7607808113098145, "loss_scale": 4.0, "retries": 0, "seconds": 276.1, "max_memory_gb": 21.36, "pair": 0.4962267130613327, "sign": 0.5606528744101524, "keep": 0.30469732359051704, "leak": 1.601968849543482, "replacement": 1.3293940275907516}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.3901658058166504, "loss_scale": 4.0, "retries": 0, "seconds": 277.2, "max_memory_gb": 21.36, "pair": 0.34256209805607796, "sign": 0.2866247147321701, "keep": 0.07594517778488807, "leak": 0.35819048807024956, "replacement": 0.6133946850895882}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 3.4613640308380127, "loss_scale": 4.0, "retries": 0, "seconds": 278.4, "max_memory_gb": 21.36, "pair": 0.433625303208828, "sign": 0.5995559096336365, "keep": 0.14144346251850948, "leak": 0.34918060241034254, "replacement": 0.8914201408624649}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 1.874016284942627, "loss_scale": 4.0, "retries": 0, "seconds": 279.5, "max_memory_gb": 21.36, "pair": 0.4640734065324068, "sign": 0.5652837841771543, "keep": 0.27680476621026173, "leak": 0.6596705261617899, "replacement": 1.0500353313982487}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 2.72001576423645, "loss_scale": 4.0, "retries": 0, "seconds": 280.8, "max_memory_gb": 21.36, "pair": 0.17152456939220428, "sign": 0.18442491441965103, "keep": 0.05132039834279567, "leak": 1.8062005788087845, "replacement": 0.7409473657608032}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 8.729917526245117, "loss_scale": 4.0, "retries": 0, "seconds": 282.0, "max_memory_gb": 21.36, "pair": 0.6406796835362911, "sign": 1.0109797520563006, "keep": 1.4558674842119217, "leak": 0.9357261210680008, "replacement": 2.108034871518612}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 4.325392246246338, "loss_scale": 4.0, "retries": 0, "seconds": 283.3, "max_memory_gb": 21.36, "pair": 0.7910960428416729, "sign": 0.9728198274970055, "keep": 0.7625944539904594, "leak": 0.4231882016174495, "replacement": 1.764600247144699}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 5.467565536499023, "loss_scale": 4.0, "retries": 0, "seconds": 284.4, "max_memory_gb": 21.36, "pair": 0.1954549071379006, "sign": 0.20404108264483511, "keep": 0.08761333976872265, "leak": 1.525293257087469, "replacement": 0.7226054221391678}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 2.9948132038116455, "loss_scale": 4.0, "retries": 0, "seconds": 285.5, "max_memory_gb": 21.36, "pair": 0.30935199186205864, "sign": 0.3535515367984772, "keep": 0.1433207653753925, "leak": 0.4171234234236181, "replacement": 0.6620689928531647}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 3.0185139179229736, "loss_scale": 4.0, "retries": 0, "seconds": 286.6, "max_memory_gb": 21.36, "pair": 0.24116641283035278, "sign": 0.16274785995483398, "keep": 0.16074566263705492, "leak": 0.5592055469751358, "replacement": 0.5427145659923553}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 4.316273212432861, "loss_scale": 4.0, "retries": 0, "seconds": 287.8, "max_memory_gb": 21.36, "pair": 0.5780943632125854, "sign": 0.84968002140522, "keep": 0.07283553248271346, "leak": 0.5780584650347009, "replacement": 1.18386672437191}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 4.886379241943359, "loss_scale": 4.0, "retries": 0, "seconds": 289.0, "max_memory_gb": 21.36, "pair": 0.22532249614596367, "sign": 0.2314147911965847, "keep": 0.14545953983906657, "leak": 1.46437918394804, "replacement": 0.7798544615507126}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 2.7710139751434326, "loss_scale": 4.0, "retries": 0, "seconds": 290.2, "max_memory_gb": 21.36, "pair": 0.3651289977133274, "sign": 0.4043091908097267, "keep": 0.17955525463912636, "leak": 0.8203337226295844, "replacement": 0.8621446490287781}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 4.364136219024658, "loss_scale": 4.0, "retries": 0, "seconds": 291.4, "max_memory_gb": 21.36, "pair": 0.3380509167909622, "sign": 0.20583605393767357, "keep": 0.020750923897139728, "leak": 0.8721669726073742, "replacement": 0.6693861335515976}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 6.54511022567749, "loss_scale": 4.0, "retries": 0, "seconds": 292.5, "max_memory_gb": 21.36, "pair": 0.2863859124481678, "sign": 0.3288883548229933, "keep": 0.27843918971484527, "leak": 1.1190011277794838, "replacement": 0.869799941778183}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 3.698011636734009, "loss_scale": 4.0, "retries": 0, "seconds": 293.6, "max_memory_gb": 21.36, "pair": 0.27764106914401054, "sign": 0.3565740855410695, "keep": 0.31846886221319437, "leak": 0.5947103127837181, "replacement": 0.7638401240110397}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 2.7595717906951904, "loss_scale": 4.0, "retries": 0, "seconds": 294.8, "max_memory_gb": 21.36, "pair": 0.4734725058078766, "sign": 0.6279358938336372, "keep": 0.1228687206748873, "leak": 0.42220527003519237, "replacement": 0.9544261693954468}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 4.252193450927734, "loss_scale": 4.0, "retries": 0, "seconds": 295.9, "max_memory_gb": 21.36, "pair": 0.424492159858346, "sign": 0.4986765496432781, "keep": 0.10741305258125067, "leak": 0.8715910476748832, "replacement": 0.9454347342252731}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 8.38437271118164, "loss_scale": 4.0, "retries": 0, "seconds": 297.1, "max_memory_gb": 21.36, "pair": 0.4735738541930914, "sign": 0.5376549670472741, "keep": 0.3010972041229252, "leak": 1.48231790214777, "replacement": 1.263529397547245}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 4.261892318725586, "loss_scale": 4.0, "retries": 0, "seconds": 298.3, "max_memory_gb": 21.36, "pair": 0.18292425014078617, "sign": 0.16155067970976233, "keep": 0.1549203055910766, "leak": 0.9265977367758751, "replacement": 0.5728091821074486}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 2.0067801475524902, "loss_scale": 4.0, "retries": 0, "seconds": 299.5, "max_memory_gb": 21.36, "pair": 0.18478925339877605, "sign": 0.18047180771827698, "keep": 0.005827072134707123, "leak": 0.7436061650514603, "replacement": 0.4638402462005615}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 5.288995265960693, "loss_scale": 4.0, "retries": 0, "seconds": 300.6, "max_memory_gb": 21.36, "pair": 0.2973104976117611, "sign": 0.36189640685915947, "keep": 0.2047147355042398, "leak": 0.8688923716545105, "replacement": 0.7978391796350479}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.3744499683380127, "loss_scale": 4.0, "retries": 0, "seconds": 301.8, "max_memory_gb": 21.36, "pair": 0.32292556017637253, "sign": 0.3187592402100563, "keep": 0.2825388861820102, "leak": 0.8813405162654817, "replacement": 0.843909777700901}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 5.363153457641602, "loss_scale": 4.0, "retries": 0, "seconds": 302.9, "max_memory_gb": 21.36, "pair": 0.47721215710043907, "sign": 0.5384729811921716, "keep": 0.43237510000471957, "leak": 0.2483962745172903, "replacement": 1.0247352719306946}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 11.116636276245117, "loss_scale": 4.0, "retries": 0, "seconds": 304.1, "max_memory_gb": 21.36, "pair": 0.5830888375639915, "sign": 0.45812701620161533, "keep": 0.485115472227335, "leak": 0.6508984379470348, "replacement": 1.2174347341060638}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 7.90048360824585, "loss_scale": 4.0, "retries": 0, "seconds": 305.3, "max_memory_gb": 21.36, "pair": 0.2369910627603531, "sign": 0.15987679734826088, "keep": 0.7742403661250137, "leak": 1.173084328765981, "replacement": 0.997320719063282}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 3.2721517086029053, "loss_scale": 4.0, "retries": 0, "seconds": 306.4, "max_memory_gb": 21.36, "pair": 0.3088780343532562, "sign": 0.2933505065739155, "keep": 0.0809344615554437, "leak": 0.9770601709315088, "replacement": 0.7402855381369591}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 7.820058822631836, "loss_scale": 4.0, "retries": 0, "seconds": 307.6, "max_memory_gb": 21.36, "pair": 0.5655317772179842, "sign": 0.5392075711861253, "keep": 1.3192981192842126, "leak": 0.2516571590676904, "replacement": 1.5576989352703094}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.6042282581329346, "loss_scale": 4.0, "retries": 0, "seconds": 308.8, "max_memory_gb": 21.36, "pair": 0.40265556052327156, "sign": 0.45603936491534114, "keep": 0.35481691674795, "leak": 0.41263650730252266, "replacement": 0.9112428203225136}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 1.8052120208740234, "loss_scale": 4.0, "retries": 0, "seconds": 310.0, "max_memory_gb": 21.36, "pair": 0.5690276175737381, "sign": 0.6975261829793453, "keep": 0.3178304198663682, "leak": 0.1895853360183537, "replacement": 1.1241022571921349}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 5.706128120422363, "loss_scale": 4.0, "retries": 0, "seconds": 311.1, "max_memory_gb": 21.36, "pair": 0.3338235095143318, "sign": 0.44617578014731407, "keep": 1.4163148230873048, "leak": 0.10721225896850228, "replacement": 1.2918719053268433}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 6.474332332611084, "loss_scale": 4.0, "retries": 0, "seconds": 312.3, "max_memory_gb": 21.36, "pair": 0.28419071435928345, "sign": 0.4370990674942732, "keep": 0.5821084270719439, "leak": 1.1920102760195732, "replacement": 1.0917970314621925}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 5.172911643981934, "loss_scale": 4.0, "retries": 0, "seconds": 313.4, "max_memory_gb": 21.36, "pair": 0.2926951125264168, "sign": 0.3059272011741996, "keep": 0.0956624117679894, "leak": 1.0824043042957783, "replacement": 0.7640910148620605}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 4.006451606750488, "loss_scale": 4.0, "retries": 0, "seconds": 314.5, "max_memory_gb": 21.36, "pair": 0.16496815718710423, "sign": 0.062451726756989956, "keep": 0.031240368261933327, "leak": 0.7878468937706202, "replacement": 0.4087759181857109}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 4.253701686859131, "loss_scale": 4.0, "retries": 0, "seconds": 315.7, "max_memory_gb": 21.36, "pair": 0.3192378133535385, "sign": 0.16147346794605255, "keep": 0.049959504744037986, "leak": 0.2836427018046379, "replacement": 0.49586496502161026}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 4.102715492248535, "loss_scale": 4.0, "retries": 0, "seconds": 316.8, "max_memory_gb": 21.36, "pair": 0.4241166263818741, "sign": 0.6426583975553513, "keep": 0.22449716133996844, "leak": 0.2227761538233608, "replacement": 0.9133884757757187}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 3.4151952266693115, "loss_scale": 4.0, "retries": 0, "seconds": 318.0, "max_memory_gb": 21.36, "pair": 0.3528502304106951, "sign": 0.42433313094079494, "keep": 1.3365609240718186, "leak": 0.5639398014172912, "replacement": 1.3742822855710983}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 3.183993339538574, "loss_scale": 4.0, "retries": 0, "seconds": 319.2, "max_memory_gb": 21.36, "pair": 0.1869714530184865, "sign": 0.09447808191180229, "keep": 0.032701895746868104, "leak": 0.41010619059670717, "replacement": 0.35308799147605896}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 4.290737152099609, "loss_scale": 4.0, "retries": 0, "seconds": 320.4, "max_memory_gb": 21.36, "pair": 0.34319693222641945, "sign": 0.2784497458487749, "keep": 0.01706305571133271, "leak": 0.5196300400421023, "replacement": 0.620860829949379}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 6.394667625427246, "loss_scale": 4.0, "retries": 0, "seconds": 321.5, "max_memory_gb": 21.36, "pair": 0.33964506536722183, "sign": 0.5517204888164997, "keep": 0.5949052011710592, "leak": 0.3301002159714699, "replacement": 0.9954829588532448}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 3.6761457920074463, "loss_scale": 4.0, "retries": 0, "seconds": 322.6, "max_memory_gb": 21.36, "pair": 0.3535083271563053, "sign": 0.28237324208021164, "keep": 0.1070429872488603, "leak": 0.5977616952732205, "replacement": 0.6976568847894669}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 2.7956504821777344, "loss_scale": 4.0, "retries": 0, "seconds": 323.7, "max_memory_gb": 21.36, "pair": 0.26104462146759033, "sign": 0.1693377709016204, "keep": 0.042687647917773575, "leak": 0.528509771451354, "replacement": 0.4991847574710846}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.4569380283355713, "loss_scale": 4.0, "retries": 0, "seconds": 324.9, "max_memory_gb": 21.36, "pair": 0.24996347725391388, "sign": 0.25062888115644455, "keep": 0.38478025881340727, "leak": 0.315862323041074, "replacement": 0.646633617579937}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.986511468887329, "loss_scale": 4.0, "retries": 0, "seconds": 326.0, "max_memory_gb": 21.36, "pair": 0.31119571533054113, "sign": 0.3503872873261571, "keep": 0.23277416481869295, "leak": 0.376574270427227, "replacement": 0.6969199925661087}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 3.9899494647979736, "loss_scale": 4.0, "retries": 0, "seconds": 327.2, "max_memory_gb": 21.36, "pair": 0.24071870930492878, "sign": 0.1327685583382845, "keep": 0.1478450265713036, "leak": 1.7012818751609302, "replacement": 0.8063459694385529}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 4.810959815979004, "loss_scale": 4.0, "retries": 0, "seconds": 328.4, "max_memory_gb": 21.36, "pair": 0.23985048197209835, "sign": 0.17203339654952288, "keep": 0.0979415065376088, "leak": 1.3331576064229012, "replacement": 0.7081273421645164}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 4.175482273101807, "loss_scale": 4.0, "retries": 0, "seconds": 329.5, "max_memory_gb": 21.36, "pair": 0.4091448113322258, "sign": 0.6095201000571251, "keep": 0.20320368860848248, "leak": 0.6127815601066686, "replacement": 0.968702107667923}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 6.099386215209961, "loss_scale": 4.0, "retries": 0, "seconds": 330.6, "max_memory_gb": 21.36, "pair": 0.46959663182497025, "sign": 0.7562352083623409, "keep": 0.052860282361507416, "leak": 0.5403031129390001, "replacement": 1.0092201605439186}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 6.173405647277832, "loss_scale": 4.0, "retries": 0, "seconds": 331.7, "max_memory_gb": 21.36, "pair": 0.3291414622217417, "sign": 0.4334150890354067, "keep": 0.4660018486902118, "leak": 0.13883105307013466, "replacement": 0.8135576844215393}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 4.899879455566406, "loss_scale": 4.0, "retries": 0, "seconds": 332.9, "max_memory_gb": 21.36, "pair": 0.41693131625652313, "sign": 0.4472486423328519, "keep": 0.03943890007212758, "leak": 0.9692123556742445, "replacement": 0.9025782085955143}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 7.256515026092529, "loss_scale": 4.0, "retries": 0, "seconds": 334.1, "max_memory_gb": 21.36, "pair": 0.20178677514195442, "sign": 0.19526306400075555, "keep": 0.14008845016360283, "leak": 1.6289110351353884, "replacement": 0.7766902819275856}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 8.225711822509766, "loss_scale": 4.0, "retries": 0, "seconds": 335.3, "max_memory_gb": 21.36, "pair": 0.3268155511468649, "sign": 0.4269259494030848, "keep": 0.3256681394414045, "leak": 1.1056952476501465, "replacement": 0.979536384344101}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 6.643019676208496, "loss_scale": 4.0, "retries": 0, "seconds": 336.4, "max_memory_gb": 21.36, "pair": 0.06460100412368774, "sign": 0.04544994980096817, "keep": 0.012722620856948197, "leak": 2.155136816203594, "replacement": 0.6324714869260788}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 4.9730224609375, "loss_scale": 4.0, "retries": 0, "seconds": 337.6, "max_memory_gb": 21.36, "pair": 0.2264843638986349, "sign": 0.1721676387824118, "keep": 0.03106348286382854, "leak": 0.9219634383916855, "replacement": 0.558590792119503}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 3.341310739517212, "loss_scale": 4.0, "retries": 0, "seconds": 338.7, "max_memory_gb": 21.36, "pair": 0.2115099374204874, "sign": 0.22564221359789371, "keep": 0.042474780697375536, "leak": 0.5220466405153275, "replacement": 0.47608010098338127}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 7.2489399909973145, "loss_scale": 4.0, "retries": 0, "seconds": 339.9, "max_memory_gb": 21.36, "pair": 0.290187232196331, "sign": 0.19025935418903828, "keep": 0.4441276267170906, "leak": 0.558638222515583, "replacement": 0.7470402717590332}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 8.281851768493652, "loss_scale": 4.0, "retries": 0, "seconds": 341.0, "max_memory_gb": 21.36, "pair": 0.32180021330714226, "sign": 0.5155865997076035, "keep": 0.4947207230143249, "leak": 1.1929032504558563, "replacement": 1.1251797527074814}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 7.597646713256836, "loss_scale": 4.0, "retries": 0, "seconds": 342.1, "max_memory_gb": 21.36, "pair": 0.519785076379776, "sign": 0.7790310680866241, "keep": 0.7249303134158254, "leak": 0.5244064300786704, "replacement": 1.4028673246502876}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 5.1490020751953125, "loss_scale": 4.0, "retries": 0, "seconds": 343.3, "max_memory_gb": 21.36, "pair": 0.47738705202937126, "sign": 0.6874239593744278, "keep": 0.3679943227325566, "leak": 0.5264420656021684, "replacement": 1.1367066949605942}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 6.149434566497803, "loss_scale": 4.0, "retries": 0, "seconds": 344.4, "max_memory_gb": 21.36, "pair": 0.48903822153806686, "sign": 0.6883394941687584, "keep": 0.879188486665953, "leak": 0.04013049418460923, "replacement": 1.2828348353505135}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 5.240923881530762, "loss_scale": 4.0, "retries": 0, "seconds": 345.6, "max_memory_gb": 21.36, "pair": 0.22327645495533943, "sign": 0.11811233218759298, "keep": 0.05323340091854334, "leak": 1.2285071103833616, "replacement": 0.6160760819911957}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 8.571062088012695, "loss_scale": 4.0, "retries": 0, "seconds": 346.7, "max_memory_gb": 21.36, "pair": 0.6130746565759182, "sign": 0.8596977572888136, "keep": 0.6055013770237565, "leak": 0.8969332803972065, "replacement": 1.5699076056480408}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 6.319057941436768, "loss_scale": 4.0, "retries": 0, "seconds": 347.9, "max_memory_gb": 21.36, "pair": 0.286675157956779, "sign": 0.37020121747627854, "keep": 0.10527951596304774, "leak": 0.6263605669955723, "replacement": 0.6810056567192078}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 7.137359142303467, "loss_scale": 4.0, "retries": 0, "seconds": 349.1, "max_memory_gb": 21.36, "pair": 0.22026570327579975, "sign": 0.06511914310976863, "keep": 0.0768120652064681, "leak": 0.42341802397277206, "replacement": 0.3970858044922352}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 5.341692924499512, "loss_scale": 4.0, "retries": 0, "seconds": 350.3, "max_memory_gb": 21.36, "pair": 0.3654074501246214, "sign": 0.25288811326026917, "keep": 0.06189878936856985, "leak": 1.0037515088915825, "replacement": 0.7737387642264366}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 5.66582727432251, "loss_scale": 4.0, "retries": 0, "seconds": 351.4, "max_memory_gb": 21.36, "pair": 0.49357507284730673, "sign": 0.7254648562520742, "keep": 0.052100920118391514, "leak": 1.1376062808558345, "replacement": 1.1667595896869898}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 8.139337539672852, "loss_scale": 4.0, "retries": 0, "seconds": 352.6, "max_memory_gb": 21.36, "pair": 0.18548481911420822, "sign": 0.12308617448434234, "keep": 0.07275887887226418, "leak": 2.5903540290892124, "replacement": 0.930995836853981}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 5.292173385620117, "loss_scale": 4.0, "retries": 0, "seconds": 353.7, "max_memory_gb": 21.36, "pair": 0.1427753483876586, "sign": 0.04313221946358681, "keep": 0.054040283896028996, "leak": 0.7200728319585323, "replacement": 0.3713798001408577}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 8.793728828430176, "loss_scale": 4.0, "retries": 0, "seconds": 354.9, "max_memory_gb": 21.36, "pair": 0.36132755503058434, "sign": 0.39314621686935425, "keep": 0.35238262824714184, "leak": 0.7105657905340195, "replacement": 0.9117333963513374}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 6.558751583099365, "loss_scale": 4.0, "retries": 0, "seconds": 356.1, "max_memory_gb": 21.36, "pair": 0.22195159271359444, "sign": 0.4404575303196907, "keep": 0.06771078516612761, "leak": 2.1528027784079313, "replacement": 1.0142364613711834}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 9.997020721435547, "loss_scale": 4.0, "retries": 0, "seconds": 357.2, "max_memory_gb": 21.36, "pair": 0.5389794334769249, "sign": 0.698832678142935, "keep": 0.5017150298226625, "leak": 0.7840151637792587, "replacement": 1.3352570533752441}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 2.5945913791656494, "loss_scale": 4.0, "retries": 0, "seconds": 358.4, "max_memory_gb": 21.36, "pair": 0.23723992332816124, "sign": 0.11136558745056391, "keep": 0.030691867956193164, "leak": 0.9592676609754562, "replacement": 0.5480855703353882}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 3.8820183277130127, "loss_scale": 4.0, "retries": 0, "seconds": 359.5, "max_memory_gb": 21.36, "pair": 0.3332038652151823, "sign": 0.4278487414121628, "keep": 0.31124537746654823, "leak": 0.3208603113889694, "replacement": 0.7829660549759865}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.262485027313232, "loss_scale": 4.0, "retries": 0, "seconds": 360.6, "max_memory_gb": 21.36, "pair": 0.29388552717864513, "sign": 0.2103351061232388, "keep": 0.07463262788951397, "leak": 0.18813178595155478, "replacement": 0.4834023416042328}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 9.458954811096191, "loss_scale": 4.0, "retries": 0, "seconds": 361.8, "max_memory_gb": 21.36, "pair": 0.3672340977936983, "sign": 0.6367652378976345, "keep": 0.752265261951834, "leak": 1.3559827208518982, "replacement": 1.400745078921318}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 6.090293884277344, "loss_scale": 4.0, "retries": 0, "seconds": 363.0, "max_memory_gb": 21.36, "pair": 0.30122561007738113, "sign": 0.3896561088040471, "keep": 0.3813263240735978, "leak": 0.5810279995203018, "replacement": 0.8319738134741783}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 4.905755996704102, "loss_scale": 4.0, "retries": 0, "seconds": 364.2, "max_memory_gb": 21.36, "pair": 0.23276962526142597, "sign": 0.2263080244883895, "keep": 0.07860597036778927, "leak": 0.6412511570379138, "replacement": 0.545539416372776}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 4.230419158935547, "loss_scale": 4.0, "retries": 0, "seconds": 365.4, "max_memory_gb": 21.36, "pair": 0.4699028544127941, "sign": 0.7210351931862533, "keep": 0.5667326264083385, "leak": 1.2068005724577233, "replacement": 1.4154868870973587}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 3.9934656620025635, "loss_scale": 4.0, "retries": 0, "seconds": 366.5, "max_memory_gb": 21.36, "pair": 0.321192380040884, "sign": 0.37152722105383873, "keep": 0.22756410809233785, "leak": 0.7690034471452236, "replacement": 0.8129888772964478}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 4.0951995849609375, "loss_scale": 4.0, "retries": 0, "seconds": 367.6, "max_memory_gb": 21.36, "pair": 0.23568664398044348, "sign": 0.24007080495357513, "keep": 0.04377823747927323, "leak": 1.0750569477677345, "replacement": 0.6463753953576088}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 7.244014739990234, "loss_scale": 4.0, "retries": 0, "seconds": 368.7, "max_memory_gb": 21.36, "pair": 0.37784528359770775, "sign": 0.31416346691548824, "keep": 0.12282598577439785, "leak": 0.4860879140906036, "replacement": 0.7178619801998138}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 8.40770435333252, "loss_scale": 4.0, "retries": 0, "seconds": 369.9, "max_memory_gb": 21.36, "pair": 0.3705153651535511, "sign": 0.4541374482214451, "keep": 0.6209674007259309, "leak": 0.5945763513445854, "replacement": 1.0567118600010872}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 5.234339714050293, "loss_scale": 4.0, "retries": 0, "seconds": 371.1, "max_memory_gb": 21.36, "pair": 0.3260985314846039, "sign": 0.2726070787757635, "keep": 0.06606515636667609, "leak": 0.3910744332242757, "replacement": 0.5932032391428947}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 3.439711093902588, "loss_scale": 4.0, "retries": 0, "seconds": 372.2, "max_memory_gb": 21.36, "pair": 0.3412835858762264, "sign": 0.49868882820010185, "keep": 0.034374472219496965, "leak": 0.3309300011023879, "replacement": 0.6905477494001389}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 5.947821617126465, "loss_scale": 4.0, "retries": 0, "seconds": 373.4, "max_memory_gb": 21.36, "pair": 0.1851351521909237, "sign": 0.15945946238934994, "keep": 0.07545182295143604, "leak": 1.8282693028450012, "replacement": 0.7596581131219864}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 3.8817315101623535, "loss_scale": 4.0, "retries": 0, "seconds": 374.5, "max_memory_gb": 21.36, "pair": 0.10145387426018715, "sign": 0.058558471500873566, "keep": 0.05213967474992387, "leak": 0.7903149500489235, "replacement": 0.354381687939167}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 9.411620140075684, "loss_scale": 4.0, "retries": 0, "seconds": 375.7, "max_memory_gb": 21.36, "pair": 0.678329274058342, "sign": 0.6768958484753966, "keep": 0.610999602911761, "leak": 0.2865909216925502, "replacement": 1.3939247727394104}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 5.179975986480713, "loss_scale": 4.0, "retries": 0, "seconds": 376.8, "max_memory_gb": 21.36, "pair": 0.2753794025629759, "sign": 0.4231195542961359, "keep": 0.1552057285880437, "leak": 0.504284193390049, "replacement": 0.6906130611896515}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 7.908730983734131, "loss_scale": 4.0, "retries": 0, "seconds": 378.0, "max_memory_gb": 21.36, "pair": 0.1981773804873228, "sign": 0.12216478493064642, "keep": 0.6754663190804422, "leak": 0.851877573877573, "replacement": 0.8099623396992683}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 6.849567413330078, "loss_scale": 4.0, "retries": 0, "seconds": 379.2, "max_memory_gb": 21.36, "pair": 0.41750677675008774, "sign": 0.564845122396946, "keep": 0.22574747615726665, "leak": 0.19256294588558376, "replacement": 0.8609438017010689}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 4.6272053718566895, "loss_scale": 4.0, "retries": 0, "seconds": 380.3, "max_memory_gb": 21.36, "pair": 0.38967330381274223, "sign": 0.4113242710009217, "keep": 0.2706318410637323, "leak": 0.45533969067037106, "replacement": 0.8444862887263298}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 4.18339729309082, "loss_scale": 4.0, "retries": 0, "seconds": 381.5, "max_memory_gb": 21.36, "pair": 0.44404910132288933, "sign": 0.5029437178745866, "keep": 0.35434799137874506, "leak": 0.3776144958101213, "replacement": 0.9670986160635948}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 6.628215312957764, "loss_scale": 4.0, "retries": 0, "seconds": 382.7, "max_memory_gb": 21.36, "pair": 0.5810458362102509, "sign": 0.6462484113872051, "keep": 0.6178502657276113, "leak": 0.3083524501416832, "replacement": 1.2901832833886147}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 6.361331462860107, "loss_scale": 4.0, "retries": 0, "seconds": 383.8, "max_memory_gb": 21.36, "pair": 0.3229852132499218, "sign": 0.33954691514372826, "keep": 0.3753751608892344, "leak": 0.3866451755166054, "replacement": 0.7771075367927551}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 4.958725929260254, "loss_scale": 4.0, "retries": 0, "seconds": 385.0, "max_memory_gb": 21.36, "pair": 0.6334025412797928, "sign": 0.7757497578859329, "keep": 0.6556208573456388, "leak": 0.32376701495377347, "replacement": 1.4300296306610107}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 5.799193859100342, "loss_scale": 4.0, "retries": 0, "seconds": 386.2, "max_memory_gb": 21.36, "pair": 0.4843081720173359, "sign": 0.5891125053167343, "keep": 0.7741519822739065, "leak": 1.1657610535621643, "replacement": 1.457380697131157}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 4.686560153961182, "loss_scale": 4.0, "retries": 0, "seconds": 387.3, "max_memory_gb": 21.36, "pair": 0.4080594852566719, "sign": 0.5361969657242298, "keep": 0.9526729171338957, "leak": 0.7539070066995919, "replacement": 1.3409712389111519}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 12.276488304138184, "loss_scale": 4.0, "retries": 0, "seconds": 388.5, "max_memory_gb": 21.36, "pair": 0.36131905764341354, "sign": 0.17820988409221172, "keep": 0.16081989897065796, "leak": 2.2684938623569906, "replacement": 1.0979574024677277}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 5.286421298980713, "loss_scale": 4.0, "retries": 0, "seconds": 389.7, "max_memory_gb": 21.36, "pair": 0.5390460230410099, "sign": 0.8537711706012487, "keep": 0.20153197273612022, "leak": 0.8422474605031312, "replacement": 1.2772594690322876}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 5.213544845581055, "loss_scale": 4.0, "retries": 0, "seconds": 390.9, "max_memory_gb": 21.36, "pair": 0.4261116720736027, "sign": 0.49679698795080185, "keep": 0.4216400943696499, "leak": 0.5988901342498139, "replacement": 1.0350527614355087}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 5.784239768981934, "loss_scale": 4.0, "retries": 0, "seconds": 392.1, "max_memory_gb": 21.36, "pair": 0.4651576764881611, "sign": 0.7531889099627733, "keep": 0.3536998121999204, "leak": 1.264675340091344, "replacement": 1.3347708508372307}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 4.39305305480957, "loss_scale": 4.0, "retries": 0, "seconds": 393.2, "max_memory_gb": 21.36, "pair": 0.31718449480831623, "sign": 0.35010131634771824, "keep": 0.6664002207107842, "leak": 0.32705070078372955, "replacement": 0.9071979261934757}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 5.597315311431885, "loss_scale": 4.0, "retries": 0, "seconds": 394.3, "max_memory_gb": 21.36, "pair": 0.2836385741829872, "sign": 0.3500802917405963, "keep": 0.32332120765931904, "leak": 0.8185835322365165, "replacement": 0.8249851614236832}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 6.182053089141846, "loss_scale": 4.0, "retries": 0, "seconds": 395.4, "max_memory_gb": 21.36, "pair": 0.5354424640536308, "sign": 0.6515394672751427, "keep": 0.6848939508199692, "leak": 1.082648715004325, "replacement": 1.4743213504552841}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 2.662724733352661, "loss_scale": 4.0, "retries": 0, "seconds": 396.6, "max_memory_gb": 21.36, "pair": 0.4135762155056, "sign": 0.4979089591652155, "keep": 0.02392421814147383, "leak": 0.3660625275806524, "replacement": 0.7660084217786789}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 3.15822696685791, "loss_scale": 4.0, "retries": 0, "seconds": 397.7, "max_memory_gb": 21.36, "pair": 0.15137485042214394, "sign": 0.11679853312671185, "keep": 0.021544366551097482, "leak": 0.8394595813006163, "replacement": 0.430411197245121}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 5.235080718994141, "loss_scale": 4.0, "retries": 0, "seconds": 398.8, "max_memory_gb": 21.36, "pair": 0.2687223330140114, "sign": 0.3468365194275975, "keep": 0.21523943729698658, "leak": 0.7182380203157663, "replacement": 0.7293198332190514}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 7.415347099304199, "loss_scale": 4.0, "retries": 0, "seconds": 400.0, "max_memory_gb": 21.36, "pair": 0.4433733746409416, "sign": 0.5946310125291348, "keep": 0.5925605350639671, "leak": 0.739273838698864, "replacement": 1.2217876464128494}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 3.3351833820343018, "loss_scale": 4.0, "retries": 0, "seconds": 401.2, "max_memory_gb": 21.36, "pair": 0.33682282269001007, "sign": 0.3745609982870519, "keep": 0.2896010894328356, "leak": 0.7294959753053263, "replacement": 0.8512778729200363}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 4.802908897399902, "loss_scale": 4.0, "retries": 0, "seconds": 402.3, "max_memory_gb": 21.36, "pair": 0.16149131674319506, "sign": 0.12524203187786043, "keep": 0.016826642211526632, "leak": 1.6392837539315224, "replacement": 0.6423466205596924}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 4.56488037109375, "loss_scale": 4.0, "retries": 0, "seconds": 403.6, "max_memory_gb": 21.36, "pair": 0.3100750809535384, "sign": 0.3502365993335843, "keep": 0.13414525537518784, "leak": 1.1014519517775625, "replacement": 0.8276289850473404}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 5.670352935791016, "loss_scale": 4.0, "retries": 0, "seconds": 404.7, "max_memory_gb": 21.36, "pair": 0.3475945256650448, "sign": 0.6746213883161545, "keep": 0.27661877404898405, "leak": 0.4689388140104711, "replacement": 0.9404493048787117}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 5.2626776695251465, "loss_scale": 4.0, "retries": 0, "seconds": 405.8, "max_memory_gb": 21.36, "pair": 0.2390437312424183, "sign": 0.1511789783835411, "keep": 0.11736893537454307, "leak": 0.15294038644788088, "replacement": 0.4115527868270874}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 7.630542755126953, "loss_scale": 4.0, "retries": 0, "seconds": 407.0, "max_memory_gb": 21.36, "pair": 0.4143196865916252, "sign": 0.6287994580343366, "keep": 0.48853589640930295, "leak": 1.2054617926478386, "replacement": 1.2743528336286545}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 7.336553573608398, "loss_scale": 4.0, "retries": 0, "seconds": 408.1, "max_memory_gb": 21.36, "pair": 0.38572338223457336, "sign": 0.6146519239991903, "keep": 0.8805749579332769, "leak": 1.015767965465784, "replacement": 1.3872788473963737}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 1.9640487432479858, "loss_scale": 4.0, "retries": 0, "seconds": 409.3, "max_memory_gb": 21.36, "pair": 0.4224836826324463, "sign": 0.4381362982094288, "keep": 0.28634391076047905, "leak": 0.19757234049029648, "replacement": 0.8341169431805611}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 5.070456504821777, "loss_scale": 4.0, "retries": 0, "seconds": 410.5, "max_memory_gb": 21.36, "pair": 0.4640294834971428, "sign": 0.457713462645188, "keep": 0.20817918714601547, "leak": 1.271425372142403, "replacement": 1.114832118153572}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 5.473617076873779, "loss_scale": 4.0, "retries": 0, "seconds": 411.7, "max_memory_gb": 21.36, "pair": 0.1375387515872717, "sign": 0.15690714865922928, "keep": 0.029548177029937506, "leak": 1.6871645748615265, "replacement": 0.6525575518608093}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 11.348252296447754, "loss_scale": 4.0, "retries": 0, "seconds": 412.8, "max_memory_gb": 21.36, "pair": 0.1518017277121544, "sign": 0.14440229604952037, "keep": 0.1607212144881487, "leak": 2.5799299478530884, "replacement": 0.9493459612131119}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 5.323300838470459, "loss_scale": 4.0, "retries": 0, "seconds": 414.0, "max_memory_gb": 21.36, "pair": 0.5073303282260895, "sign": 0.6804923266172409, "keep": 0.5875361495418474, "leak": 0.36241980424165376, "replacement": 1.2319495007395744}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 2.5704894065856934, "loss_scale": 4.0, "retries": 0, "seconds": 415.2, "max_memory_gb": 21.36, "pair": 0.35942288488149643, "sign": 0.4766990579664707, "keep": 0.08490857400465757, "leak": 0.834879744797945, "replacement": 0.8489466682076454}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 7.6605987548828125, "loss_scale": 4.0, "retries": 0, "seconds": 416.3, "max_memory_gb": 21.36, "pair": 0.18356730416417122, "sign": 0.19128178246319294, "keep": 0.03617723690695129, "leak": 1.5024676769971848, "replacement": 0.6729137152433395}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 5.573115825653076, "loss_scale": 4.0, "retries": 0, "seconds": 417.5, "max_memory_gb": 21.36, "pair": 0.29115717858076096, "sign": 0.49608090333640575, "keep": 0.017836876679211855, "leak": 0.6432760208845139, "replacement": 0.7089350819587708}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 4.877575874328613, "loss_scale": 4.0, "retries": 0, "seconds": 418.6, "max_memory_gb": 21.36, "pair": 0.41005177423357964, "sign": 0.42628306755796075, "keep": 1.589395776274614, "leak": 0.1220260919071734, "replacement": 1.4483977779746056}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 8.493833541870117, "loss_scale": 4.0, "retries": 0, "seconds": 419.7, "max_memory_gb": 21.36, "pair": 0.4252011813223362, "sign": 0.8080350570380688, "keep": 0.06085876328870654, "leak": 0.5949947480112314, "replacement": 1.0083967372775078}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 2.9198529720306396, "loss_scale": 4.0, "retries": 0, "seconds": 420.9, "max_memory_gb": 21.36, "pair": 0.2377955373376608, "sign": 0.06375635135918856, "keep": 0.0736604689154774, "leak": 0.558300661155954, "replacement": 0.4460791125893593}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 2.828958034515381, "loss_scale": 4.0, "retries": 0, "seconds": 422.1, "max_memory_gb": 21.36, "pair": 0.32062765397131443, "sign": 0.3775177039206028, "keep": 0.24660887947538868, "leak": 0.23775316774845123, "replacement": 0.6921292096376419}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 5.927920818328857, "loss_scale": 4.0, "retries": 0, "seconds": 423.2, "max_memory_gb": 21.36, "pair": 0.26725102588534355, "sign": 0.294392978772521, "keep": 0.050103478133678436, "leak": 1.6062724888324738, "replacement": 0.8410673812031746}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 3.0113823413848877, "loss_scale": 4.0, "retries": 0, "seconds": 424.4, "max_memory_gb": 21.36, "pair": 0.40007491037249565, "sign": 0.4071331024169922, "keep": 0.3516098251566291, "leak": 0.3692421279847622, "replacement": 0.8717569336295128}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 2.484997034072876, "loss_scale": 4.0, "retries": 0, "seconds": 425.5, "max_memory_gb": 21.36, "pair": 0.1911902278661728, "sign": 0.12027407949790359, "keep": 0.05422527942573652, "leak": 0.2357091624289751, "replacement": 0.3373671919107437}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 6.1099419593811035, "loss_scale": 4.0, "retries": 0, "seconds": 426.6, "max_memory_gb": 21.36, "pair": 0.43674697540700436, "sign": 0.3609100887551904, "keep": 0.05912447423906997, "leak": 0.4856328093446791, "replacement": 0.7681724578142166}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 11.6134033203125, "loss_scale": 4.0, "retries": 0, "seconds": 427.8, "max_memory_gb": 21.36, "pair": 0.19007454439997673, "sign": 0.27304246462881565, "keep": 1.223852900788188, "leak": 1.1197643538471311, "replacement": 1.2184633165597916}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.776763439178467, "loss_scale": 4.0, "retries": 0, "seconds": 428.9, "max_memory_gb": 21.36, "pair": 0.30331543646752834, "sign": 0.09894415410235524, "keep": 0.22300315857864916, "leak": 0.876583126373589, "replacement": 0.6834348887205124}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 8.31057071685791, "loss_scale": 4.0, "retries": 0, "seconds": 430.0, "max_memory_gb": 21.36, "pair": 0.3022050620056689, "sign": 0.23362487740814686, "keep": 0.07552011293591931, "leak": 1.076974008261459, "replacement": 0.7260210439562798}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 2.6182169914245605, "loss_scale": 4.0, "retries": 0, "seconds": 431.2, "max_memory_gb": 21.36, "pair": 0.5535994619131088, "sign": 0.7663986161351204, "keep": 0.05161812505684793, "leak": 0.3147704713046551, "replacement": 1.0413004979491234}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 4.66807746887207, "loss_scale": 4.0, "retries": 0, "seconds": 432.3, "max_memory_gb": 21.36, "pair": 0.2627468407154083, "sign": 0.30292644118890166, "keep": 0.4708408883307129, "leak": 0.47075285017490387, "replacement": 0.767318706959486}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 2.5278854370117188, "loss_scale": 4.0, "retries": 0, "seconds": 433.5, "max_memory_gb": 21.36, "pair": 0.3880983479321003, "sign": 0.4699472077190876, "keep": 0.259647571539972, "leak": 0.8136734694999177, "replacement": 0.9563141167163849}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.2950563430786133, "loss_scale": 4.0, "retries": 0, "seconds": 434.7, "max_memory_gb": 21.36, "pair": 0.2627793401479721, "sign": 0.22377412021160126, "keep": 0.1255446292925626, "leak": 0.1574702076613903, "replacement": 0.47680629417300224}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 3.099259853363037, "loss_scale": 4.0, "retries": 0, "seconds": 435.8, "max_memory_gb": 21.36, "pair": 0.29033802449703217, "sign": 0.347099713049829, "keep": 0.05065142991952598, "leak": 0.491612296551466, "replacement": 0.6121166795492172}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 5.028525352478027, "loss_scale": 4.0, "retries": 0, "seconds": 437.0, "max_memory_gb": 21.36, "pair": 0.2653629444539547, "sign": 0.1869663605466485, "keep": 0.17101819347590208, "leak": 0.608247451717034, "replacement": 0.5964170694351196}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 3.3851301670074463, "loss_scale": 4.0, "retries": 0, "seconds": 438.1, "max_memory_gb": 21.36, "pair": 0.18497169390320778, "sign": 0.25621766969561577, "keep": 0.11033497704192996, "leak": 0.2822399288415909, "replacement": 0.43880801275372505}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 7.1448798179626465, "loss_scale": 4.0, "retries": 0, "seconds": 439.3, "max_memory_gb": 21.36, "pair": 0.2119608335196972, "sign": 0.31150338519364595, "keep": 0.08263985067605972, "leak": 1.3531561717391014, "replacement": 0.7473215013742447}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 2.7262494564056396, "loss_scale": 4.0, "retries": 0, "seconds": 440.4, "max_memory_gb": 21.36, "pair": 0.5567594766616821, "sign": 0.7630100846290588, "keep": 0.4136268300935626, "leak": 0.12027750792913139, "replacement": 1.1751473173499107}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 5.629456520080566, "loss_scale": 4.0, "retries": 0, "seconds": 441.6, "max_memory_gb": 21.36, "pair": 0.45253701880574226, "sign": 0.6136347893625498, "keep": 0.9893934074789286, "leak": 0.3790825565115483, "replacement": 1.3488217927515507}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 3.485666036605835, "loss_scale": 4.0, "retries": 0, "seconds": 442.7, "max_memory_gb": 21.36, "pair": 0.2650163844227791, "sign": 0.2877695672214031, "keep": 0.14083653781563044, "leak": 0.6640387307852507, "replacement": 0.6453291326761246}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 8.231958389282227, "loss_scale": 4.0, "retries": 0, "seconds": 443.8, "max_memory_gb": 21.36, "pair": 0.0785004086792469, "sign": 0.07140028942376375, "keep": 0.020006010308861732, "leak": 1.3575804024003446, "replacement": 0.46359865739941597}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 3.01786732673645, "loss_scale": 4.0, "retries": 0, "seconds": 445.0, "max_memory_gb": 21.36, "pair": 0.3738573119044304, "sign": 0.46781909465789795, "keep": 0.2765202655282337, "leak": 0.30868306231013776, "replacement": 0.8231977745890617}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 3.4463236331939697, "loss_scale": 4.0, "retries": 0, "seconds": 446.2, "max_memory_gb": 21.36, "pair": 0.18158805929124355, "sign": 0.1836712653748691, "keep": 0.052331288083223626, "leak": 0.21760183200240135, "replacement": 0.35398978739976883}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 5.925828456878662, "loss_scale": 4.0, "retries": 0, "seconds": 447.4, "max_memory_gb": 21.36, "pair": 0.24463844671845436, "sign": 0.46024274453520775, "keep": 0.08688961493317038, "leak": 0.6178956180810928, "replacement": 0.6726785376667976}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 2.9428622722625732, "loss_scale": 4.0, "retries": 0, "seconds": 448.5, "max_memory_gb": 21.36, "pair": 0.13100353628396988, "sign": 0.16876663640141487, "keep": 0.04956325513194315, "leak": 0.5234497617930174, "replacement": 0.37103091180324554}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 5.492605209350586, "loss_scale": 4.0, "retries": 0, "seconds": 449.7, "max_memory_gb": 21.36, "pair": 0.12901445478200912, "sign": 0.09513832535594702, "keep": 0.07826076308265328, "leak": 1.698470251634717, "replacement": 0.640331543982029}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 6.733942031860352, "loss_scale": 4.0, "retries": 0, "seconds": 450.9, "max_memory_gb": 21.36, "pair": 0.20295459032058716, "sign": 0.3370535299181938, "keep": 0.16278517455793917, "leak": 1.3003191407769918, "replacement": 0.7779537290334702}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 6.543365955352783, "loss_scale": 4.0, "retries": 0, "seconds": 452.1, "max_memory_gb": 21.36, "pair": 0.5240155048668385, "sign": 0.7568396423012018, "keep": 0.7517118155956268, "leak": 0.2494562790961936, "replacement": 1.3406553640961647}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 3.034310817718506, "loss_scale": 4.0, "retries": 0, "seconds": 453.3, "max_memory_gb": 21.36, "pair": 0.08479003049433231, "sign": 0.07007372099906206, "keep": 0.07168082660064101, "leak": 0.30439168214797974, "replacement": 0.23176522552967072}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 2.983154773712158, "loss_scale": 4.0, "retries": 0, "seconds": 454.4, "max_memory_gb": 21.36, "pair": 0.13884548097848892, "sign": 0.19274931121617556, "keep": 0.09108428377658129, "leak": 0.23247875599190593, "replacement": 0.3388819620013237}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 6.926449298858643, "loss_scale": 4.0, "retries": 0, "seconds": 455.6, "max_memory_gb": 21.36, "pair": 0.4206216260790825, "sign": 0.49431291595101357, "keep": 0.154661706648767, "leak": 0.5191696712281555, "replacement": 0.8749013617634773}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 3.089836597442627, "loss_scale": 4.0, "retries": 0, "seconds": 456.7, "max_memory_gb": 21.36, "pair": 0.19134446047246456, "sign": 0.18211789429187775, "keep": 0.03329849417787045, "leak": 0.6495203897356987, "replacement": 0.4614327475428581}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 3.988429069519043, "loss_scale": 4.0, "retries": 0, "seconds": 457.8, "max_memory_gb": 21.36, "pair": 0.19616806507110596, "sign": 0.06082545407116413, "keep": 0.07379017770290375, "leak": 0.2989981498394627, "replacement": 0.33822541683912277}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 5.116646766662598, "loss_scale": 4.0, "retries": 0, "seconds": 459.0, "max_memory_gb": 21.36, "pair": 0.33197323977947235, "sign": 0.46800181083381176, "keep": 0.18551679467782378, "leak": 0.4853722592815757, "replacement": 0.7800756059587002}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 6.621747016906738, "loss_scale": 4.0, "retries": 0, "seconds": 460.2, "max_memory_gb": 21.36, "pair": 0.35378923639655113, "sign": 0.40652607288211584, "keep": 0.20312847063178197, "leak": 0.7364608161151409, "replacement": 0.8427317216992378}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 8.849384307861328, "loss_scale": 4.0, "retries": 0, "seconds": 461.3, "max_memory_gb": 21.36, "pair": 0.379663759842515, "sign": 0.7168350666761398, "keep": 0.9930760082788765, "leak": 0.9332075947895646, "replacement": 1.4679211899638176}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 8.73312759399414, "loss_scale": 4.0, "retries": 0, "seconds": 462.5, "max_memory_gb": 21.36, "pair": 0.38960445299744606, "sign": 0.3100838400423527, "keep": 0.19554270897060633, "leak": 0.7784291990101337, "replacement": 0.8370250165462494}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 14.024895668029785, "loss_scale": 4.0, "retries": 0, "seconds": 463.8, "max_memory_gb": 21.36, "pair": 0.31181230675429106, "sign": 0.585965882986784, "keep": 0.5408254984067753, "leak": 1.3363663144409657, "replacement": 1.209299635142088}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 17.539226531982422, "loss_scale": 4.0, "retries": 0, "seconds": 464.9, "max_memory_gb": 21.36, "pair": 0.793543191626668, "sign": 0.916504692286253, "keep": 0.4978085160255432, "leak": 0.35841628155321814, "replacement": 1.59030382335186}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 6.041408538818359, "loss_scale": 4.0, "retries": 0, "seconds": 466.1, "max_memory_gb": 21.36, "pair": 0.1262268559075892, "sign": 0.10817330819554627, "keep": 0.02139160456135869, "leak": 1.4429511334747076, "replacement": 0.5517470985651016}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 7.577330112457275, "loss_scale": 4.0, "retries": 0, "seconds": 467.3, "max_memory_gb": 21.36, "pair": 0.32254623528569937, "sign": 0.46913658641278744, "keep": 0.3980686850845814, "leak": 0.4314044415950775, "replacement": 0.8640000075101852}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 2.7714550495147705, "loss_scale": 4.0, "retries": 0, "seconds": 468.4, "max_memory_gb": 21.36, "pair": 0.26037855818867683, "sign": 0.21415568282827735, "keep": 0.05923579534282908, "leak": 0.4345051869750023, "replacement": 0.5057005845010281}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 4.2753095626831055, "loss_scale": 4.0, "retries": 0, "seconds": 469.6, "max_memory_gb": 21.36, "pair": 0.5727393962442875, "sign": 0.7384098805487156, "keep": 0.1667806375771761, "leak": 0.4717054031789303, "replacement": 1.143261007964611}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 13.413191795349121, "loss_scale": 4.0, "retries": 0, "seconds": 470.8, "max_memory_gb": 21.36, "pair": 0.6526850610971451, "sign": 0.9980927398428321, "keep": 2.439792881719768, "leak": 0.6636446118354797, "replacement": 2.53753899037838}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 6.846108913421631, "loss_scale": 4.0, "retries": 0, "seconds": 471.9, "max_memory_gb": 21.36, "pair": 0.37412811536341906, "sign": 0.509739792207256, "keep": 0.6478135858196765, "leak": 0.645952878636308, "replacement": 1.1143930405378342}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 5.434689998626709, "loss_scale": 4.0, "retries": 0, "seconds": 473.1, "max_memory_gb": 21.36, "pair": 0.3788800295442343, "sign": 0.44687229581177235, "keep": 0.3098716721578967, "leak": 1.3695195070758928, "replacement": 1.0996318981051445}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 4.603961944580078, "loss_scale": 4.0, "retries": 0, "seconds": 474.3, "max_memory_gb": 21.36, "pair": 0.10911394841969013, "sign": 0.1135007431730628, "keep": 0.20773060945793986, "leak": 1.300983240827918, "replacement": 0.594975434243679}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 9.571776390075684, "loss_scale": 4.0, "retries": 0, "seconds": 475.5, "max_memory_gb": 21.36, "pair": 0.3860844671726227, "sign": 0.8039331249892712, "keep": 1.2713946968578966, "leak": 0.2680840604007244, "replacement": 1.4907693937420845}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 5.213472843170166, "loss_scale": 4.0, "retries": 0, "seconds": 476.6, "max_memory_gb": 21.36, "pair": 0.4478355422616005, "sign": 0.6621972173452377, "keep": 0.24450899608200416, "leak": 0.18526384024880826, "replacement": 0.9475046545267105}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 7.55253267288208, "loss_scale": 4.0, "retries": 0, "seconds": 477.8, "max_memory_gb": 21.36, "pair": 0.13498372212052345, "sign": 0.1443921346217394, "keep": 0.03466467000544071, "leak": 1.2851008613361046, "replacement": 0.5457873269915581}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 9.515116691589355, "loss_scale": 4.0, "retries": 0, "seconds": 478.9, "max_memory_gb": 21.36, "pair": 0.4486722256988287, "sign": 0.4038165600504726, "keep": 0.9668034538626671, "leak": 0.5387838571332395, "replacement": 1.2686782144010067}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 6.7600250244140625, "loss_scale": 4.0, "retries": 0, "seconds": 480.0, "max_memory_gb": 21.36, "pair": 0.4035658407956362, "sign": 0.3859913405030966, "keep": 1.5423593847081065, "leak": 0.3060213742501219, "replacement": 1.4442465081810951}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 4.348627090454102, "loss_scale": 4.0, "retries": 0, "seconds": 481.1, "max_memory_gb": 21.36, "pair": 0.21778167225420475, "sign": 0.14094869466498494, "keep": 0.07537212182069197, "leak": 0.3914092667400837, "replacement": 0.4237943924963474}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 12.509077072143555, "loss_scale": 4.0, "retries": 0, "seconds": 482.2, "max_memory_gb": 21.36, "pair": 0.8168941512703896, "sign": 1.2420172020792961, "keep": 1.6223255922086537, "leak": 0.6144653186202049, "replacement": 2.402681812644005}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 1.854705572128296, "loss_scale": 4.0, "retries": 0, "seconds": 483.4, "max_memory_gb": 21.36, "pair": 0.10678358748555183, "sign": 0.16561508364975452, "keep": 0.022034310270100832, "leak": 0.31839695386588573, "replacement": 0.2802075296640396}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 7.744966506958008, "loss_scale": 4.0, "retries": 0, "seconds": 484.4, "max_memory_gb": 21.36, "pair": 0.22460380010306835, "sign": 0.48402608186006546, "keep": 0.039941991737578064, "leak": 2.830811783671379, "replacement": 1.1942908242344856}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 9.14770221710205, "loss_scale": 4.0, "retries": 0, "seconds": 485.6, "max_memory_gb": 21.36, "pair": 0.2801984744146466, "sign": 0.4883853131905198, "keep": 0.973801143001765, "leak": 1.2364021092653275, "replacement": 1.3203922435641289}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 9.260637283325195, "loss_scale": 4.0, "retries": 0, "seconds": 486.7, "max_memory_gb": 21.36, "pair": 0.1433215495198965, "sign": 0.2042787680402398, "keep": 0.03780701069626957, "leak": 1.983435183763504, "replacement": 0.7602232396602631}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 5.022984504699707, "loss_scale": 4.0, "retries": 0, "seconds": 487.9, "max_memory_gb": 21.36, "pair": 0.3869985770434141, "sign": 0.48693792428821325, "keep": 0.5750971564557403, "leak": 0.3414347367361188, "replacement": 1.0033747777342796}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 12.13808822631836, "loss_scale": 4.0, "retries": 0, "seconds": 489.0, "max_memory_gb": 21.36, "pair": 0.6309831626713276, "sign": 0.9228912442922592, "keep": 1.1847301349043846, "leak": 0.5409139096736908, "replacement": 1.8200222626328468}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 4.199766635894775, "loss_scale": 4.0, "retries": 0, "seconds": 490.2, "max_memory_gb": 21.36, "pair": 0.15818964503705502, "sign": 0.08470848854631186, "keep": 0.0777676310390234, "leak": 0.36450292356312275, "replacement": 0.33055343478918076}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 4.024703502655029, "loss_scale": 4.0, "retries": 0, "seconds": 491.3, "max_memory_gb": 21.36, "pair": 0.21736076101660728, "sign": 0.180232722312212, "keep": 0.07171838614158332, "leak": 0.4576307905954309, "replacement": 0.4577440023422241}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 4.8911519050598145, "loss_scale": 4.0, "retries": 0, "seconds": 492.5, "max_memory_gb": 21.36, "pair": 0.4694239906966686, "sign": 0.5883419448509812, "keep": 0.3791309604421258, "leak": 0.513182021677494, "replacement": 1.081455945968628}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 4.809195518493652, "loss_scale": 4.0, "retries": 0, "seconds": 493.7, "max_memory_gb": 21.36, "pair": 0.21929273381829262, "sign": 0.26406180299818516, "keep": 0.11022600065916777, "leak": 1.913586363196373, "replacement": 0.8848332315683365}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 2.803251028060913, "loss_scale": 4.0, "retries": 0, "seconds": 494.9, "max_memory_gb": 21.36, "pair": 0.1971923839300871, "sign": 0.1154337003827095, "keep": 0.016279227333143353, "leak": 0.2774933120672358, "replacement": 0.3324221782386303}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 7.228039741516113, "loss_scale": 4.0, "retries": 0, "seconds": 496.0, "max_memory_gb": 21.36, "pair": 0.23090253211557865, "sign": 0.08385805319994688, "keep": 0.03228483448037878, "leak": 0.1932686511427164, "replacement": 0.3372911289334297}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 4.068978309631348, "loss_scale": 4.0, "retries": 0, "seconds": 497.2, "max_memory_gb": 21.36, "pair": 0.1838351972401142, "sign": 0.2134211235679686, "keep": 0.24796794611029327, "leak": 0.746089793741703, "replacement": 0.6010521873831749}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.75801944732666, "loss_scale": 4.0, "retries": 0, "seconds": 498.3, "max_memory_gb": 21.36, "pair": 0.13634523190557957, "sign": 0.06330772861838341, "keep": 0.20892231032485142, "leak": 0.918589930050075, "replacement": 0.5021077431738377}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 2.7887933254241943, "loss_scale": 4.0, "retries": 0, "seconds": 499.5, "max_memory_gb": 21.36, "pair": 0.14626025408506393, "sign": 0.17848990391939878, "keep": 0.10807664575986564, "leak": 0.5093641169369221, "replacement": 0.41688455641269684}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 3.5378828048706055, "loss_scale": 4.0, "retries": 0, "seconds": 500.6, "max_memory_gb": 21.36, "pair": 0.16570141166448593, "sign": 0.23434310220181942, "keep": 0.05262019531801343, "leak": 0.9681314155459404, "replacement": 0.5512159131467342}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 4.7047295570373535, "loss_scale": 4.0, "retries": 0, "seconds": 501.7, "max_memory_gb": 21.36, "pair": 0.16519033908843994, "sign": 0.13133670203387737, "keep": 0.015039628517115489, "leak": 0.21003783773630857, "replacement": 0.2908879593014717}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 10.450770378112793, "loss_scale": 4.0, "retries": 0, "seconds": 502.9, "max_memory_gb": 21.36, "pair": 0.5966737605631351, "sign": 0.4990228242240846, "keep": 0.5614577984670177, "leak": 0.495534336194396, "replacement": 1.250797614455223}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 3.598651170730591, "loss_scale": 4.0, "retries": 0, "seconds": 504.0, "max_memory_gb": 21.36, "pair": 0.17641553282737732, "sign": 0.20896986685693264, "keep": 0.11629027128219604, "leak": 0.5989103056490421, "replacement": 0.4887731745839119}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 8.866508483886719, "loss_scale": 4.0, "retries": 0, "seconds": 505.2, "max_memory_gb": 21.36, "pair": 0.30897964909672737, "sign": 0.4845826691016555, "keep": 0.5788633534684777, "leak": 0.8812179714441299, "replacement": 1.0610071271657944}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.4537509679794312, "loss_scale": 4.0, "retries": 0, "seconds": 506.3, "max_memory_gb": 21.36, "pair": 0.09701719367876649, "sign": 0.07887108065187931, "keep": 0.07824094183160923, "leak": 0.25310844743216876, "replacement": 0.2388503160327673}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 5.682249546051025, "loss_scale": 4.0, "retries": 0, "seconds": 507.5, "max_memory_gb": 21.36, "pair": 0.31199673376977444, "sign": 0.41931488970294595, "keep": 0.5907885897904634, "leak": 0.558388459845446, "replacement": 0.956645593047142}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 2.9267184734344482, "loss_scale": 4.0, "retries": 0, "seconds": 508.6, "max_memory_gb": 21.36, "pair": 0.22116070613265038, "sign": 0.32161586731672287, "keep": 0.01725642615929246, "leak": 0.2920451876707375, "replacement": 0.46360816061496735}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 10.492966651916504, "loss_scale": 4.0, "retries": 0, "seconds": 509.8, "max_memory_gb": 21.36, "pair": 0.657023224979639, "sign": 0.7404917608946562, "keep": 0.9070090983586852, "leak": 1.1051188632845879, "replacement": 1.7570533081889153}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 2.520250082015991, "loss_scale": 4.0, "retries": 0, "seconds": 510.9, "max_memory_gb": 21.36, "pair": 0.42447127774357796, "sign": 0.43817881308496, "keep": 0.07624968909658492, "leak": 0.7261887602508068, "replacement": 0.8632327392697334}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 2.7425315380096436, "loss_scale": 4.0, "retries": 0, "seconds": 512.1, "max_memory_gb": 21.36, "pair": 0.20507570821791887, "sign": 0.31103353574872017, "keep": 0.0431737809558399, "leak": 0.42898564599454403, "replacement": 0.48942576348781586}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 6.292919158935547, "loss_scale": 4.0, "retries": 0, "seconds": 513.2, "max_memory_gb": 21.36, "pair": 0.21609089523553848, "sign": 0.4198650764301419, "keep": 0.20813505817204714, "leak": 0.2750407326966524, "replacement": 0.5988511517643929}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 6.591038227081299, "loss_scale": 4.0, "retries": 0, "seconds": 514.4, "max_memory_gb": 21.36, "pair": 0.6344932122156024, "sign": 0.8653661329299212, "keep": 1.0325703092385083, "leak": 0.7456191112287343, "replacement": 1.7698661908507347}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 3.1940231323242188, "loss_scale": 4.0, "retries": 0, "seconds": 515.5, "max_memory_gb": 21.36, "pair": 0.16046405583620071, "sign": 0.1966842357069254, "keep": 0.299315053736791, "leak": 0.33076826014439575, "replacement": 0.49115578830242157}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 4.726510524749756, "loss_scale": 4.0, "retries": 0, "seconds": 516.7, "max_memory_gb": 21.36, "pair": 0.42655488662421703, "sign": 0.5742826852947474, "keep": 0.32806215184973553, "leak": 1.1018788495566696, "replacement": 1.1531970016658306}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 5.346139907836914, "loss_scale": 4.0, "retries": 0, "seconds": 517.8, "max_memory_gb": 21.36, "pair": 0.22892526350915432, "sign": 0.27462203055620193, "keep": 0.23016799369361252, "leak": 0.12838746135730617, "replacement": 0.5134171396493912}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 2.299520492553711, "loss_scale": 4.0, "retries": 0, "seconds": 518.9, "max_memory_gb": 21.36, "pair": 0.11485573463141918, "sign": 0.17013954184949398, "keep": 0.03706796039477922, "leak": 0.6824244186282158, "replacement": 0.38906558230519295}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 3.3041462898254395, "loss_scale": 4.0, "retries": 0, "seconds": 520.0, "max_memory_gb": 21.36, "pair": 0.22546981181949377, "sign": 0.1428917907178402, "keep": 0.1937553400057368, "leak": 0.5135232512839139, "replacement": 0.5221741795539856}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 3.9494900703430176, "loss_scale": 4.0, "retries": 0, "seconds": 521.2, "max_memory_gb": 21.36, "pair": 0.13181474059820175, "sign": 0.11789241898804903, "keep": 0.03351080967695452, "leak": 0.7523336838930845, "replacement": 0.39559976756572723}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 5.982326984405518, "loss_scale": 4.0, "retries": 0, "seconds": 522.3, "max_memory_gb": 21.36, "pair": 0.3473319299519062, "sign": 0.43875565007328987, "keep": 0.5282502700574696, "leak": 1.0472752396017313, "replacement": 1.0926537290215492}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 12.0418062210083, "loss_scale": 4.0, "retries": 0, "seconds": 523.4, "max_memory_gb": 21.36, "pair": 0.3475304916501045, "sign": 0.31638482585549355, "keep": 0.10612298524938524, "leak": 0.3829790958989179, "replacement": 0.654529195278883}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 8.279337882995605, "loss_scale": 4.0, "retries": 0, "seconds": 524.4, "max_memory_gb": 21.36, "pair": 0.2733459882438183, "sign": 0.22442507278174162, "keep": 0.44125575199723244, "leak": 0.4419075110927224, "replacement": 0.7166632786393166}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 5.845198631286621, "loss_scale": 4.0, "retries": 0, "seconds": 525.6, "max_memory_gb": 21.36, "pair": 0.18740174733102322, "sign": 0.23422100394964218, "keep": 0.30869069951586425, "leak": 1.206371145322919, "replacement": 0.7604504004120827}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 6.484842300415039, "loss_scale": 4.0, "retries": 0, "seconds": 526.7, "max_memory_gb": 21.36, "pair": 0.3263575490564108, "sign": 0.8080948079004884, "keep": 0.04131866618990898, "leak": 1.4007289323490113, "replacement": 1.1012464836239815}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 3.172844171524048, "loss_scale": 4.0, "retries": 0, "seconds": 527.9, "max_memory_gb": 21.36, "pair": 0.11215496994554996, "sign": 0.2071161698549986, "keep": 0.020499348887824453, "leak": 0.3193404357880354, "replacement": 0.30579783394932747}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 15.88509464263916, "loss_scale": 4.0, "retries": 0, "seconds": 529.0, "max_memory_gb": 21.36, "pair": 0.3928194008767605, "sign": 0.5374430511146784, "keep": 0.43365206895396113, "leak": 1.0324159285519272, "replacement": 1.1364709585905075}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 7.4041428565979, "loss_scale": 4.0, "retries": 0, "seconds": 530.2, "max_memory_gb": 21.36, "pair": 0.2627613991498947, "sign": 0.32141151931136847, "keep": 0.12482103030197322, "leak": 0.337746350094676, "replacement": 0.5703142434358597}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 3.138530969619751, "loss_scale": 4.0, "retries": 0, "seconds": 531.4, "max_memory_gb": 21.36, "pair": 0.21730299480259418, "sign": 0.25973211601376534, "keep": 0.0340721650281921, "leak": 0.3476434089243412, "replacement": 0.4511159807443619}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 9.125554084777832, "loss_scale": 4.0, "retries": 0, "seconds": 532.6, "max_memory_gb": 21.36, "pair": 0.2927124258130789, "sign": 0.6577857472002506, "keep": 0.6814026217907667, "leak": 0.703232710249722, "replacement": 1.1381148286163807}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 2.922652244567871, "loss_scale": 4.0, "retries": 0, "seconds": 533.7, "max_memory_gb": 21.36, "pair": 0.1603158377110958, "sign": 0.22042137011885643, "keep": 0.02810133300954476, "leak": 0.2831518417224288, "replacement": 0.3553651496767998}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 3.964736223220825, "loss_scale": 4.0, "retries": 0, "seconds": 534.9, "max_memory_gb": 21.36, "pair": 0.15686994791030884, "sign": 0.1926562562584877, "keep": 0.03245710168266669, "leak": 0.03461385547416285, "replacement": 0.278080090880394}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 3.697441816329956, "loss_scale": 4.0, "retries": 0, "seconds": 536.0, "max_memory_gb": 21.36, "pair": 0.15537234768271446, "sign": 0.19389893813058734, "keep": 0.062049519940046594, "leak": 0.849561233073473, "replacement": 0.49573688581585884}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 8.01469612121582, "loss_scale": 4.0, "retries": 0, "seconds": 537.2, "max_memory_gb": 21.36, "pair": 0.12474363669753075, "sign": 0.18158331234008074, "keep": 0.1576866233954206, "leak": 1.1801556162536144, "replacement": 0.5894174948334694}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 5.150734901428223, "loss_scale": 4.0, "retries": 0, "seconds": 538.4, "max_memory_gb": 21.36, "pair": 0.1911148466169834, "sign": 0.20783666241914034, "keep": 0.11630677327048033, "leak": 0.6463029338046908, "replacement": 0.514762282371521}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 5.038284778594971, "loss_scale": 4.0, "retries": 0, "seconds": 539.5, "max_memory_gb": 21.36, "pair": 0.16011845972388983, "sign": 0.13169599883258343, "keep": 0.045209853909909725, "leak": 1.0126745663583279, "replacement": 0.5017400272190571}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 4.8177809715271, "loss_scale": 4.0, "retries": 0, "seconds": 540.6, "max_memory_gb": 21.36, "pair": 0.16374776884913445, "sign": 0.1364045194350183, "keep": 0.022847899701446295, "leak": 0.10833021386815744, "replacement": 0.2704565264284611}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 4.681056022644043, "loss_scale": 4.0, "retries": 0, "seconds": 541.8, "max_memory_gb": 21.36, "pair": 0.2744504604488611, "sign": 0.1428979206830263, "keep": 0.35161583276931196, "leak": 0.9586938358843327, "replacement": 0.7613807767629623}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 3.4574551582336426, "loss_scale": 4.0, "retries": 0, "seconds": 542.9, "max_memory_gb": 21.36, "pair": 0.11568301357328892, "sign": 0.07290795398876071, "keep": 0.010278946021571755, "leak": 0.4139317898079753, "replacement": 0.2607594169676304}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 9.702428817749023, "loss_scale": 4.0, "retries": 0, "seconds": 544.1, "max_memory_gb": 21.36, "pair": 0.3598398296162486, "sign": 0.5434980597347021, "keep": 0.5330424709245563, "leak": 0.5563397593796253, "replacement": 1.0371949970722198}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 9.077548027038574, "loss_scale": 4.0, "retries": 0, "seconds": 545.2, "max_memory_gb": 21.36, "pair": 0.2590651046484709, "sign": 0.31027068896219134, "keep": 0.09983631002251059, "leak": 0.44552642945200205, "replacement": 0.5755002051591873}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 8.870269775390625, "loss_scale": 4.0, "retries": 0, "seconds": 546.4, "max_memory_gb": 21.36, "pair": 0.1772168166935444, "sign": 0.13281810004264116, "keep": 0.3437533221731428, "leak": 0.19732833094894886, "replacement": 0.46483462303876877}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 7.715615272521973, "loss_scale": 4.0, "retries": 0, "seconds": 547.5, "max_memory_gb": 21.36, "pair": 0.2454653149470687, "sign": 0.40591611713171005, "keep": 0.3902982786967186, "leak": 0.5176393091678619, "replacement": 0.7729823365807533}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 11.942307472229004, "loss_scale": 4.0, "retries": 0, "seconds": 548.7, "max_memory_gb": 21.36, "pair": 0.32741368655115366, "sign": 0.4807252362370491, "keep": 0.4971346147940494, "leak": 0.4565826468169689, "replacement": 0.930489256978035}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 2.851574420928955, "loss_scale": 4.0, "retries": 0, "seconds": 549.8, "max_memory_gb": 21.36, "pair": 0.15875163581222296, "sign": 0.2382999211549759, "keep": 0.12479468691162765, "leak": 1.1172825507819653, "replacement": 0.619619570672512}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 8.222089767456055, "loss_scale": 4.0, "retries": 0, "seconds": 550.9, "max_memory_gb": 21.36, "pair": 0.23143278155475855, "sign": 0.34480152232572436, "keep": 0.025234008819097653, "leak": 1.1333902939222753, "replacement": 0.6997981071472168}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 8.422516822814941, "loss_scale": 4.0, "retries": 0, "seconds": 552.1, "max_memory_gb": 21.36, "pair": 0.33108335826545954, "sign": 0.28727214923128486, "keep": 0.04146551003213972, "leak": 0.581071587279439, "replacement": 0.6407201066613197}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 5.71329402923584, "loss_scale": 4.0, "retries": 0, "seconds": 553.3, "max_memory_gb": 21.36, "pair": 0.45448138285428286, "sign": 0.48819673620164394, "keep": 0.33439949341118336, "leak": 0.11182068148627877, "replacement": 0.8937346562743187}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 3.7820115089416504, "loss_scale": 4.0, "retries": 0, "seconds": 554.4, "max_memory_gb": 21.36, "pair": 0.3542969450354576, "sign": 0.47970390506088734, "keep": 0.38094548135995865, "leak": 0.11299482511822134, "replacement": 0.8128703460097313}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 5.425405502319336, "loss_scale": 4.0, "retries": 0, "seconds": 555.6, "max_memory_gb": 21.36, "pair": 0.08614478167146444, "sign": 0.11725162155926228, "keep": 0.053671361922170036, "leak": 0.9816014198586345, "replacement": 0.41700662299990654}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 3.651526927947998, "loss_scale": 4.0, "retries": 0, "seconds": 556.6, "max_memory_gb": 21.36, "pair": 0.17411555722355843, "sign": 0.35936974361538887, "keep": 0.1969364754622802, "leak": 0.8328474415466189, "replacement": 0.6604805253446102}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 8.126781463623047, "loss_scale": 4.0, "retries": 0, "seconds": 557.8, "max_memory_gb": 21.36, "pair": 0.32104891538619995, "sign": 0.6759440861642361, "keep": 0.4442427158355713, "leak": 0.2510110803414136, "replacement": 0.9438951015472412}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 14.21174430847168, "loss_scale": 4.0, "retries": 0, "seconds": 559.0, "max_memory_gb": 21.36, "pair": 0.6283130533993244, "sign": 0.9815521575510502, "keep": 1.5006226394325495, "leak": 0.5468498878180981, "replacement": 2.006112888455391}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 11.28780746459961, "loss_scale": 4.0, "retries": 0, "seconds": 560.1, "max_memory_gb": 21.36, "pair": 0.24903398007154465, "sign": 0.5233221650123596, "keep": 0.0406965171569027, "leak": 1.0599518716335297, "replacement": 0.7960313074290752}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 8.357171058654785, "loss_scale": 4.0, "retries": 0, "seconds": 561.2, "max_memory_gb": 21.36, "pair": 0.34945218823850155, "sign": 0.6162229934707284, "keep": 0.7605261681601405, "leak": 0.48439817130565643, "replacement": 1.1589263081550598}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 2.7297260761260986, "loss_scale": 4.0, "retries": 0, "seconds": 562.3, "max_memory_gb": 21.36, "pair": 0.15766151808202267, "sign": 0.12959363590925932, "keep": 0.06850227748509496, "leak": 0.4753827229142189, "replacement": 0.375555157661438}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 9.02043628692627, "loss_scale": 4.0, "retries": 0, "seconds": 563.5, "max_memory_gb": 21.36, "pair": 0.2554700141772628, "sign": 0.39490577206015587, "keep": 0.1884000226855278, "leak": 0.8556002918630838, "replacement": 0.7610229998826981}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 9.219917297363281, "loss_scale": 4.0, "retries": 0, "seconds": 564.7, "max_memory_gb": 21.36, "pair": 0.13103993237018585, "sign": 0.24002020247280598, "keep": 0.45638861070619896, "leak": 1.4469941812567413, "replacement": 0.8409928604960442}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 8.856350898742676, "loss_scale": 4.0, "retries": 0, "seconds": 565.8, "max_memory_gb": 21.36, "pair": 0.2550597870722413, "sign": 0.3517201389186084, "keep": 0.025414570700377226, "leak": 1.2852857608813792, "replacement": 0.7649485841393471}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 11.364032745361328, "loss_scale": 4.0, "retries": 0, "seconds": 566.9, "max_memory_gb": 21.36, "pair": 0.129440494813025, "sign": 0.3599452469497919, "keep": 0.2896959760983009, "leak": 3.0975552573800087, "replacement": 1.2286498919129372}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 4.370754718780518, "loss_scale": 4.0, "retries": 0, "seconds": 568.1, "max_memory_gb": 21.36, "pair": 0.1531067956238985, "sign": 0.12095041573047638, "keep": 0.013728149409871548, "leak": 0.9330341555178165, "replacement": 0.453704621642828}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 10.533390998840332, "loss_scale": 4.0, "retries": 0, "seconds": 569.3, "max_memory_gb": 21.36, "pair": 0.5421921573579311, "sign": 0.810173362493515, "keep": 0.39144973739166744, "leak": 0.6564419195055962, "replacement": 1.307114154100418}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 5.009902000427246, "loss_scale": 4.0, "retries": 0, "seconds": 570.4, "max_memory_gb": 21.36, "pair": 0.061152401845902205, "sign": 0.2337140473537147, "keep": 0.04103042325004935, "leak": 1.1172257284633815, "replacement": 0.47783106565475464}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 5.713217735290527, "loss_scale": 4.0, "retries": 0, "seconds": 571.6, "max_memory_gb": 21.36, "pair": 0.10184311773627996, "sign": 0.2581705073826015, "keep": 0.02489810367114842, "leak": 0.9392059929668903, "replacement": 0.478178933262825}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 3.982637643814087, "loss_scale": 4.0, "retries": 0, "seconds": 572.7, "max_memory_gb": 21.36, "pair": 0.12151582632213831, "sign": 0.05794199276715517, "keep": 0.0902237863629125, "leak": 0.8744900301098824, "replacement": 0.41422123461961746}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 17.430084228515625, "loss_scale": 4.0, "retries": 0, "seconds": 573.9, "max_memory_gb": 21.36, "pair": 0.345119615085423, "sign": 0.762786727398634, "keep": 1.3876367677003145, "leak": 1.5383373732911423, "replacement": 1.80491578951478}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 5.613132476806641, "loss_scale": 4.0, "retries": 0, "seconds": 575.0, "max_memory_gb": 21.36, "pair": 0.43961510621011257, "sign": 0.41603224724531174, "keep": 0.3425784781575203, "leak": 0.8903980366885662, "replacement": 1.041519969701767}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 12.172733306884766, "loss_scale": 4.0, "retries": 0, "seconds": 576.1, "max_memory_gb": 21.36, "pair": 0.4603787478990853, "sign": 0.423069117590785, "keep": 0.4755824994063005, "leak": 0.42236241698265076, "replacement": 1.0152951590716839}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 13.693584442138672, "loss_scale": 4.0, "retries": 0, "seconds": 577.2, "max_memory_gb": 21.36, "pair": 0.3751449938863516, "sign": 0.566776409978047, "keep": 0.8293416466913186, "leak": 0.7534239972010255, "replacement": 1.2615600749850273}
diff --git a/runs/single_llava/replacement/L22_s13/adapter/README.md b/runs/single_llava/replacement/L22_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L22_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L22_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L22_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..653eb1948b838539f0b9ee085d43a2e16d821e1a
--- /dev/null
+++ b/runs/single_llava/replacement/L22_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.22.mlp.up_proj",
+    "model.language_model.layers.22.mlp.down_proj",
+    "model.language_model.layers.22.self_attn.q_proj",
+    "model.language_model.layers.22.self_attn.v_proj",
+    "model.language_model.layers.22.self_attn.k_proj",
+    "model.language_model.layers.22.self_attn.o_proj",
+    "model.language_model.layers.22.mlp.gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L22_s13/config.json b/runs/single_llava/replacement/L22_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2e91c92691bd3c09b70f49059a44415f1c95e253
--- /dev/null
+++ b/runs/single_llava/replacement/L22_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 22,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L22_s13/steps.jsonl b/runs/single_llava/replacement/L22_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..1338a2021238752da442ffaa30c6cf93239bb96c
--- /dev/null
+++ b/runs/single_llava/replacement/L22_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.21827401220798492, "loss_scale": 4.0, "retries": 0, "seconds": 1.3, "max_memory_gb": 19.6, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.36282867193222046, "loss_scale": 4.0, "retries": 0, "seconds": 2.3, "max_memory_gb": 19.6, "pair": 1.0988498330116272, "sign": 0.5367308147251606, "keep": 5.373994210269302, "leak": 3.8046042674100278e-06, "replacement": 4.0542131960392}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.6386926174163818, "loss_scale": 4.0, "retries": 0, "seconds": 3.4, "max_memory_gb": 19.64, "pair": 1.100272685289383, "sign": 0.5417513307183981, "keep": 10.804645895957947, "leak": 3.448211924705902e-05, "replacement": 6.77347993850708}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.5073074102401733, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 19.65, "pair": 1.0972115695476532, "sign": 0.472273170016706, "keep": 7.6726133562624454, "leak": 3.451119374631162e-05, "replacement": 5.169663429260254}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.2320050597190857, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 19.65, "pair": 1.0979028344154358, "sign": 0.7763271182775497, "keep": 0.1590296719223261, "leak": 4.503357686758136e-10, "replacement": 1.565581202507019}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.3231787383556366, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 19.65, "pair": 1.0976023077964783, "sign": 0.738011509180069, "keep": 2.9359959736466408, "leak": 1.1491787919482022e-05, "replacement": 2.9346089959144592}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.5870041251182556, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 19.65, "pair": 1.091917246580124, "sign": 0.547561576589942, "keep": 6.882066131569445, "leak": 5.524154863678632e-05, "replacement": 4.80674484372139}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.5132049918174744, "loss_scale": 4.0, "retries": 0, "seconds": 8.9, "max_memory_gb": 19.65, "pair": 1.0959461629390717, "sign": 0.5897704660892487, "keep": 7.391685023903847, "leak": 5.1626166282403574e-05, "replacement": 5.086686730384827}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.5005217790603638, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 19.65, "pair": 1.0966856181621552, "sign": 0.6194050619378686, "keep": 5.713660314679146, "leak": 2.348212429215124e-05, "replacement": 4.263224273920059}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.31804773211479187, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 19.65, "pair": 1.0928255915641785, "sign": 0.780846856534481, "keep": 3.272034913301468, "leak": 4.335988666070989e-05, "replacement": 3.119277447462082}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.5172642469406128, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 19.65, "pair": 1.0941901803016663, "sign": 0.454619063064456, "keep": 5.2704895585775375, "leak": 3.281848989900027e-05, "replacement": 3.956752598285675}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.34039464592933655, "loss_scale": 4.0, "retries": 0, "seconds": 13.2, "max_memory_gb": 19.65, "pair": 1.0924333333969116, "sign": 0.6114772707223892, "keep": 3.0614251592196524, "leak": 1.81185162091424e-05, "replacement": 2.9288889467716217}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.5506093502044678, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 19.65, "pair": 1.09315025806427, "sign": 0.5124420076608658, "keep": 5.356360588222742, "leak": 0.00014884740203147118, "replacement": 4.02758863568306}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.7876222729682922, "loss_scale": 4.0, "retries": 0, "seconds": 15.3, "max_memory_gb": 19.65, "pair": 1.0894046127796173, "sign": 0.3997959438711405, "keep": 7.794744320213795, "leak": 9.425414486941008e-05, "replacement": 5.186698228120804}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.5115678310394287, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 19.65, "pair": 1.0883879363536835, "sign": 0.6898387335240841, "keep": 5.195585459470749, "leak": 7.962656991367112e-05, "replacement": 4.0311199426651}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.39592409133911133, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 19.65, "pair": 1.0932626128196716, "sign": 0.7851675972342491, "keep": 2.768678106367588, "leak": 4.456775945982372e-05, "replacement": 2.870196580886841}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.3938556909561157, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 19.65, "pair": 1.0926975011825562, "sign": 0.7162787914276123, "keep": 3.1958674285560846, "leak": 0.00017248300810024375, "replacement": 3.048813760280609}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.7760584950447083, "loss_scale": 4.0, "retries": 0, "seconds": 19.6, "max_memory_gb": 19.69, "pair": 1.087206393480301, "sign": 0.5423915758728981, "keep": 7.206585116684437, "leak": 7.836497297830647e-05, "replacement": 4.961714297533035}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.2870208024978638, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 19.69, "pair": 1.0750096440315247, "sign": 0.542625367641449, "keep": 10.711562156677246, "leak": 0.0009425546522834338, "replacement": 6.702339053153992}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.3584109842777252, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 19.69, "pair": 1.0902361273765564, "sign": 0.43145086616277695, "keep": 0.024104315089061856, "leak": 9.837427933234721e-05, "replacement": 1.318038284778595}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.78764808177948, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 19.69, "pair": 1.079669862985611, "sign": 0.37693000212311745, "keep": 5.555431473068893, "leak": 0.00012706335883194697, "replacement": 4.045882225036621}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.7600749731063843, "loss_scale": 4.0, "retries": 0, "seconds": 24.0, "max_memory_gb": 19.69, "pair": 1.0818899273872375, "sign": 0.5756285935640335, "keep": 5.948131941258907, "leak": 0.0006658202860307938, "replacement": 4.3439366817474365}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.8534553647041321, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 19.69, "pair": 1.07257878780365, "sign": 0.2546280976384878, "keep": 5.319567944854498, "leak": 0.000662872300381423, "replacement": 3.8598426282405853}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.4668635129928589, "loss_scale": 4.0, "retries": 0, "seconds": 26.2, "max_memory_gb": 19.69, "pair": 1.0694679915904999, "sign": 0.26869038864970207, "keep": 10.068029284477234, "leak": 0.00043528524838620797, "replacement": 6.237936735153198}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.1665644645690918, "loss_scale": 4.0, "retries": 0, "seconds": 27.3, "max_memory_gb": 19.69, "pair": 1.0704871118068695, "sign": 0.39685091376304626, "keep": 8.479089468717575, "leak": 0.0008335725158430396, "replacement": 5.5086658000946045}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.4272059202194214, "loss_scale": 4.0, "retries": 0, "seconds": 28.4, "max_memory_gb": 19.69, "pair": 1.0618039667606354, "sign": 0.6637219041585922, "keep": 10.55212414264679, "leak": 0.0016902072384255007, "replacement": 6.6701496839523315}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.1326003074645996, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 19.69, "pair": 1.0775183737277985, "sign": 0.601575218141079, "keep": 8.450670793652534, "leak": 0.0009481056440563407, "replacement": 5.603878289461136}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.9304652214050293, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 19.69, "pair": 1.0718731582164764, "sign": 0.7829947341233492, "keep": 5.65893742442131, "leak": 0.0023585069495135613, "replacement": 4.2934287786483765}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 0.8470264673233032, "loss_scale": 4.0, "retries": 0, "seconds": 31.7, "max_memory_gb": 19.69, "pair": 1.077626347541809, "sign": 0.8358170483261347, "keep": 5.235740737058222, "leak": 0.0010301653965143487, "replacement": 4.113662868738174}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.041122555732727, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 19.69, "pair": 1.0573176443576813, "sign": 0.34935329016298056, "keep": 5.553970415145159, "leak": 0.002727651600707759, "replacement": 4.009661465883255}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.0467408895492554, "loss_scale": 4.0, "retries": 0, "seconds": 33.9, "max_memory_gb": 19.69, "pair": 1.0646196603775024, "sign": 0.8055845946073532, "keep": 5.723664551973343, "leak": 0.0021561449220826034, "replacement": 4.329783260822296}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.8760021924972534, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 19.69, "pair": 1.0496091842651367, "sign": 0.3983446769416332, "keep": 10.771084070205688, "leak": 0.005663961521349847, "replacement": 6.63573956489563}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 0.9864806532859802, "loss_scale": 4.0, "retries": 0, "seconds": 36.1, "max_memory_gb": 19.69, "pair": 1.0652150511741638, "sign": 0.6287090722471476, "keep": 4.92938172724098, "leak": 0.0026528103824043914, "replacement": 3.844923585653305}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.1421632766723633, "loss_scale": 4.0, "retries": 0, "seconds": 37.1, "max_memory_gb": 19.69, "pair": 1.054421067237854, "sign": 0.5318186655640602, "keep": 5.314850183203816, "leak": 0.002993598023749655, "replacement": 3.978504002094269}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.211608648300171, "loss_scale": 4.0, "retries": 0, "seconds": 38.2, "max_memory_gb": 19.69, "pair": 1.042170226573944, "sign": 0.4342209827154875, "keep": 5.534949406050146, "leak": 0.004529554336841102, "replacement": 4.027887850999832}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.1698497533798218, "loss_scale": 4.0, "retries": 0, "seconds": 39.3, "max_memory_gb": 19.69, "pair": 1.0453442931175232, "sign": 0.4385496638715267, "keep": 5.191076919436455, "leak": 0.003023596886123414, "replacement": 3.860913336277008}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.2987874746322632, "loss_scale": 4.0, "retries": 0, "seconds": 40.3, "max_memory_gb": 19.69, "pair": 1.047953486442566, "sign": 0.5911042913794518, "keep": 5.02898383140564, "leak": 0.0034255199100670097, "replacement": 3.858853906393051}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 2.1204993724823, "loss_scale": 4.0, "retries": 0, "seconds": 41.4, "max_memory_gb": 19.69, "pair": 1.0259131789207458, "sign": 0.7871599830687046, "keep": 9.529724836349487, "leak": 0.010303653602022678, "replacement": 6.186931490898132}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.896328330039978, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 19.69, "pair": 1.0562783777713776, "sign": 0.5631388053297997, "keep": 2.9300999687984586, "leak": 0.005312452085490804, "replacement": 2.80422580242157}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.2649723291397095, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 19.69, "pair": 1.0455373227596283, "sign": 0.5033796355128288, "keep": 5.550899088382721, "leak": 0.003978988839662634, "replacement": 4.07367131114006}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.6828874945640564, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 19.69, "pair": 1.081476092338562, "sign": 1.025972232222557, "keep": 2.9698561429977417, "leak": 0.002591430120711502, "replacement": 3.0800381004810333}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.4929006099700928, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 19.69, "pair": 1.0286138951778412, "sign": 0.4032723642885685, "keep": 4.790248030796647, "leak": 0.004033283912576735, "replacement": 3.6263823807239532}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 1.45932137966156, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 19.69, "pair": 1.0346806645393372, "sign": 0.517982043325901, "keep": 5.006622714921832, "leak": 0.0033830093598226085, "replacement": 3.7978287637233734}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 0.9349641799926758, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 19.69, "pair": 1.0461203157901764, "sign": 0.5817684307694435, "keep": 2.7906985580921173, "leak": 0.007413945277221501, "replacement": 2.734207361936569}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 0.8813662528991699, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 19.69, "pair": 1.0633102357387543, "sign": 0.8930447325110435, "keep": 2.8594807237386703, "leak": 0.003668230704590769, "replacement": 2.94049009680748}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.5067983865737915, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 19.69, "pair": 1.0225342512130737, "sign": 0.7250798307359219, "keep": 4.657495558261871, "leak": 0.014419223205095477, "replacement": 3.717426687479019}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 1.430254578590393, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 19.69, "pair": 0.999305009841919, "sign": 0.6914528422057629, "keep": 4.181411092635244, "leak": 0.011516774280607933, "replacement": 3.4386161863803864}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 2.453305959701538, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 19.69, "pair": 0.9611286371946335, "sign": 0.2058368492871523, "keep": 7.751987420022488, "leak": 0.025339370360597968, "replacement": 4.946375489234924}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 2.1420035362243652, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 19.69, "pair": 0.9976001232862473, "sign": 0.5980876516550779, "keep": 7.338552251458168, "leak": 0.01100515597499907, "replacement": 4.968671381473541}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 2.400963068008423, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 19.69, "pair": 1.0057522505521774, "sign": 0.4538526386022568, "keep": 7.4320983393117785, "leak": 0.012440817314200103, "replacement": 4.951837986707687}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.004052758216858, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 19.69, "pair": 1.0343399941921234, "sign": 0.5831253454089165, "keep": 2.769472320564091, "leak": 0.010427981014800025, "replacement": 2.7132457494735718}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 1.7172200679779053, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 19.69, "pair": 1.0120113641023636, "sign": 0.7091199532151222, "keep": 4.838810957968235, "leak": 0.010349589021643624, "replacement": 3.7885642051696777}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 2.5349438190460205, "loss_scale": 4.0, "retries": 0, "seconds": 57.0, "max_memory_gb": 19.69, "pair": 0.9998449832201004, "sign": 0.7682584151625633, "keep": 7.590638816356659, "leak": 0.033121103542725905, "replacement": 5.187573879957199}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 2.478639841079712, "loss_scale": 4.0, "retries": 0, "seconds": 58.1, "max_memory_gb": 19.69, "pair": 0.9769615978002548, "sign": 0.6513910591602325, "keep": 7.982729941606522, "leak": 0.03646893202676438, "replacement": 5.3031392097473145}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 1.9267281293869019, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 19.69, "pair": 0.994696855545044, "sign": 0.4760551005601883, "keep": 5.049231108743697, "leak": 0.025331849785288796, "replacement": 3.7636728286743164}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 3.6993660926818848, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 19.69, "pair": 0.9026085734367371, "sign": 0.2019138839095831, "keep": 9.826736211776733, "leak": 0.026635099202394485, "replacement": 5.9235923290252686}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.074583888053894, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 19.69, "pair": 1.0212906002998352, "sign": 0.7888341061770916, "keep": 2.5265528270974755, "leak": 0.007476304948795587, "replacement": 2.6808530390262604}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 2.0279605388641357, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 19.69, "pair": 0.9658686965703964, "sign": 0.2819015998393297, "keep": 4.94162594305817, "leak": 0.05038533796323463, "replacement": 3.590228796005249}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.272412657737732, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 19.69, "pair": 1.0353639423847198, "sign": 0.6690678223967552, "keep": 2.847230613231659, "leak": 0.002857658535845964, "replacement": 2.7942275404930115}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 1.9467196464538574, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 19.69, "pair": 0.9726015329360962, "sign": 0.6346973739564419, "keep": 4.9323542239144444, "leak": 0.011144210875499994, "replacement": 3.758913427591324}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 2.859941005706787, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 19.69, "pair": 0.9449798911809921, "sign": 0.6353709753602743, "keep": 7.20356023311615, "leak": 0.04225401087751379, "replacement": 4.875009000301361}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.2521318197250366, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 19.69, "pair": 1.0080243796110153, "sign": 0.7165469788014889, "keep": 2.7451660092920065, "leak": 0.009184530244965572, "replacement": 2.74117711186409}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 2.9965310096740723, "loss_scale": 4.0, "retries": 0, "seconds": 67.7, "max_memory_gb": 19.69, "pair": 1.0016444325447083, "sign": 0.6019442304968834, "keep": 7.510532078333199, "leak": 0.016971080240182346, "replacement": 5.062125355005264}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 3.3062541484832764, "loss_scale": 4.0, "retries": 0, "seconds": 68.8, "max_memory_gb": 19.72, "pair": 0.9255415499210358, "sign": 0.5044884411618114, "keep": 7.446355087682605, "leak": 0.07575802414794452, "replacement": 4.9199027717113495}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 2.986992359161377, "loss_scale": 4.0, "retries": 0, "seconds": 69.9, "max_memory_gb": 19.72, "pair": 0.9445035755634308, "sign": 0.567790050059557, "keep": 7.576854556798935, "leak": 0.04994304290448781, "replacement": 5.0293115973472595}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 0.9684534072875977, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 19.72, "pair": 1.0407598614692688, "sign": 0.6407578215003014, "keep": 1.9026122929062694, "leak": 0.006694558640447212, "replacement": 2.314118653535843}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 3.3692121505737305, "loss_scale": 4.0, "retries": 0, "seconds": 72.0, "max_memory_gb": 19.72, "pair": 0.8466807454824448, "sign": 0.38971228897571564, "keep": 7.091333478689194, "leak": 0.08866486102078852, "replacement": 4.609369933605194}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.2932318449020386, "loss_scale": 4.0, "retries": 0, "seconds": 73.1, "max_memory_gb": 19.72, "pair": 0.9989123493432999, "sign": 0.5068860054016113, "keep": 2.4608304055873305, "leak": 0.023205273709208996, "replacement": 2.48857182264328}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 2.351226568222046, "loss_scale": 4.0, "retries": 0, "seconds": 74.2, "max_memory_gb": 19.72, "pair": 0.9406886398792267, "sign": 0.7579131349921227, "keep": 4.629569103941321, "leak": 0.01877502153297428, "replacement": 3.6391235888004303}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 2.4586503505706787, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 19.72, "pair": 0.9579570293426514, "sign": 0.6535549759864807, "keep": 4.728093781508505, "leak": 0.06239166067825863, "replacement": 3.664379298686981}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 3.795180082321167, "loss_scale": 4.0, "retries": 0, "seconds": 76.3, "max_memory_gb": 19.72, "pair": 0.8112309873104095, "sign": 0.18623326160013676, "keep": 7.00801351852715, "leak": 0.09024794935248792, "replacement": 4.430916458368301}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 4.206559658050537, "loss_scale": 4.0, "retries": 0, "seconds": 77.5, "max_memory_gb": 19.72, "pair": 0.8218519389629364, "sign": 0.31730057671666145, "keep": 8.120225429534912, "leak": 0.07185232918709517, "replacement": 5.05857789516449}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.7041163444519043, "loss_scale": 4.0, "retries": 0, "seconds": 78.5, "max_memory_gb": 19.72, "pair": 1.0454820692539215, "sign": 0.8111248537898064, "keep": 0.4380194819532335, "leak": 0.003623886578282054, "replacement": 1.670960247516632}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.507996082305908, "loss_scale": 4.0, "retries": 0, "seconds": 79.6, "max_memory_gb": 19.72, "pair": 0.9141182750463486, "sign": 0.5058811195194721, "keep": 4.681253492832184, "leak": 0.05090013702010765, "replacement": 3.5204105973243713}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.6174894571304321, "loss_scale": 4.0, "retries": 0, "seconds": 80.6, "max_memory_gb": 19.72, "pair": 0.9617003053426743, "sign": 0.9044117145240307, "keep": 3.412693217396736, "leak": 0.062048279834016284, "replacement": 3.1357648074626923}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 2.434119939804077, "loss_scale": 4.0, "retries": 0, "seconds": 81.7, "max_memory_gb": 19.72, "pair": 0.8750949949026108, "sign": 0.45677399169653654, "keep": 4.387466806918383, "leak": 0.08964264567475766, "replacement": 3.319626033306122}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 3.808978796005249, "loss_scale": 4.0, "retries": 0, "seconds": 82.7, "max_memory_gb": 19.72, "pair": 0.8420888781547546, "sign": 0.4981135744601488, "keep": 7.205746650695801, "leak": 0.213227433465363, "replacement": 4.747325658798218}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 2.4666223526000977, "loss_scale": 4.0, "retries": 0, "seconds": 83.8, "max_memory_gb": 19.72, "pair": 0.9264040738344193, "sign": 0.738337330520153, "keep": 3.7878333516418934, "leak": 0.13835181067315716, "replacement": 3.2240774035453796}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.5858746767044067, "loss_scale": 4.0, "retries": 0, "seconds": 84.8, "max_memory_gb": 19.72, "pair": 0.9492457509040833, "sign": 0.5855162627995014, "keep": 1.9594716420397162, "leak": 0.03574875650519971, "replacement": 2.2306768596172333}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 2.6120121479034424, "loss_scale": 4.0, "retries": 0, "seconds": 85.9, "max_memory_gb": 19.72, "pair": 0.8441849201917648, "sign": 0.3794603571295738, "keep": 4.294302839087322, "leak": 0.15296743507497013, "replacement": 3.2193083465099335}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 4.061007499694824, "loss_scale": 4.0, "retries": 0, "seconds": 86.9, "max_memory_gb": 19.72, "pair": 0.8012431114912033, "sign": 0.5523610040545464, "keep": 6.311735473573208, "leak": 0.1535785110463621, "replacement": 4.27168595790863}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 4.503197193145752, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 19.72, "pair": 0.795691043138504, "sign": 0.3906004149466753, "keep": 6.281546376645565, "leak": 0.1251842679685069, "replacement": 4.163060426712036}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 3.697981595993042, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 19.72, "pair": 0.7885397225618362, "sign": 0.5959604829549789, "keep": 5.495441567152739, "leak": 0.18637504009529948, "replacement": 3.8808343708515167}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.7843343019485474, "loss_scale": 4.0, "retries": 0, "seconds": 90.2, "max_memory_gb": 19.72, "pair": 0.9525207579135895, "sign": 0.8415795788168907, "keep": 2.308938290691003, "leak": 0.0898177201161161, "replacement": 2.550234079360962}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.887539803981781, "loss_scale": 4.0, "retries": 0, "seconds": 91.3, "max_memory_gb": 19.72, "pair": 1.008851319551468, "sign": 0.579473964869976, "keep": 0.15057705156505108, "leak": 0.011917857824300881, "replacement": 1.3768563568592072}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 4.295867919921875, "loss_scale": 4.0, "retries": 0, "seconds": 92.4, "max_memory_gb": 19.72, "pair": 0.8855354636907578, "sign": 0.7515526302158833, "keep": 5.9298195987939835, "leak": 0.13206961071045953, "replacement": 4.259239137172699}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 4.318226337432861, "loss_scale": 4.0, "retries": 0, "seconds": 93.4, "max_memory_gb": 19.72, "pair": 0.8001737147569656, "sign": 0.5103698875755072, "keep": 6.00930181145668, "leak": 0.3315700995735824, "replacement": 4.142902135848999}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 3.3896665573120117, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 19.72, "pair": 0.8667624369263649, "sign": 0.6377326641231775, "keep": 4.02760008815676, "leak": 0.1352658224462857, "replacement": 3.233245223760605}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.7985671162605286, "loss_scale": 4.0, "retries": 0, "seconds": 95.5, "max_memory_gb": 19.72, "pair": 1.0101073384284973, "sign": 0.6994777843356133, "keep": 0.21819131670054048, "leak": 0.014690102240820124, "replacement": 1.47261443734169}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 3.2275891304016113, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 19.72, "pair": 0.8619315326213837, "sign": 0.631865419447422, "keep": 3.3693986763246357, "leak": 0.07124703648150899, "replacement": 2.880375415086746}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.5384202003479004, "loss_scale": 4.0, "retries": 0, "seconds": 97.7, "max_memory_gb": 19.72, "pair": 1.0543963611125946, "sign": 1.022382527589798, "keep": 0.631535716354847, "leak": 0.0035686951482603035, "replacement": 1.882247656583786}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.0703253746032715, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 19.72, "pair": 0.9396661669015884, "sign": 0.6401507630944252, "keep": 1.9831535457633436, "leak": 0.2216681478312239, "replacement": 2.3067353665828705}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 3.647904396057129, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 19.72, "pair": 0.7543030306696892, "sign": 0.306764492765069, "keep": 3.505926350597292, "leak": 0.1916411097627133, "replacement": 2.708558738231659}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 5.463846683502197, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 19.72, "pair": 0.7133538126945496, "sign": 0.4069956596940756, "keep": 5.923389360308647, "leak": 0.21882003991322563, "replacement": 3.9332513213157654}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 3.5884132385253906, "loss_scale": 4.0, "retries": 0, "seconds": 102.0, "max_memory_gb": 19.72, "pair": 0.7989285737276077, "sign": 0.512617664411664, "keep": 3.925603376701474, "leak": 0.12689775347462273, "replacement": 3.0497635304927826}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 5.553086757659912, "loss_scale": 4.0, "retries": 0, "seconds": 103.1, "max_memory_gb": 19.72, "pair": 0.6938196271657944, "sign": 0.5084367133677006, "keep": 5.767749294638634, "leak": 0.5069249889952516, "replacement": 3.9586437940597534}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.124635696411133, "loss_scale": 4.0, "retries": 0, "seconds": 104.2, "max_memory_gb": 19.72, "pair": 0.8686315417289734, "sign": 0.5390169285237789, "keep": 2.341592188924551, "leak": 0.17319412634242326, "replacement": 2.3522347509860992}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 3.7944045066833496, "loss_scale": 4.0, "retries": 0, "seconds": 105.2, "max_memory_gb": 19.72, "pair": 0.8517959266901016, "sign": 0.5340351946651936, "keep": 3.4270413825288415, "leak": 0.2804309665225446, "replacement": 2.902442067861557}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.6209195852279663, "loss_scale": 4.0, "retries": 0, "seconds": 106.3, "max_memory_gb": 19.72, "pair": 1.015088677406311, "sign": 0.8744811713695526, "keep": 0.4126913635991514, "leak": 0.007801707513863221, "replacement": 1.660625398159027}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 5.102719306945801, "loss_scale": 4.0, "retries": 0, "seconds": 107.4, "max_memory_gb": 19.72, "pair": 0.7848903834819794, "sign": 0.7075117565691471, "keep": 5.206940047442913, "leak": 0.2780181516063749, "replacement": 3.8116208910942078}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 5.775768280029297, "loss_scale": 4.0, "retries": 0, "seconds": 108.5, "max_memory_gb": 19.72, "pair": 0.5617048069834709, "sign": 0.17951529240235686, "keep": 5.092654153704643, "leak": 0.23857448203489184, "replacement": 3.2574331760406494}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 5.926365852355957, "loss_scale": 4.0, "retries": 0, "seconds": 109.6, "max_memory_gb": 19.72, "pair": 0.701550155878067, "sign": 0.35669897496700287, "keep": 5.417045325040817, "leak": 0.2862731082714163, "replacement": 3.6599906384944916}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 4.241710662841797, "loss_scale": 4.0, "retries": 0, "seconds": 110.7, "max_memory_gb": 19.72, "pair": 0.7854948192834854, "sign": 0.398441806435585, "keep": 3.1998703062999994, "leak": 0.2786156840156764, "replacement": 2.6543048918247223}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.6070728302001953, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 19.72, "pair": 0.9656609892845154, "sign": 0.9375515058636665, "keep": 1.7739812445361167, "leak": 0.10006137809250504, "replacement": 2.34644278883934}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 5.640140056610107, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 19.72, "pair": 0.6787950545549393, "sign": 0.4075595699250698, "keep": 4.703527685254812, "leak": 0.8015991080319509, "replacement": 3.4347383975982666}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 6.309455871582031, "loss_scale": 4.0, "retries": 0, "seconds": 113.9, "max_memory_gb": 19.72, "pair": 0.7230034023523331, "sign": 0.6306890230625868, "keep": 4.457393303513527, "leak": 0.7396334955410566, "replacement": 3.451952964067459}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.6101675033569336, "loss_scale": 4.0, "retries": 0, "seconds": 115.0, "max_memory_gb": 19.72, "pair": 0.9240071177482605, "sign": 0.8517160639166832, "keep": 1.7227573245763779, "leak": 0.2424342781305313, "replacement": 2.271852433681488}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.3382408618927, "loss_scale": 4.0, "retries": 0, "seconds": 116.0, "max_memory_gb": 19.72, "pair": 0.8713590055704117, "sign": 0.7154879420995712, "keep": 1.7418579757213593, "leak": 0.4469600999218528, "replacement": 2.2117719650268555}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 6.677730083465576, "loss_scale": 4.0, "retries": 0, "seconds": 117.1, "max_memory_gb": 19.72, "pair": 0.564897395670414, "sign": 0.44763090182095766, "keep": 4.3940697610378265, "leak": 1.1421617759278888, "replacement": 3.271288275718689}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 5.167717933654785, "loss_scale": 4.0, "retries": 0, "seconds": 118.2, "max_memory_gb": 19.72, "pair": 0.7650102600455284, "sign": 0.5486353263258934, "keep": 3.091362565755844, "leak": 0.154238736264233, "replacement": 2.623568832874298}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 4.808498382568359, "loss_scale": 4.0, "retries": 0, "seconds": 119.2, "max_memory_gb": 19.72, "pair": 0.6794262453913689, "sign": 0.4397299876436591, "keep": 2.5551127353683114, "leak": 0.18801229260861874, "replacement": 2.2238506376743317}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 4.937582492828369, "loss_scale": 4.0, "retries": 0, "seconds": 120.3, "max_memory_gb": 19.72, "pair": 0.790260449051857, "sign": 0.48502933606505394, "keep": 2.8187653212808073, "leak": 0.21855192724615335, "replacement": 2.496795803308487}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 6.039135932922363, "loss_scale": 4.0, "retries": 0, "seconds": 121.4, "max_memory_gb": 19.72, "pair": 0.7358993142843246, "sign": 0.5564802326261997, "keep": 2.876349799335003, "leak": 0.2912067389661388, "replacement": 2.525116056203842}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 7.357533931732178, "loss_scale": 4.0, "retries": 0, "seconds": 122.5, "max_memory_gb": 19.72, "pair": 0.6517917290329933, "sign": 0.5288138762116432, "keep": 4.188626766204834, "leak": 0.8587008113972843, "replacement": 3.2251872420310974}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 4.730763912200928, "loss_scale": 4.0, "retries": 0, "seconds": 123.6, "max_memory_gb": 19.72, "pair": 0.7469134852290154, "sign": 0.5762656964361668, "keep": 2.7934385240077972, "leak": 0.670682510266488, "replacement": 2.5994362235069275}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.9172454476356506, "loss_scale": 4.0, "retries": 0, "seconds": 124.6, "max_memory_gb": 19.72, "pair": 0.9664271175861359, "sign": 0.7094531133770943, "keep": 0.2304022484458983, "leak": 0.025651607453255565, "replacement": 1.442767709493637}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 5.087007522583008, "loss_scale": 4.0, "retries": 0, "seconds": 125.7, "max_memory_gb": 19.72, "pair": 0.7165934890508652, "sign": 0.45285073295235634, "keep": 2.5750553894322366, "leak": 0.429591606836766, "replacement": 2.337944447994232}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 2.8976809978485107, "loss_scale": 4.0, "retries": 0, "seconds": 126.7, "max_memory_gb": 19.72, "pair": 0.7875602170825005, "sign": 0.5410196809098125, "keep": 1.4957685410045087, "leak": 0.27191373095365634, "replacement": 1.8739327490329742}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 3.0494301319122314, "loss_scale": 4.0, "retries": 0, "seconds": 127.8, "max_memory_gb": 19.72, "pair": 0.8581762164831161, "sign": 0.7871287614107132, "keep": 1.159433638677001, "leak": 0.24689730201498605, "replacement": 1.893181711435318}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 5.236283302307129, "loss_scale": 4.0, "retries": 0, "seconds": 128.9, "max_memory_gb": 19.72, "pair": 0.687077596783638, "sign": 0.5253013726323843, "keep": 2.3017945173196495, "leak": 0.48371646262239665, "replacement": 2.221554696559906}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.319505453109741, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 19.72, "pair": 0.861788809299469, "sign": 0.7538435682654381, "keep": 1.6481688721105456, "leak": 1.1386663889861666, "replacement": 2.3474615812301636}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 5.131263732910156, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 19.72, "pair": 0.7401293739676476, "sign": 0.6954122725874186, "keep": 1.6934504231903702, "leak": 0.39521053972202935, "replacement": 2.0333633422851562}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 3.0630242824554443, "loss_scale": 4.0, "retries": 0, "seconds": 132.2, "max_memory_gb": 19.72, "pair": 0.8563762381672859, "sign": 0.7933380380272865, "keep": 1.81360624730587, "leak": 0.2343583722540643, "replacement": 2.218437999486923}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 3.584606885910034, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 19.72, "pair": 0.953281819820404, "sign": 1.0040638744831085, "keep": 0.9872814789414406, "leak": 0.012648999923840165, "replacement": 1.9521167278289795}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 3.4409420490264893, "loss_scale": 4.0, "retries": 0, "seconds": 134.3, "max_memory_gb": 19.72, "pair": 0.8889951407909393, "sign": 0.826347604393959, "keep": 0.9691947754472494, "leak": 0.10937860092963092, "replacement": 1.8141109645366669}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 5.874814033508301, "loss_scale": 4.0, "retries": 0, "seconds": 135.4, "max_memory_gb": 19.72, "pair": 0.697861485183239, "sign": 0.6163752134889364, "keep": 2.188108891248703, "leak": 0.1395572665446707, "replacement": 2.134992778301239}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 3.0515592098236084, "loss_scale": 4.0, "retries": 0, "seconds": 136.4, "max_memory_gb": 19.72, "pair": 0.7132179327309132, "sign": 0.38744067400693893, "keep": 0.6479048498440534, "leak": 0.2972576208412647, "replacement": 1.305205076932907}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 5.534583568572998, "loss_scale": 4.0, "retries": 0, "seconds": 137.5, "max_memory_gb": 19.72, "pair": 0.581725649535656, "sign": 0.34314935142174363, "keep": 1.3303177400957793, "leak": 0.9178381953388453, "replacement": 1.6479187309741974}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 9.406667709350586, "loss_scale": 4.0, "retries": 0, "seconds": 138.6, "max_memory_gb": 19.72, "pair": 0.4596419706940651, "sign": 0.2650827653706074, "keep": 2.0394947826862335, "leak": 1.6267440989613533, "replacement": 2.018616795539856}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 6.5970563888549805, "loss_scale": 4.0, "retries": 0, "seconds": 139.7, "max_memory_gb": 19.72, "pair": 0.7394865080714226, "sign": 0.8476404100656509, "keep": 1.9868374839425087, "leak": 0.8045228514238261, "replacement": 2.3578562140464783}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 3.0063846111297607, "loss_scale": 4.0, "retries": 0, "seconds": 140.7, "max_memory_gb": 19.72, "pair": 0.8433028906583786, "sign": 0.7191199883818626, "keep": 0.974275678396225, "leak": 0.12860374117327233, "replacement": 1.7221516370773315}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 6.87430477142334, "loss_scale": 4.0, "retries": 0, "seconds": 141.8, "max_memory_gb": 19.72, "pair": 0.4644690304994583, "sign": 0.14031630661338568, "keep": 2.0958276903256774, "leak": 0.3305564168840647, "replacement": 1.6651801019906998}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 6.1949286460876465, "loss_scale": 4.0, "retries": 0, "seconds": 142.9, "max_memory_gb": 19.72, "pair": 0.5225663222372532, "sign": 0.4126110188663006, "keep": 1.4958427250385284, "leak": 2.10503458463063, "replacement": 2.0030518770217896}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 1.044476866722107, "loss_scale": 4.0, "retries": 0, "seconds": 144.0, "max_memory_gb": 19.72, "pair": 0.8981849402189255, "sign": 0.7114098742604256, "keep": 0.3718949081376195, "leak": 0.08692553573928308, "replacement": 1.4615687429904938}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 6.188833713531494, "loss_scale": 4.0, "retries": 0, "seconds": 145.0, "max_memory_gb": 19.72, "pair": 0.47520552948117256, "sign": 0.4487670389935374, "keep": 1.2461220026016235, "leak": 2.59607039031107, "replacement": 1.9716676771640778}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 3.2569363117218018, "loss_scale": 4.0, "retries": 0, "seconds": 146.1, "max_memory_gb": 19.72, "pair": 0.6484036445617676, "sign": 0.4747921684756875, "keep": 0.6511236764490604, "leak": 0.06472444554674439, "replacement": 1.2275426983833313}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 3.3599419593811035, "loss_scale": 4.0, "retries": 0, "seconds": 147.2, "max_memory_gb": 19.72, "pair": 0.6332733854651451, "sign": 0.4709006156772375, "keep": 0.8763485364615917, "leak": 2.0443335961026605, "replacement": 1.8179814219474792}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 4.749118328094482, "loss_scale": 4.0, "retries": 0, "seconds": 148.3, "max_memory_gb": 19.72, "pair": 0.6028258055448532, "sign": 0.3211522502824664, "keep": 0.7104445435106754, "leak": 0.09998434106819332, "replacement": 1.143620252609253}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 6.230330467224121, "loss_scale": 4.0, "retries": 0, "seconds": 149.4, "max_memory_gb": 19.72, "pair": 0.6048314198851585, "sign": 0.6577744781970978, "keep": 1.3029930964112282, "leak": 0.9983020320069045, "replacement": 1.8347906470298767}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 3.7341480255126953, "loss_scale": 4.0, "retries": 0, "seconds": 150.5, "max_memory_gb": 19.72, "pair": 0.6727612465620041, "sign": 0.6112961452454329, "keep": 0.39733144268393517, "leak": 0.9499477166682482, "replacement": 1.414562001824379}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 3.3570871353149414, "loss_scale": 4.0, "retries": 0, "seconds": 151.6, "max_memory_gb": 19.72, "pair": 0.6317089796066284, "sign": 0.6648860652931035, "keep": 0.3948752610012889, "leak": 2.167455663322471, "replacement": 1.703453540802002}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.5871758460998535, "loss_scale": 4.0, "retries": 0, "seconds": 152.7, "max_memory_gb": 19.72, "pair": 0.6045303046703339, "sign": 0.41501977294683456, "keep": 0.4028054620139301, "leak": 0.34260051287738236, "replacement": 1.099093034863472}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.9165555238723755, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 19.72, "pair": 0.9588912278413773, "sign": 1.0195207670331001, "keep": 0.24347775022033602, "leak": 0.07023206082135403, "replacement": 1.6079485416412354}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 3.694408416748047, "loss_scale": 4.0, "retries": 0, "seconds": 154.9, "max_memory_gb": 19.72, "pair": 0.5453804358839989, "sign": 0.4007372558116913, "keep": 0.6499989069998264, "leak": 1.5447648517147172, "replacement": 1.4569397866725922}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 3.2264175415039062, "loss_scale": 4.0, "retries": 0, "seconds": 156.0, "max_memory_gb": 19.72, "pair": 0.4816708527505398, "sign": 0.2937806979753077, "keep": 0.2819530952256173, "leak": 1.2961021475493908, "replacement": 1.093563288450241}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 3.1610846519470215, "loss_scale": 4.0, "retries": 0, "seconds": 157.1, "max_memory_gb": 19.72, "pair": 0.5956125520169735, "sign": 0.5525420689955354, "keep": 0.380904009565711, "leak": 1.199579262174666, "replacement": 1.3622304201126099}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.6446635723114014, "loss_scale": 4.0, "retries": 0, "seconds": 158.2, "max_memory_gb": 19.72, "pair": 0.6107249781489372, "sign": 0.5256206262856722, "keep": 0.4769330443814397, "leak": 0.8408250795328058, "replacement": 1.322208121418953}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.5475865602493286, "loss_scale": 4.0, "retries": 0, "seconds": 159.2, "max_memory_gb": 19.72, "pair": 0.6368871107697487, "sign": 0.32038337364792824, "keep": 0.10747863235883415, "leak": 0.19464581832289696, "replacement": 0.8994795978069305}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 1.3612488508224487, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 19.72, "pair": 0.703035481274128, "sign": 0.518455570563674, "keep": 0.2553590298630297, "leak": 0.0718066340778023, "replacement": 1.10789442807436}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 3.5473599433898926, "loss_scale": 4.0, "retries": 0, "seconds": 161.4, "max_memory_gb": 19.72, "pair": 0.4080688841640949, "sign": 0.3388898875564337, "keep": 0.486225001513958, "leak": 3.0776358080329373, "replacement": 1.5900352746248245}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 2.607003688812256, "loss_scale": 4.0, "retries": 0, "seconds": 162.5, "max_memory_gb": 19.72, "pair": 0.7058972865343094, "sign": 0.4613342210650444, "keep": 0.4006753540597856, "leak": 0.1574480589479208, "replacement": 1.1762640476226807}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 3.5355701446533203, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 19.72, "pair": 0.37950874492526054, "sign": 0.2616642341017723, "keep": 0.1821915078908205, "leak": 2.967048367485404, "replacement": 1.34319868683815}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 2.721421957015991, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 19.72, "pair": 0.4905477464199066, "sign": 0.29058967158198357, "keep": 0.10076599847525358, "leak": 1.9317690543830395, "replacement": 1.169167846441269}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 3.0769875049591064, "loss_scale": 4.0, "retries": 0, "seconds": 165.8, "max_memory_gb": 19.72, "pair": 0.4369095265865326, "sign": 0.23951582796871662, "keep": 0.8964753467589617, "leak": 1.1601767954416573, "replacement": 1.2949493303894997}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.8830976486206055, "loss_scale": 4.0, "retries": 0, "seconds": 166.9, "max_memory_gb": 19.72, "pair": 0.6156540438532829, "sign": 0.5817537922412157, "keep": 1.2567794960923493, "leak": 1.7764985114336014, "replacement": 1.9790453165769577}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.9284385442733765, "loss_scale": 4.0, "retries": 0, "seconds": 168.0, "max_memory_gb": 19.72, "pair": 0.6699632816016674, "sign": 0.5563134662806988, "keep": 0.312463978305459, "leak": 0.5587460988026578, "replacement": 1.2440384924411774}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 2.242655038833618, "loss_scale": 4.0, "retries": 0, "seconds": 169.1, "max_memory_gb": 19.72, "pair": 0.9153204262256622, "sign": 1.188806876540184, "keep": 0.3231075629591942, "leak": 0.08031823113560677, "replacement": 1.6913572251796722}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.2626245021820068, "loss_scale": 4.0, "retries": 0, "seconds": 170.1, "max_memory_gb": 19.72, "pair": 0.7444278560578823, "sign": 0.7088922187685966, "keep": 0.4736059121787548, "leak": 0.20398680624202825, "replacement": 1.3866735994815826}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 3.177424669265747, "loss_scale": 4.0, "retries": 0, "seconds": 171.3, "max_memory_gb": 19.72, "pair": 0.448124997317791, "sign": 0.2128786575049162, "keep": 0.0906513319350779, "leak": 2.0083261355757713, "replacement": 1.1019715517759323}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 2.873924970626831, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 19.72, "pair": 0.440137829631567, "sign": 0.35378763591870666, "keep": 0.22403058409690857, "leak": 1.9110304564128455, "replacement": 1.2068045511841774}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 1.9114176034927368, "loss_scale": 4.0, "retries": 0, "seconds": 173.4, "max_memory_gb": 19.72, "pair": 0.6940792091190815, "sign": 0.6136101335287094, "keep": 0.36329890601336956, "leak": 0.6115049513973645, "replacement": 1.3354099839925766}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 1.3259226083755493, "loss_scale": 4.0, "retries": 0, "seconds": 174.5, "max_memory_gb": 19.72, "pair": 0.6531927473843098, "sign": 0.6072738580405712, "keep": 0.12661117315292358, "leak": 0.2415310386568308, "replacement": 1.0805180072784424}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 3.6166772842407227, "loss_scale": 4.0, "retries": 0, "seconds": 175.5, "max_memory_gb": 19.72, "pair": 0.4192679189145565, "sign": 0.33350085094571114, "keep": 0.12085145246237516, "leak": 1.0389276421628892, "replacement": 0.9061759859323502}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 2.969087839126587, "loss_scale": 4.0, "retries": 0, "seconds": 176.6, "max_memory_gb": 19.72, "pair": 0.7176646254956722, "sign": 0.869399013929069, "keep": 0.4672650396823883, "leak": 0.8382422177237459, "replacement": 1.5955572128295898}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 2.9669134616851807, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 19.72, "pair": 0.7100745029747486, "sign": 0.6556495986878872, "keep": 0.28803339321166277, "leak": 1.5677531627006829, "replacement": 1.5738542973995209}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 2.2255687713623047, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 19.72, "pair": 0.6255413293838501, "sign": 0.5537921041250229, "keep": 0.2758697848767042, "leak": 1.089199746507802, "replacement": 1.312672197818756}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 2.5052177906036377, "loss_scale": 4.0, "retries": 0, "seconds": 179.8, "max_memory_gb": 19.72, "pair": 0.4305178225040436, "sign": 0.40437750704586506, "keep": 0.4833050803281367, "leak": 1.5402661866974086, "replacement": 1.25942562520504}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.9498376846313477, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 19.72, "pair": 0.5660595186054707, "sign": 0.503658015280962, "keep": 0.1800467437133193, "leak": 1.5899015590548515, "replacement": 1.305387258529663}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 3.979184865951538, "loss_scale": 4.0, "retries": 0, "seconds": 182.0, "max_memory_gb": 19.72, "pair": 0.4452265128493309, "sign": 0.4604194890707731, "keep": 0.6740074176341295, "leak": 2.375836078543216, "replacement": 1.6063989698886871}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 2.6707870960235596, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 19.72, "pair": 0.35049325227737427, "sign": 0.16118282545357943, "keep": 0.039066229946911335, "leak": 1.7223028391599655, "replacement": 0.8811934590339661}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 2.918332576751709, "loss_scale": 4.0, "retries": 0, "seconds": 184.2, "max_memory_gb": 19.72, "pair": 0.5158771239221096, "sign": 0.4059705650433898, "keep": 0.07345925713889301, "leak": 1.719451555982232, "replacement": 1.185454934835434}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 5.5089497566223145, "loss_scale": 4.0, "retries": 0, "seconds": 185.2, "max_memory_gb": 19.72, "pair": 0.25595754384994507, "sign": 0.09505743812769651, "keep": 0.13741296576336026, "leak": 4.35425466299057, "replacement": 1.46075639128685}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 5.038496494293213, "loss_scale": 4.0, "retries": 0, "seconds": 186.2, "max_memory_gb": 19.72, "pair": 0.4361249953508377, "sign": 0.4387437626719475, "keep": 0.3081532698124647, "leak": 2.0714846216142178, "replacement": 1.3274446427822113}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 1.7386903762817383, "loss_scale": 4.0, "retries": 0, "seconds": 187.3, "max_memory_gb": 19.72, "pair": 0.5257331058382988, "sign": 0.32096587866544724, "keep": 0.2810765369795263, "leak": 0.7313472144305706, "replacement": 1.0095910876989365}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 3.378901958465576, "loss_scale": 4.0, "retries": 0, "seconds": 188.3, "max_memory_gb": 19.72, "pair": 0.5550391897559166, "sign": 0.5666480902582407, "keep": 0.3173931841738522, "leak": 1.6795339023228735, "replacement": 1.416943296790123}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 5.132011890411377, "loss_scale": 4.0, "retries": 0, "seconds": 189.4, "max_memory_gb": 19.72, "pair": 0.31433913111686707, "sign": 0.2402433082461357, "keep": 0.04586149984970689, "leak": 2.8153168857097626, "replacement": 1.161220744252205}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 1.5739672183990479, "loss_scale": 4.0, "retries": 0, "seconds": 190.4, "max_memory_gb": 19.72, "pair": 0.4640921391546726, "sign": 0.32735538855195045, "keep": 0.28009174671024084, "leak": 0.44309774900466437, "replacement": 0.8785901889204979}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.262540817260742, "loss_scale": 4.0, "retries": 0, "seconds": 191.5, "max_memory_gb": 19.72, "pair": 0.6857286170125008, "sign": 0.6051150113344193, "keep": 0.3647501387167722, "leak": 1.525389825925231, "replacement": 1.5520086586475372}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 1.5567032098770142, "loss_scale": 4.0, "retries": 0, "seconds": 192.5, "max_memory_gb": 19.72, "pair": 0.7344949245452881, "sign": 0.7693889662623405, "keep": 0.6017391378991306, "leak": 0.6301858984206774, "replacement": 1.5776054412126541}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 2.5874826908111572, "loss_scale": 4.0, "retries": 0, "seconds": 193.6, "max_memory_gb": 19.72, "pair": 0.591648280620575, "sign": 0.5001062480732799, "keep": 0.5919928904622793, "leak": 0.3514420051360503, "replacement": 1.2255583629012108}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 3.269671678543091, "loss_scale": 4.0, "retries": 0, "seconds": 194.7, "max_memory_gb": 19.72, "pair": 0.32358845323324203, "sign": 0.1795138856396079, "keep": 0.09901895001530647, "leak": 1.9845576913794503, "replacement": 0.9589943364262581}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 1.61197829246521, "loss_scale": 4.0, "retries": 0, "seconds": 195.8, "max_memory_gb": 19.72, "pair": 0.718233585357666, "sign": 0.6085522994399071, "keep": 0.2296126449946314, "leak": 0.5476551016326994, "replacement": 1.2742298543453217}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.7700417041778564, "loss_scale": 4.0, "retries": 0, "seconds": 196.9, "max_memory_gb": 19.72, "pair": 0.4164930172264576, "sign": 0.17609059158712626, "keep": 0.02080676646437496, "leak": 0.9038641080260277, "replacement": 0.7409077435731888}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 3.202472686767578, "loss_scale": 4.0, "retries": 0, "seconds": 198.0, "max_memory_gb": 19.72, "pair": 0.5093562304973602, "sign": 0.3152730776928365, "keep": 0.06745848420541734, "leak": 1.2370665771886706, "replacement": 1.009988695383072}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 3.3255763053894043, "loss_scale": 4.0, "retries": 0, "seconds": 199.1, "max_memory_gb": 19.72, "pair": 0.41529905423521996, "sign": 0.28424966987222433, "keep": 0.04296148242428899, "leak": 2.7800610456615686, "replacement": 1.2739198505878448}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 1.7189258337020874, "loss_scale": 4.0, "retries": 0, "seconds": 200.1, "max_memory_gb": 19.72, "pair": 0.47228164970874786, "sign": 0.24964974261820316, "keep": 0.08133346773684025, "leak": 0.4256494417786598, "replacement": 0.7441856116056442}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 3.5031886100769043, "loss_scale": 4.0, "retries": 0, "seconds": 201.3, "max_memory_gb": 19.72, "pair": 0.3940686583518982, "sign": 0.09516472462564707, "keep": 0.21266506100073457, "leak": 0.8794062961824238, "replacement": 0.7678351402282715}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 1.1025278568267822, "loss_scale": 4.0, "retries": 0, "seconds": 202.3, "max_memory_gb": 19.72, "pair": 0.7404449880123138, "sign": 0.6492502950131893, "keep": 0.27035294787492603, "leak": 0.22403128983569331, "replacement": 1.25625441968441}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.4647510051727295, "loss_scale": 4.0, "retries": 0, "seconds": 203.4, "max_memory_gb": 19.72, "pair": 0.40138981118798256, "sign": 0.2422154676169157, "keep": 0.04936636844649911, "leak": 1.0727147534489632, "replacement": 0.8153593987226486}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 3.6099557876586914, "loss_scale": 4.0, "retries": 0, "seconds": 204.6, "max_memory_gb": 19.72, "pair": 0.2494201846420765, "sign": 0.14797094278037548, "keep": 0.033699785359203815, "leak": 1.4714873172342777, "replacement": 0.7081273645162582}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 1.4457875490188599, "loss_scale": 4.0, "retries": 0, "seconds": 205.7, "max_memory_gb": 19.72, "pair": 0.6986477300524712, "sign": 0.6239438112825155, "keep": 0.593351268209517, "leak": 0.31984745524823666, "replacement": 1.3872571736574173}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 2.529139518737793, "loss_scale": 4.0, "retries": 0, "seconds": 206.7, "max_memory_gb": 19.72, "pair": 0.676074206829071, "sign": 0.5298372879624367, "keep": 0.36076833587139845, "leak": 0.5607680183566117, "replacement": 1.2615690678358078}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 1.922540307044983, "loss_scale": 4.0, "retries": 0, "seconds": 207.8, "max_memory_gb": 19.72, "pair": 0.4475688934326172, "sign": 0.2095533199608326, "keep": 0.05118467484135181, "leak": 0.5444104745984077, "replacement": 0.7140404880046844}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 1.6995943784713745, "loss_scale": 4.0, "retries": 0, "seconds": 208.8, "max_memory_gb": 19.72, "pair": 0.7148268148303032, "sign": 0.7578691467642784, "keep": 0.581749468576163, "leak": 0.5007591061294079, "replacement": 1.5098259001970291}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 3.307910203933716, "loss_scale": 4.0, "retries": 0, "seconds": 210.0, "max_memory_gb": 19.72, "pair": 0.5268727093935013, "sign": 0.36910721752792597, "keep": 0.17899882746860385, "leak": 1.3444892594998237, "replacement": 1.1370480358600616}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 4.155867099761963, "loss_scale": 4.0, "retries": 0, "seconds": 211.2, "max_memory_gb": 19.72, "pair": 0.24575135856866837, "sign": 0.11753485212102532, "keep": 0.05160882486961782, "leak": 1.4989480823278427, "replacement": 0.705060213804245}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 2.4635210037231445, "loss_scale": 4.0, "retries": 0, "seconds": 212.3, "max_memory_gb": 19.72, "pair": 0.6403587237000465, "sign": 0.5752233937382698, "keep": 0.22265773452818394, "leak": 0.5430264440365136, "replacement": 1.1750558763742447}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 1.8261324167251587, "loss_scale": 4.0, "retries": 0, "seconds": 213.3, "max_memory_gb": 19.72, "pair": 0.636540062725544, "sign": 0.6138433963060379, "keep": 0.3270901623181999, "leak": 0.63423230079934, "replacement": 1.2655649483203888}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.291814088821411, "loss_scale": 4.0, "retries": 0, "seconds": 214.4, "max_memory_gb": 19.72, "pair": 0.3847927041351795, "sign": 0.25361074693500996, "keep": 0.16425672266632318, "leak": 1.004747398197651, "replacement": 0.8449133038520813}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 1.2897815704345703, "loss_scale": 4.0, "retries": 0, "seconds": 215.5, "max_memory_gb": 19.72, "pair": 0.7444909736514091, "sign": 0.7207529526203871, "keep": 0.47075405437499285, "leak": 0.19032835317329955, "replacement": 1.387826532125473}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 1.8677500486373901, "loss_scale": 4.0, "retries": 0, "seconds": 216.6, "max_memory_gb": 19.72, "pair": 0.6074994914233685, "sign": 0.6641314122825861, "keep": 0.3088848479092121, "leak": 0.5389940738677979, "replacement": 1.2287560775876045}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 2.3855040073394775, "loss_scale": 4.0, "retries": 0, "seconds": 217.6, "max_memory_gb": 19.72, "pair": 0.5295582413673401, "sign": 0.3990187495946884, "keep": 0.005564114602748305, "leak": 0.43287935853004456, "replacement": 0.8400695323944092}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 1.7077546119689941, "loss_scale": 4.0, "retries": 0, "seconds": 218.7, "max_memory_gb": 19.72, "pair": 0.5151975490152836, "sign": 0.5216904580593109, "keep": 0.29404291696846485, "leak": 0.7186568570905365, "replacement": 1.1027284562587738}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 4.1089982986450195, "loss_scale": 4.0, "retries": 0, "seconds": 219.7, "max_memory_gb": 19.72, "pair": 0.6085083559155464, "sign": 0.5235611703246832, "keep": 0.22475913725793362, "leak": 0.5710425723809749, "replacement": 1.1254291832447052}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 2.5926856994628906, "loss_scale": 4.0, "retries": 0, "seconds": 220.8, "max_memory_gb": 19.72, "pair": 0.4393158629536629, "sign": 0.43829482793807983, "keep": 0.18636962218442932, "leak": 1.0155943613499403, "replacement": 1.0055466890335083}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 3.1055219173431396, "loss_scale": 4.0, "retries": 0, "seconds": 221.9, "max_memory_gb": 19.72, "pair": 0.46661243215203285, "sign": 0.540656540542841, "keep": 0.16290877177380025, "leak": 1.6262212470173836, "replacement": 1.2249503880739212}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 2.456057548522949, "loss_scale": 4.0, "retries": 0, "seconds": 223.0, "max_memory_gb": 19.72, "pair": 0.9211079180240631, "sign": 1.016754925251007, "keep": 1.4546753880567849, "leak": 0.1887336407089606, "replacement": 2.204006567597389}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 3.7495062351226807, "loss_scale": 4.0, "retries": 0, "seconds": 224.1, "max_memory_gb": 19.72, "pair": 0.2655314914882183, "sign": 0.09534858772531152, "keep": 0.026128396159037948, "leak": 1.5808249339461327, "replacement": 0.7214762195944786}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 1.961362361907959, "loss_scale": 4.0, "retries": 0, "seconds": 225.2, "max_memory_gb": 19.72, "pair": 0.6299049369990826, "sign": 0.6376357693225145, "keep": 0.40564341098070145, "leak": 0.6537191530223936, "replacement": 1.3149743378162384}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 1.8643115758895874, "loss_scale": 4.0, "retries": 0, "seconds": 226.3, "max_memory_gb": 19.72, "pair": 0.5348451733589172, "sign": 0.5327038131654263, "keep": 0.4653823649859987, "leak": 1.0197702893055975, "replacement": 1.2888308465480804}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 2.440568447113037, "loss_scale": 4.0, "retries": 0, "seconds": 227.4, "max_memory_gb": 19.72, "pair": 0.5587106607854366, "sign": 0.5296529079787433, "keep": 0.2978000286966562, "leak": 0.6397610598360188, "replacement": 1.1323773860931396}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 2.478548288345337, "loss_scale": 4.0, "retries": 0, "seconds": 228.5, "max_memory_gb": 19.72, "pair": 0.6001117005944252, "sign": 0.7854717522859573, "keep": 0.6159188523888588, "leak": 0.22102042520418763, "replacement": 1.356062114238739}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 3.694347381591797, "loss_scale": 4.0, "retries": 0, "seconds": 229.6, "max_memory_gb": 19.72, "pair": 0.5188647657632828, "sign": 0.48084072954952717, "keep": 0.5907429084181786, "leak": 1.4748684070073068, "replacement": 1.4233736991882324}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 4.768309593200684, "loss_scale": 4.0, "retries": 0, "seconds": 230.7, "max_memory_gb": 19.72, "pair": 0.6998720094561577, "sign": 0.693548783659935, "keep": 1.081603528233245, "leak": 0.20315201848279685, "replacement": 1.6382361054420471}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 2.101442575454712, "loss_scale": 4.0, "retries": 0, "seconds": 231.8, "max_memory_gb": 19.72, "pair": 0.7751860246062279, "sign": 0.9156597852706909, "keep": 0.8623602159204893, "leak": 0.3832645110960584, "replacement": 1.7600121647119522}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.4250153303146362, "loss_scale": 4.0, "retries": 0, "seconds": 232.9, "max_memory_gb": 19.72, "pair": 0.4724300019443035, "sign": 0.36756937205791473, "keep": 0.48394676693715155, "leak": 0.31043360475450754, "replacement": 0.9757964536547661}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 3.2418572902679443, "loss_scale": 4.0, "retries": 0, "seconds": 234.0, "max_memory_gb": 19.72, "pair": 0.3297121748328209, "sign": 0.24889895413070917, "keep": 0.11115784465800971, "leak": 1.261565200984478, "replacement": 0.8251318335533142}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 2.1048169136047363, "loss_scale": 4.0, "retries": 0, "seconds": 235.0, "max_memory_gb": 19.72, "pair": 0.8004484474658966, "sign": 1.0302804708480835, "keep": 0.29392528266180307, "leak": 0.2605503093218431, "replacement": 1.5276888608932495}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 1.8188563585281372, "loss_scale": 4.0, "retries": 0, "seconds": 236.1, "max_memory_gb": 19.72, "pair": 0.6443998292088509, "sign": 0.5984517633914948, "keep": 0.2094397689215839, "leak": 0.2593214947846718, "replacement": 1.113175943493843}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 3.749356746673584, "loss_scale": 4.0, "retries": 0, "seconds": 237.2, "max_memory_gb": 19.72, "pair": 0.42856721580028534, "sign": 0.4951852560043335, "keep": 0.26734645786928013, "leak": 0.628080777823925, "replacement": 0.9668533056974411}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 3.9201712608337402, "loss_scale": 4.0, "retries": 0, "seconds": 238.3, "max_memory_gb": 19.72, "pair": 0.38687166571617126, "sign": 0.30600312910974026, "keep": 0.24192925821989775, "leak": 0.966520678717643, "replacement": 0.9024680107831955}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.791966199874878, "loss_scale": 4.0, "retries": 0, "seconds": 239.4, "max_memory_gb": 19.72, "pair": 0.3549182489514351, "sign": 0.19856658205389977, "keep": 0.45559096056967974, "leak": 1.4334892297629267, "replacement": 1.0403692871332169}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 1.2656702995300293, "loss_scale": 4.0, "retries": 0, "seconds": 240.5, "max_memory_gb": 19.72, "pair": 0.6063526645302773, "sign": 0.6733219847083092, "keep": 0.4056796688819304, "leak": 0.41837792748992797, "replacement": 1.2504480332136154}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 2.2374813556671143, "loss_scale": 4.0, "retries": 0, "seconds": 241.6, "max_memory_gb": 19.72, "pair": 0.4999691806733608, "sign": 0.5127230640500784, "keep": 0.21858217136468738, "leak": 0.7762839458882809, "replacement": 1.0596927478909492}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 2.4770915508270264, "loss_scale": 4.0, "retries": 0, "seconds": 242.7, "max_memory_gb": 19.72, "pair": 0.5342228636145592, "sign": 0.5285667702555656, "keep": 0.201190163847059, "leak": 0.943391351174796, "replacement": 1.1349491626024246}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 3.0196001529693604, "loss_scale": 4.0, "retries": 0, "seconds": 243.8, "max_memory_gb": 19.72, "pair": 0.44100888073444366, "sign": 0.4054477699100971, "keep": 0.3070883802138269, "leak": 0.8297727443277836, "replacement": 1.0047201216220856}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.684234380722046, "loss_scale": 4.0, "retries": 0, "seconds": 244.8, "max_memory_gb": 19.72, "pair": 0.3266170509159565, "sign": 0.16950441617518663, "keep": 0.02236326228012331, "leak": 0.9221429228782654, "replacement": 0.6530866175889969}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 2.498838186264038, "loss_scale": 4.0, "retries": 0, "seconds": 245.9, "max_memory_gb": 19.72, "pair": 0.4581635966897011, "sign": 0.4831850528717041, "keep": 0.10570487566292286, "leak": 0.6920220851898193, "replacement": 0.9256141036748886}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 2.65183162689209, "loss_scale": 4.0, "retries": 0, "seconds": 247.0, "max_memory_gb": 19.72, "pair": 0.5626347325742245, "sign": 0.6738497242331505, "keep": 0.5093127135187387, "leak": 0.7433524411171675, "replacement": 1.3400540798902512}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 2.078535795211792, "loss_scale": 4.0, "retries": 0, "seconds": 247.9, "max_memory_gb": 19.72, "pair": 0.7287441492080688, "sign": 0.7171195447444916, "keep": 0.670092350745108, "leak": 0.18714496633037925, "replacement": 1.4691363275051117}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 1.785610556602478, "loss_scale": 4.0, "retries": 0, "seconds": 249.0, "max_memory_gb": 19.72, "pair": 0.6816822290420532, "sign": 0.7975067049264908, "keep": 0.46343281166628003, "leak": 0.3436188604682684, "replacement": 1.39805668592453}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 4.8159308433532715, "loss_scale": 4.0, "retries": 0, "seconds": 250.1, "max_memory_gb": 19.72, "pair": 0.42539408802986145, "sign": 0.5908128153532743, "keep": 0.040010749362409115, "leak": 0.9312786087393761, "replacement": 0.973625473678112}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 3.8990421295166016, "loss_scale": 4.0, "retries": 0, "seconds": 251.1, "max_memory_gb": 19.72, "pair": 0.41680416092276573, "sign": 0.19664495810866356, "keep": 0.3360972870141268, "leak": 1.2147409170866013, "replacement": 0.9868604987859726}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.061372756958008, "loss_scale": 4.0, "retries": 0, "seconds": 252.2, "max_memory_gb": 19.72, "pair": 0.4302521664649248, "sign": 0.4697474054992199, "keep": 0.39527403988176957, "leak": 1.8212784174829721, "replacement": 1.318082481622696}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 1.9340978860855103, "loss_scale": 4.0, "retries": 0, "seconds": 253.3, "max_memory_gb": 19.72, "pair": 0.5340532287955284, "sign": 0.6025886237621307, "keep": 0.5442557046189904, "leak": 0.49045753199607134, "replacement": 1.2300898432731628}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 2.138751983642578, "loss_scale": 4.0, "retries": 0, "seconds": 254.4, "max_memory_gb": 19.72, "pair": 0.5517639070749283, "sign": 0.4899633936583996, "keep": 0.17115328862564638, "leak": 0.6056710171453688, "replacement": 1.0337399989366531}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 3.138674736022949, "loss_scale": 4.0, "retries": 0, "seconds": 255.4, "max_memory_gb": 19.72, "pair": 0.5594804584980011, "sign": 0.4683200791478157, "keep": 0.2512885386822745, "leak": 0.4921043159556575, "replacement": 1.0423108339309692}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 2.189049243927002, "loss_scale": 4.0, "retries": 0, "seconds": 256.5, "max_memory_gb": 19.72, "pair": 0.3454539403319359, "sign": 0.24954804126173258, "keep": 0.10451398603618145, "leak": 0.48573256145755295, "replacement": 0.6439180970191956}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 3.0873167514801025, "loss_scale": 4.0, "retries": 0, "seconds": 257.6, "max_memory_gb": 19.72, "pair": 0.4960791915655136, "sign": 0.6159750260412693, "keep": 0.43922425247728825, "leak": 1.517613511532545, "replacement": 1.4030821323394775}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.3054027557373047, "loss_scale": 4.0, "retries": 0, "seconds": 258.6, "max_memory_gb": 19.72, "pair": 0.39152196794748306, "sign": 0.3353394567966461, "keep": 0.17574704228900373, "leak": 0.6046267375349998, "replacement": 0.7982219010591507}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 1.8317835330963135, "loss_scale": 4.0, "retries": 0, "seconds": 259.7, "max_memory_gb": 19.72, "pair": 0.6381551772356033, "sign": 0.7355040647089481, "keep": 0.33885750640183687, "leak": 0.3656642222777009, "replacement": 1.2667520493268967}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 1.782407283782959, "loss_scale": 4.0, "retries": 0, "seconds": 260.8, "max_memory_gb": 19.72, "pair": 0.5741188898682594, "sign": 0.6326486291363835, "keep": 0.3598567155422643, "leak": 0.3734718835912645, "replacement": 1.1637395471334457}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 2.8406741619110107, "loss_scale": 4.0, "retries": 0, "seconds": 261.9, "max_memory_gb": 19.72, "pair": 0.2606836911290884, "sign": 0.3121389066800475, "keep": 0.14520408469252288, "leak": 1.618000090122223, "replacement": 0.8938551843166351}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 3.2815895080566406, "loss_scale": 4.0, "retries": 0, "seconds": 263.0, "max_memory_gb": 19.72, "pair": 0.5876971650868654, "sign": 0.790026543661952, "keep": 0.820924652274698, "leak": 0.33810320124030113, "replacement": 1.4776984974741936}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 3.8985724449157715, "loss_scale": 4.0, "retries": 0, "seconds": 264.1, "max_memory_gb": 19.72, "pair": 0.8509881645441055, "sign": 1.0122637003660202, "keep": 0.6755392956547439, "leak": 0.6815542550757527, "replacement": 1.8652782142162323}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 2.8837807178497314, "loss_scale": 4.0, "retries": 0, "seconds": 265.2, "max_memory_gb": 19.72, "pair": 0.2922776583582163, "sign": 0.20452639926224947, "keep": 0.11026797047816217, "leak": 0.9415900558233261, "replacement": 0.6850723773241043}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 1.8020793199539185, "loss_scale": 4.0, "retries": 0, "seconds": 266.3, "max_memory_gb": 19.72, "pair": 0.4240947812795639, "sign": 0.4241131395101547, "keep": 0.23314236925216392, "leak": 0.5841288343071938, "replacement": 0.8987547606229782}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.2763428688049316, "loss_scale": 4.0, "retries": 0, "seconds": 267.3, "max_memory_gb": 19.72, "pair": 0.2994946949183941, "sign": 0.1849162308499217, "keep": 0.14184296689927578, "leak": 0.47352556698024273, "replacement": 0.5812556892633438}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 4.370803356170654, "loss_scale": 4.0, "retries": 0, "seconds": 268.4, "max_memory_gb": 19.72, "pair": 0.6161376684904099, "sign": 0.8631465137004852, "keep": 0.11621412634849548, "leak": 0.5530224475078285, "replacement": 1.2440735697746277}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 4.540993690490723, "loss_scale": 4.0, "retries": 0, "seconds": 269.5, "max_memory_gb": 19.72, "pair": 0.2353682741522789, "sign": 0.2239762879908085, "keep": 0.22691554098855704, "leak": 1.7354744374752045, "replacement": 0.8946827799081802}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 2.3862411975860596, "loss_scale": 4.0, "retries": 0, "seconds": 270.7, "max_memory_gb": 19.72, "pair": 0.4172469265758991, "sign": 0.4472149983048439, "keep": 0.22668489895295352, "leak": 1.0435010022483766, "replacement": 1.0150721371173859}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 2.4862723350524902, "loss_scale": 4.0, "retries": 0, "seconds": 271.7, "max_memory_gb": 19.72, "pair": 0.3021260052919388, "sign": 0.23027156479656696, "keep": 0.026067076716572046, "leak": 0.6459957994520664, "replacement": 0.5917942672967911}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 3.219541549682617, "loss_scale": 4.0, "retries": 0, "seconds": 272.9, "max_memory_gb": 19.72, "pair": 0.34170858189463615, "sign": 0.32314572390168905, "keep": 0.1743937471183017, "leak": 0.9007780808024108, "replacement": 0.8156728520989418}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 2.7549753189086914, "loss_scale": 4.0, "retries": 0, "seconds": 273.9, "max_memory_gb": 19.72, "pair": 0.4474090337753296, "sign": 0.5693939533084631, "keep": 0.4935929588973522, "leak": 0.7455922290682793, "replacement": 1.1653005629777908}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 1.9195138216018677, "loss_scale": 4.0, "retries": 0, "seconds": 275.0, "max_memory_gb": 19.72, "pair": 0.6084754317998886, "sign": 0.74144097417593, "keep": 0.2624799737241119, "leak": 0.5235152924433351, "replacement": 1.2413146942853928}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 2.495887041091919, "loss_scale": 4.0, "retries": 0, "seconds": 276.1, "max_memory_gb": 19.72, "pair": 0.5591292530298233, "sign": 0.5726439449936152, "keep": 0.1771923885680735, "leak": 0.6124773602932692, "replacement": 1.0871667563915253}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 2.3691930770874023, "loss_scale": 4.0, "retries": 0, "seconds": 277.1, "max_memory_gb": 19.72, "pair": 0.4550884924829006, "sign": 0.4925996921956539, "keep": 0.1415816338849254, "leak": 0.6795598135795444, "replacement": 0.9420690536499023}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 3.275482654571533, "loss_scale": 4.0, "retries": 0, "seconds": 278.3, "max_memory_gb": 19.72, "pair": 0.2392045222222805, "sign": 0.19950163550674915, "keep": 0.13514374231453985, "leak": 1.1246446147561073, "replacement": 0.6876883506774902}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 1.4083786010742188, "loss_scale": 4.0, "retries": 0, "seconds": 279.4, "max_memory_gb": 19.72, "pair": 0.22400161996483803, "sign": 0.2050485834479332, "keep": 0.014128951821476221, "leak": 0.9573708176612854, "replacement": 0.572933092713356}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 2.384272813796997, "loss_scale": 4.0, "retries": 0, "seconds": 280.4, "max_memory_gb": 19.72, "pair": 0.43581547960639, "sign": 0.42548209242522717, "keep": 0.3038529879413545, "leak": 0.6648522950708866, "replacement": 0.9666960909962654}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 2.0569324493408203, "loss_scale": 4.0, "retries": 0, "seconds": 281.5, "max_memory_gb": 19.72, "pair": 0.3861341327428818, "sign": 0.3393813753500581, "keep": 0.21549919620156288, "leak": 0.9664796198921977, "replacement": 0.9051943123340607}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 2.056236505508423, "loss_scale": 4.0, "retries": 0, "seconds": 282.6, "max_memory_gb": 19.72, "pair": 0.4644216075539589, "sign": 0.5339753106236458, "keep": 0.5572573525132611, "leak": 0.6506723873317242, "replacement": 1.1727060303092003}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 2.5472562313079834, "loss_scale": 4.0, "retries": 0, "seconds": 283.7, "max_memory_gb": 19.72, "pair": 0.4595189578831196, "sign": 0.4159788638353348, "keep": 0.4986448744311929, "leak": 0.7762175276875496, "replacement": 1.110885128378868}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 3.9460842609405518, "loss_scale": 4.0, "retries": 0, "seconds": 284.7, "max_memory_gb": 19.72, "pair": 0.2592761367559433, "sign": 0.17932939250022173, "keep": 0.8278637686744332, "leak": 1.587546981871128, "replacement": 1.1597594767808914}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 2.623094320297241, "loss_scale": 4.0, "retries": 0, "seconds": 285.8, "max_memory_gb": 19.72, "pair": 0.3722771182656288, "sign": 0.38723196648061275, "keep": 0.14825651817955077, "leak": 1.2350970655679703, "replacement": 0.9487956166267395}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 4.764170169830322, "loss_scale": 4.0, "retries": 0, "seconds": 286.9, "max_memory_gb": 19.72, "pair": 0.47251588851213455, "sign": 0.45977755822241306, "keep": 0.997752757743001, "leak": 0.34232211753260344, "replacement": 1.286861576139927}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.699312210083008, "loss_scale": 4.0, "retries": 0, "seconds": 288.0, "max_memory_gb": 19.72, "pair": 0.4464824013411999, "sign": 0.45993198081851006, "keep": 0.2519459545146674, "leak": 0.8039461891166866, "replacement": 1.0034078806638718}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 1.1831861734390259, "loss_scale": 4.0, "retries": 0, "seconds": 289.1, "max_memory_gb": 19.72, "pair": 0.6544056683778763, "sign": 0.7599161863327026, "keep": 0.4024661745643243, "leak": 0.29574953217525035, "replacement": 1.3095342814922333}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.063706874847412, "loss_scale": 4.0, "retries": 0, "seconds": 290.1, "max_memory_gb": 19.72, "pair": 0.4662225469946861, "sign": 0.5780764129012823, "keep": 1.2143310122191906, "leak": 0.5985939028832945, "replacement": 1.5120747238397598}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 3.044891834259033, "loss_scale": 4.0, "retries": 0, "seconds": 291.2, "max_memory_gb": 19.72, "pair": 0.31645561940968037, "sign": 0.3891461305320263, "keep": 0.35694309789687395, "leak": 0.9480064660310745, "replacement": 0.9265018701553345}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 2.9793484210968018, "loss_scale": 4.0, "retries": 0, "seconds": 292.2, "max_memory_gb": 19.72, "pair": 0.32259101793169975, "sign": 0.39901172649115324, "keep": 0.2319150564726442, "leak": 1.116302895359695, "replacement": 0.9171301275491714}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 3.2842891216278076, "loss_scale": 4.0, "retries": 0, "seconds": 293.3, "max_memory_gb": 19.72, "pair": 0.30332184210419655, "sign": 0.09917845763266087, "keep": 0.025196240283548832, "leak": 0.8860230222344398, "replacement": 0.587014950811863}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 1.830773949623108, "loss_scale": 4.0, "retries": 0, "seconds": 294.4, "max_memory_gb": 19.72, "pair": 0.2747744619846344, "sign": 0.16096259467303753, "keep": 0.029273670341353863, "leak": 0.3048929926007986, "replacement": 0.4461158439517021}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 2.8383188247680664, "loss_scale": 4.0, "retries": 0, "seconds": 295.5, "max_memory_gb": 19.72, "pair": 0.5913479812443256, "sign": 0.7072070762515068, "keep": 0.2880493518896401, "leak": 0.35294796942616813, "replacement": 1.177213191986084}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 3.8141541481018066, "loss_scale": 4.0, "retries": 0, "seconds": 296.6, "max_memory_gb": 19.72, "pair": 0.41019069589674473, "sign": 0.452747480943799, "keep": 1.2628947850316763, "leak": 0.8209289620281197, "replacement": 1.4732440784573555}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 2.9950380325317383, "loss_scale": 4.0, "retries": 0, "seconds": 297.6, "max_memory_gb": 19.72, "pair": 0.24908526428043842, "sign": 0.1293716747313738, "keep": 0.01797047263244167, "leak": 0.6387901082634926, "replacement": 0.4824538826942444}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 3.0092201232910156, "loss_scale": 4.0, "retries": 0, "seconds": 298.7, "max_memory_gb": 19.72, "pair": 0.3330954425036907, "sign": 0.30811429768800735, "keep": 0.014756181393750012, "leak": 0.7578469291329384, "replacement": 0.6839923784136772}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 2.298259973526001, "loss_scale": 4.0, "retries": 0, "seconds": 299.8, "max_memory_gb": 19.72, "pair": 0.3666691705584526, "sign": 0.48793452978134155, "keep": 0.339439318748191, "leak": 0.3777057812549174, "replacement": 0.8747825175523758}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 2.9843292236328125, "loss_scale": 4.0, "retries": 0, "seconds": 300.9, "max_memory_gb": 19.72, "pair": 0.4316530227661133, "sign": 0.32721368968486786, "keep": 0.1883466828148812, "leak": 0.8036202443763614, "replacement": 0.8903382867574692}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 2.9980719089508057, "loss_scale": 4.0, "retries": 0, "seconds": 301.9, "max_memory_gb": 19.72, "pair": 0.28078196570277214, "sign": 0.2585039557889104, "keep": 0.032573544478509575, "leak": 0.8704065009951591, "replacement": 0.6439223438501358}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.31394624710083, "loss_scale": 4.0, "retries": 0, "seconds": 303.0, "max_memory_gb": 19.72, "pair": 0.375732634216547, "sign": 0.3033604212105274, "keep": 0.44528052204987034, "leak": 0.3955240655923262, "replacement": 0.8489341139793396}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.0266940593719482, "loss_scale": 4.0, "retries": 0, "seconds": 304.1, "max_memory_gb": 19.72, "pair": 0.3865575082600117, "sign": 0.39561916049569845, "keep": 0.26562410668702796, "leak": 0.5793864727020264, "replacement": 0.8620257824659348}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 3.90679931640625, "loss_scale": 4.0, "retries": 0, "seconds": 305.2, "max_memory_gb": 19.72, "pair": 0.15610701590776443, "sign": 0.10126730613410473, "keep": 0.13214907003566623, "leak": 1.7831916362047195, "replacement": 0.7186130881309509}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.9799203872680664, "loss_scale": 4.0, "retries": 0, "seconds": 306.3, "max_memory_gb": 19.72, "pair": 0.3441524989902973, "sign": 0.3499301695264876, "keep": 0.13536341907456517, "leak": 0.7125672474503517, "replacement": 0.764941118657589}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 4.6651835441589355, "loss_scale": 4.0, "retries": 0, "seconds": 307.4, "max_memory_gb": 19.72, "pair": 0.4941267892718315, "sign": 0.7649660930037498, "keep": 0.1714898084755987, "leak": 1.0977245830581523, "replacement": 1.2367858365178108}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 5.495343208312988, "loss_scale": 4.0, "retries": 0, "seconds": 308.5, "max_memory_gb": 19.72, "pair": 0.4323628358542919, "sign": 0.7671374306082726, "keep": 0.04066856176359579, "leak": 0.8438708586618304, "replacement": 1.0472335070371628}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 4.003182411193848, "loss_scale": 4.0, "retries": 0, "seconds": 309.7, "max_memory_gb": 19.72, "pair": 0.4871748834848404, "sign": 0.49366969987750053, "keep": 0.4812110015191138, "leak": 0.4739579437300563, "replacement": 1.0931047052145004}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 3.805339813232422, "loss_scale": 4.0, "retries": 0, "seconds": 310.7, "max_memory_gb": 19.72, "pair": 0.30920829996466637, "sign": 0.4897984527051449, "keep": 0.020465038251131773, "leak": 0.9943710081279278, "replacement": 0.8129327967762947}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 3.3338847160339355, "loss_scale": 4.0, "retries": 0, "seconds": 311.8, "max_memory_gb": 19.72, "pair": 0.3000745289027691, "sign": 0.2659249734133482, "keep": 0.18411462660878897, "leak": 0.9065352901816368, "replacement": 0.7517281472682953}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 2.619194269180298, "loss_scale": 4.0, "retries": 0, "seconds": 312.8, "max_memory_gb": 19.72, "pair": 0.47939225286245346, "sign": 0.48883690871298313, "keep": 0.45350088184932247, "leak": 0.2226606123149395, "replacement": 1.0062263309955597}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 4.381894111633301, "loss_scale": 4.0, "retries": 0, "seconds": 313.9, "max_memory_gb": 19.72, "pair": 0.13324508629739285, "sign": 0.0716627947986126, "keep": 0.012805573060177267, "leak": 1.3918678611516953, "replacement": 0.5234462395310402}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 3.560112476348877, "loss_scale": 4.0, "retries": 0, "seconds": 315.0, "max_memory_gb": 19.72, "pair": 0.32609159871935844, "sign": 0.22338471934199333, "keep": 0.1076406454667449, "leak": 0.6201324760913849, "replacement": 0.6466374024748802}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 2.9022068977355957, "loss_scale": 4.0, "retries": 0, "seconds": 316.0, "max_memory_gb": 19.72, "pair": 0.2669629957526922, "sign": 0.2475854605436325, "keep": 0.12701283441856503, "leak": 0.7086879685521126, "replacement": 0.6314341351389885}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 4.24224853515625, "loss_scale": 4.0, "retries": 0, "seconds": 317.1, "max_memory_gb": 19.72, "pair": 0.34730610996484756, "sign": 0.20914993714541197, "keep": 0.4758855551481247, "leak": 0.3188985805027187, "replacement": 0.7695484757423401}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 3.3299989700317383, "loss_scale": 4.0, "retries": 0, "seconds": 318.1, "max_memory_gb": 19.72, "pair": 0.47509803250432014, "sign": 0.6842088811099529, "keep": 0.7138871508650482, "leak": 1.0012281769886613, "replacement": 1.4244530126452446}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 3.3785836696624756, "loss_scale": 4.0, "retries": 0, "seconds": 319.2, "max_memory_gb": 19.72, "pair": 0.5046032927930355, "sign": 0.6986463144421577, "keep": 0.4779437342658639, "leak": 0.852912300972605, "replacement": 1.3061263114213943}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 2.535417079925537, "loss_scale": 4.0, "retries": 0, "seconds": 320.2, "max_memory_gb": 19.72, "pair": 0.4714694172143936, "sign": 0.6656187474727631, "keep": 0.11473512987140566, "leak": 0.5874152984470129, "replacement": 1.0085001736879349}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 4.824328899383545, "loss_scale": 4.0, "retries": 0, "seconds": 321.3, "max_memory_gb": 19.72, "pair": 0.5746914222836494, "sign": 0.7718285471200943, "keep": 0.9752838846179657, "leak": 0.12051681149750948, "replacement": 1.478376843035221}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 5.09034538269043, "loss_scale": 4.0, "retries": 0, "seconds": 322.4, "max_memory_gb": 19.72, "pair": 0.23030203580856323, "sign": 0.12159311585128307, "keep": 0.05102225160226226, "leak": 1.670209364965558, "replacement": 0.7341620773077011}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 3.153096914291382, "loss_scale": 4.0, "retries": 0, "seconds": 323.5, "max_memory_gb": 19.72, "pair": 0.6465517710894346, "sign": 0.8262895233929157, "keep": 0.4405972622334957, "leak": 1.0760005599877331, "replacement": 1.548995390534401}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 5.613398551940918, "loss_scale": 4.0, "retries": 0, "seconds": 324.5, "max_memory_gb": 19.72, "pair": 0.39211326837539673, "sign": 0.4212189093232155, "keep": 0.19195879623293877, "leak": 0.5434433920308948, "replacement": 0.8345629647374153}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 3.9503164291381836, "loss_scale": 4.0, "retries": 0, "seconds": 325.6, "max_memory_gb": 19.72, "pair": 0.34855588898062706, "sign": 0.080650445073843, "keep": 0.04403808817733079, "leak": 0.3582271058112383, "replacement": 0.5004569590091705}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 2.652975559234619, "loss_scale": 4.0, "retries": 0, "seconds": 326.7, "max_memory_gb": 19.72, "pair": 0.2593742925673723, "sign": 0.1795404157601297, "keep": 0.03526160982437432, "leak": 1.278652936220169, "replacement": 0.6864385604858398}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 4.081050395965576, "loss_scale": 4.0, "retries": 0, "seconds": 327.8, "max_memory_gb": 19.72, "pair": 0.36614030599594116, "sign": 0.5402694866061211, "keep": 0.054629553109407425, "leak": 1.2295586000254843, "replacement": 0.9709794893860817}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 5.10621976852417, "loss_scale": 4.0, "retries": 0, "seconds": 328.9, "max_memory_gb": 19.72, "pair": 0.1902200784534216, "sign": 0.1289934073574841, "keep": 0.08174543647328392, "leak": 2.4782753414474428, "replacement": 0.9151583164930344}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 3.9085099697113037, "loss_scale": 4.0, "retries": 0, "seconds": 329.9, "max_memory_gb": 19.72, "pair": 0.20931731536984444, "sign": 0.0520830936729908, "keep": 0.0725763076916337, "leak": 1.0250150598585606, "replacement": 0.5279007852077484}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 2.689566135406494, "loss_scale": 4.0, "retries": 0, "seconds": 331.1, "max_memory_gb": 19.72, "pair": 0.38508546352386475, "sign": 0.34123471193015575, "keep": 0.12732412456534803, "leak": 0.6808041501790285, "replacement": 0.789565920829773}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 5.161557197570801, "loss_scale": 4.0, "retries": 0, "seconds": 332.2, "max_memory_gb": 19.72, "pair": 0.4081605803221464, "sign": 0.6535196341574192, "keep": 0.062270582013297826, "leak": 1.395563516067341, "replacement": 1.1149466186761856}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 3.0987327098846436, "loss_scale": 4.0, "retries": 0, "seconds": 333.2, "max_memory_gb": 19.72, "pair": 0.5529862381517887, "sign": 0.6343652415089309, "keep": 0.3481047721579671, "leak": 0.37959232786670327, "replacement": 1.139119267463684}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 2.2688097953796387, "loss_scale": 4.0, "retries": 0, "seconds": 334.3, "max_memory_gb": 19.72, "pair": 0.21256201341748238, "sign": 0.11201972793787718, "keep": 0.03145812964066863, "leak": 0.8317901641130447, "replacement": 0.4922484904527664}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 1.6457942724227905, "loss_scale": 4.0, "retries": 0, "seconds": 335.3, "max_memory_gb": 19.72, "pair": 0.37026694416999817, "sign": 0.4290356133133173, "keep": 0.208126665384043, "leak": 0.5599376820027828, "replacement": 0.8288324847817421}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.282020092010498, "loss_scale": 4.0, "retries": 0, "seconds": 336.4, "max_memory_gb": 19.72, "pair": 0.31007829308509827, "sign": 0.3102358505129814, "keep": 0.07567230594577268, "leak": 0.4380964762531221, "replacement": 0.612556517124176}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 7.562952995300293, "loss_scale": 4.0, "retries": 0, "seconds": 337.5, "max_memory_gb": 19.72, "pair": 0.4499779939651489, "sign": 0.6669053817167878, "keep": 0.6173862800933421, "leak": 1.4641306400299072, "replacement": 1.4581564664840698}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 3.7607030868530273, "loss_scale": 4.0, "retries": 0, "seconds": 338.6, "max_memory_gb": 19.72, "pair": 0.38707862235605717, "sign": 0.49887070897966623, "keep": 0.3240828870330006, "leak": 0.8124732486903667, "replacement": 1.001673735678196}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 4.93881368637085, "loss_scale": 4.0, "retries": 0, "seconds": 339.7, "max_memory_gb": 19.72, "pair": 0.23643654212355614, "sign": 0.22848437819629908, "keep": 0.06358609907329082, "leak": 1.1893173474818468, "replacement": 0.6798011288046837}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 3.45601224899292, "loss_scale": 4.0, "retries": 0, "seconds": 340.8, "max_memory_gb": 19.72, "pair": 0.6121390387415886, "sign": 0.7358965268358588, "keep": 0.7162409462034702, "leak": 0.6315108418639284, "replacement": 1.496085450053215}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 3.3155088424682617, "loss_scale": 4.0, "retries": 0, "seconds": 341.8, "max_memory_gb": 19.72, "pair": 0.415976345539093, "sign": 0.42200047336518764, "keep": 0.28927953727543354, "leak": 0.9219511831179261, "replacement": 1.0021041482686996}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 3.455828905105591, "loss_scale": 4.0, "retries": 0, "seconds": 342.8, "max_memory_gb": 19.72, "pair": 0.395036693662405, "sign": 0.32408037036657333, "keep": 0.07594129908829927, "leak": 0.5222208457998931, "replacement": 0.7256027534604073}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 3.0858068466186523, "loss_scale": 4.0, "retries": 0, "seconds": 343.9, "max_memory_gb": 19.72, "pair": 0.41384582594037056, "sign": 0.3571961857378483, "keep": 0.18958052806556225, "leak": 0.41823435202240944, "replacement": 0.7917927503585815}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 3.272233486175537, "loss_scale": 4.0, "retries": 0, "seconds": 345.0, "max_memory_gb": 19.72, "pair": 0.41282155364751816, "sign": 0.4052233174443245, "keep": 0.38789955561514944, "leak": 0.4955277368426323, "replacement": 0.9332649558782578}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 6.423754692077637, "loss_scale": 4.0, "retries": 0, "seconds": 346.1, "max_memory_gb": 19.72, "pair": 0.34891871735453606, "sign": 0.31505705416202545, "keep": 0.09276016312651336, "leak": 0.666931489482522, "replacement": 0.7195601835846901}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 6.172214984893799, "loss_scale": 4.0, "retries": 0, "seconds": 347.1, "max_memory_gb": 19.72, "pair": 0.3446127474308014, "sign": 0.5420793443918228, "keep": 0.033406097907572985, "leak": 0.82731057330966, "replacement": 0.8391830623149872}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 5.584362506866455, "loss_scale": 4.0, "retries": 0, "seconds": 348.2, "max_memory_gb": 19.72, "pair": 0.2123381309211254, "sign": 0.2134192930534482, "keep": 0.07264338713139296, "leak": 1.9391955137252808, "replacement": 0.8401683419942856}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 2.6658315658569336, "loss_scale": 4.0, "retries": 0, "seconds": 349.3, "max_memory_gb": 19.72, "pair": 0.19756301492452621, "sign": 0.08276114985346794, "keep": 0.06745384336682037, "leak": 0.5020254777246009, "replacement": 0.3981768861413002}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 4.424250602722168, "loss_scale": 4.0, "retries": 0, "seconds": 350.3, "max_memory_gb": 19.72, "pair": 0.5671653803437948, "sign": 0.6296153834555298, "keep": 0.5608864618698135, "leak": 0.43129237621542416, "replacement": 1.2702394351363182}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 3.1623480319976807, "loss_scale": 4.0, "retries": 0, "seconds": 351.4, "max_memory_gb": 19.72, "pair": 0.4290075972676277, "sign": 0.5665440037846565, "keep": 0.14583332452457398, "leak": 0.33186366548761725, "replacement": 0.8681621998548508}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 4.218449592590332, "loss_scale": 4.0, "retries": 0, "seconds": 352.4, "max_memory_gb": 19.72, "pair": 0.24181028082966805, "sign": 0.1233982015401125, "keep": 0.3850555745884776, "leak": 0.6828606287599541, "replacement": 0.6667523458600044}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 3.094296932220459, "loss_scale": 4.0, "retries": 0, "seconds": 353.5, "max_memory_gb": 19.72, "pair": 0.5061257556080818, "sign": 0.597962848842144, "keep": 0.19657110457774252, "leak": 0.19201891354168765, "replacement": 0.9513974562287331}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 3.4946229457855225, "loss_scale": 4.0, "retries": 0, "seconds": 354.6, "max_memory_gb": 19.72, "pair": 0.3690754100680351, "sign": 0.40397415310144424, "keep": 0.2631547330529429, "leak": 0.8145268000662327, "replacement": 0.9062715321779251}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 4.438591480255127, "loss_scale": 4.0, "retries": 0, "seconds": 355.7, "max_memory_gb": 19.72, "pair": 0.4731803461909294, "sign": 0.5343422656878829, "keep": 0.32061595004051924, "leak": 0.5260571949183941, "replacement": 1.0321737453341484}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 2.6168289184570312, "loss_scale": 4.0, "retries": 0, "seconds": 356.7, "max_memory_gb": 19.72, "pair": 0.5852872282266617, "sign": 0.6220826059579849, "keep": 0.5503095604944974, "leak": 0.3837822100613266, "replacement": 1.2674288526177406}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 3.7316925525665283, "loss_scale": 4.0, "retries": 0, "seconds": 357.8, "max_memory_gb": 19.72, "pair": 0.30147040635347366, "sign": 0.3070860728621483, "keep": 0.32996698352508247, "leak": 0.5380495004355907, "replacement": 0.7545093297958374}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 2.6443333625793457, "loss_scale": 4.0, "retries": 0, "seconds": 358.9, "max_memory_gb": 19.72, "pair": 0.6253501810133457, "sign": 0.7598570175468922, "keep": 0.5686717302305624, "leak": 0.43592814062776597, "replacement": 1.3985965847969055}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 4.437154293060303, "loss_scale": 4.0, "retries": 0, "seconds": 360.0, "max_memory_gb": 19.72, "pair": 0.4382808580994606, "sign": 0.5388600267469883, "keep": 0.715542615391314, "leak": 1.4599362909793854, "replacement": 1.4304662495851517}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 5.57177734375, "loss_scale": 4.0, "retries": 0, "seconds": 361.1, "max_memory_gb": 19.72, "pair": 0.543769795447588, "sign": 0.5801566801965237, "keep": 1.0865788934752345, "leak": 0.1928709615021944, "replacement": 1.4253553077578545}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 3.728806734085083, "loss_scale": 4.0, "retries": 0, "seconds": 362.2, "max_memory_gb": 19.72, "pair": 0.2756474297493696, "sign": 0.17279408872127533, "keep": 0.04470295074861497, "leak": 1.6312144650146365, "replacement": 0.7921995893120766}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 3.032122850418091, "loss_scale": 4.0, "retries": 0, "seconds": 363.3, "max_memory_gb": 19.72, "pair": 0.5466759316623211, "sign": 0.9723551319912076, "keep": 0.3889222275465727, "leak": 0.5224151597358286, "replacement": 1.3579184338450432}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 2.3134539127349854, "loss_scale": 4.0, "retries": 0, "seconds": 364.5, "max_memory_gb": 19.72, "pair": 0.4474497213959694, "sign": 0.5112653933465481, "keep": 0.31621949980035424, "leak": 0.4900509813451208, "replacement": 0.9837049022316933}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 3.221216917037964, "loss_scale": 4.0, "retries": 0, "seconds": 365.6, "max_memory_gb": 19.72, "pair": 0.4957847259938717, "sign": 0.7793047018349171, "keep": 0.33491747453808784, "leak": 0.9476044662296772, "replacement": 1.2897969484329224}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 3.0602285861968994, "loss_scale": 4.0, "retries": 0, "seconds": 366.6, "max_memory_gb": 19.72, "pair": 0.32150956708937883, "sign": 0.3432028819806874, "keep": 0.5677552362903953, "leak": 0.3241073675453663, "replacement": 0.858015451580286}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 4.7708024978637695, "loss_scale": 4.0, "retries": 0, "seconds": 367.7, "max_memory_gb": 19.72, "pair": 0.346770204603672, "sign": 0.3727294933050871, "keep": 0.3864686593879014, "leak": 0.9219606960177771, "replacement": 0.9568594545125961}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 5.55471658706665, "loss_scale": 4.0, "retries": 0, "seconds": 368.8, "max_memory_gb": 19.72, "pair": 0.5800199750810862, "sign": 0.6512935850769281, "keep": 0.5448987446725368, "leak": 0.622130798175931, "replacement": 1.3336488604545593}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 4.026869297027588, "loss_scale": 4.0, "retries": 0, "seconds": 369.9, "max_memory_gb": 19.72, "pair": 0.4742000102996826, "sign": 0.5350651629269123, "keep": 0.02296188985928893, "leak": 0.3881787247955799, "replacement": 0.8502582535147667}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 3.499239921569824, "loss_scale": 4.0, "retries": 0, "seconds": 370.9, "max_memory_gb": 19.72, "pair": 0.21000397577881813, "sign": 0.13778289407491684, "keep": 0.024239272286649793, "leak": 0.929274458438158, "replacement": 0.5233336612582207}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 3.0312154293060303, "loss_scale": 4.0, "retries": 0, "seconds": 372.0, "max_memory_gb": 19.72, "pair": 0.3705386146903038, "sign": 0.38268835190683603, "keep": 0.24457711214199662, "leak": 0.4699656721204519, "replacement": 0.8016627505421638}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 3.7778024673461914, "loss_scale": 4.0, "retries": 0, "seconds": 373.1, "max_memory_gb": 19.72, "pair": 0.4018822368234396, "sign": 0.6205906085669994, "keep": 0.6518810980487615, "leak": 0.5791188180446625, "replacement": 1.1828978061676025}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 2.72563099861145, "loss_scale": 4.0, "retries": 0, "seconds": 374.1, "max_memory_gb": 19.72, "pair": 0.375921405851841, "sign": 0.3790845759212971, "keep": 0.30249038245528936, "leak": 0.582799803436501, "replacement": 0.862408846616745}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 3.821986675262451, "loss_scale": 4.0, "retries": 0, "seconds": 375.3, "max_memory_gb": 19.72, "pair": 0.15368910506367683, "sign": 0.16583239706233144, "keep": 0.022343597607687116, "leak": 1.3689273558557034, "replacement": 0.5900089293718338}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 3.53867244720459, "loss_scale": 4.0, "retries": 0, "seconds": 376.4, "max_memory_gb": 19.72, "pair": 0.3268479648977518, "sign": 0.3759535364806652, "keep": 0.1757259122096002, "leak": 1.1471812437521294, "replacement": 0.8894830048084259}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 3.411163330078125, "loss_scale": 4.0, "retries": 0, "seconds": 377.5, "max_memory_gb": 19.72, "pair": 0.38974815234541893, "sign": 0.7001643478870392, "keep": 0.30048387590795755, "leak": 0.30648005846887827, "replacement": 0.9666922762989998}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 2.370478391647339, "loss_scale": 4.0, "retries": 0, "seconds": 378.5, "max_memory_gb": 19.72, "pair": 0.22209842037409544, "sign": 0.15212851017713547, "keep": 0.09857761999592185, "leak": 0.3170226626098156, "replacement": 0.4267071485519409}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 3.947650909423828, "loss_scale": 4.0, "retries": 0, "seconds": 379.6, "max_memory_gb": 19.72, "pair": 0.4610657813027501, "sign": 0.6395858018659055, "keep": 0.5469045429490507, "leak": 0.6224259026348591, "replacement": 1.2099174782633781}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 7.5987067222595215, "loss_scale": 4.0, "retries": 0, "seconds": 380.6, "max_memory_gb": 19.72, "pair": 0.3369486853480339, "sign": 0.593888197094202, "keep": 0.782615031581372, "leak": 1.2163051813840866, "replacement": 1.3292766064405441}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 2.008678913116455, "loss_scale": 4.0, "retries": 0, "seconds": 381.7, "max_memory_gb": 19.72, "pair": 0.4410337544977665, "sign": 0.45182348787784576, "keep": 0.33501080062706023, "leak": 0.21996963676065207, "replacement": 0.8894432932138443}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 4.784430503845215, "loss_scale": 4.0, "retries": 0, "seconds": 382.8, "max_memory_gb": 19.72, "pair": 0.47518351674079895, "sign": 0.54330874979496, "keep": 0.27368256996851414, "leak": 1.3950235699303448, "replacement": 1.232435055077076}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 2.9459643363952637, "loss_scale": 4.0, "retries": 0, "seconds": 383.9, "max_memory_gb": 19.72, "pair": 0.20467692241072655, "sign": 0.18473584204912186, "keep": 0.034518289379775524, "leak": 1.115697294473648, "replacement": 0.5932283103466034}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 8.975424766540527, "loss_scale": 4.0, "retries": 0, "seconds": 385.0, "max_memory_gb": 19.72, "pair": 0.20115857105702162, "sign": 0.13458843622356653, "keep": 0.12357715424150229, "leak": 3.357915759086609, "replacement": 1.1697202920913696}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 3.5369462966918945, "loss_scale": 4.0, "retries": 0, "seconds": 386.1, "max_memory_gb": 19.72, "pair": 0.5241193175315857, "sign": 0.6739458795636892, "keep": 0.5008127759210765, "leak": 0.22962594195269048, "replacement": 1.1689050793647766}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 3.1112587451934814, "loss_scale": 4.0, "retries": 0, "seconds": 387.2, "max_memory_gb": 19.72, "pair": 0.3551865927875042, "sign": 0.5197876878082752, "keep": 0.060883654514327645, "leak": 1.126586139202118, "replacement": 0.9271687865257263}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 3.7219364643096924, "loss_scale": 4.0, "retries": 0, "seconds": 388.3, "max_memory_gb": 19.72, "pair": 0.24554525315761566, "sign": 0.2029833346605301, "keep": 0.0866195005364716, "leak": 1.088666956871748, "replacement": 0.6625134125351906}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 5.376177787780762, "loss_scale": 4.0, "retries": 0, "seconds": 389.4, "max_memory_gb": 19.72, "pair": 0.22338019125163555, "sign": 0.4129238147288561, "keep": 0.028757783700712025, "leak": 0.9598607085645199, "replacement": 0.6841861680150032}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 6.825112819671631, "loss_scale": 4.0, "retries": 0, "seconds": 390.5, "max_memory_gb": 19.72, "pair": 0.42272597923874855, "sign": 0.4453857624903321, "keep": 2.0404304643161595, "leak": 0.26494045928120613, "replacement": 1.731869176030159}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 4.870507717132568, "loss_scale": 4.0, "retries": 0, "seconds": 391.5, "max_memory_gb": 19.72, "pair": 0.5775871723890305, "sign": 1.1123317256569862, "keep": 0.05737725738435984, "leak": 0.5249796770513058, "replacement": 1.29368656873703}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 3.8611605167388916, "loss_scale": 4.0, "retries": 0, "seconds": 392.6, "max_memory_gb": 19.72, "pair": 0.15700649097561836, "sign": 0.06212840415537357, "keep": 0.06484379223547876, "leak": 1.150285318493843, "replacement": 0.5080639198422432}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 2.1655116081237793, "loss_scale": 4.0, "retries": 0, "seconds": 393.7, "max_memory_gb": 19.72, "pair": 0.3175471182912588, "sign": 0.37175592593848705, "keep": 0.27885204344056547, "leak": 0.3587885480374098, "replacement": 0.7325482442975044}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 7.861630439758301, "loss_scale": 4.0, "retries": 0, "seconds": 394.9, "max_memory_gb": 19.72, "pair": 0.24364664033055305, "sign": 0.314527977257967, "keep": 0.05286026978865266, "leak": 2.38786331564188, "replacement": 1.0243065878748894}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 2.611848831176758, "loss_scale": 4.0, "retries": 0, "seconds": 395.9, "max_memory_gb": 19.72, "pair": 0.34800041280686855, "sign": 0.40616627782583237, "keep": 0.41073355777189136, "leak": 0.6420724168419838, "replacement": 0.9169684499502182}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 3.020681142807007, "loss_scale": 4.0, "retries": 0, "seconds": 396.9, "max_memory_gb": 19.72, "pair": 0.17325102537870407, "sign": 0.11905249953269958, "keep": 0.10097175103146583, "leak": 0.3735271953046322, "replacement": 0.37664493918418884}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 3.0955424308776855, "loss_scale": 4.0, "retries": 0, "seconds": 398.0, "max_memory_gb": 19.72, "pair": 0.3814165070652962, "sign": 0.37224322301335633, "keep": 0.08492755779298022, "leak": 1.0278723808005452, "replacement": 0.8669699877500534}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 5.190154075622559, "loss_scale": 4.0, "retries": 0, "seconds": 399.1, "max_memory_gb": 19.72, "pair": 0.1864603515714407, "sign": 0.26528345234692097, "keep": 0.9340436877682805, "leak": 1.2429570369422436, "replacement": 1.0968632251024246}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 2.5366811752319336, "loss_scale": 4.0, "retries": 0, "seconds": 400.2, "max_memory_gb": 19.72, "pair": 0.1883332021534443, "sign": 0.07379627134650946, "keep": 0.16064053471200168, "leak": 1.1095311790704727, "replacement": 0.5829344168305397}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 3.9438352584838867, "loss_scale": 4.0, "retries": 0, "seconds": 401.3, "max_memory_gb": 19.72, "pair": 0.24452560767531395, "sign": 0.2347845509648323, "keep": 0.09603983868146315, "leak": 0.6590368747711182, "replacement": 0.5746970027685165}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 2.70308518409729, "loss_scale": 4.0, "retries": 0, "seconds": 402.3, "max_memory_gb": 19.72, "pair": 0.5605535581707954, "sign": 0.7892131246626377, "keep": 0.07837291562464088, "leak": 0.4959590742364526, "replacement": 1.1183363422751427}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 3.0776865482330322, "loss_scale": 4.0, "retries": 0, "seconds": 403.5, "max_memory_gb": 19.72, "pair": 0.2765584159642458, "sign": 0.3044871096499264, "keep": 0.46812895545735955, "leak": 0.6879553347826004, "replacement": 0.8348552957177162}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 2.898355484008789, "loss_scale": 4.0, "retries": 0, "seconds": 404.6, "max_memory_gb": 19.72, "pair": 0.38038917630910873, "sign": 0.47318560630083084, "keep": 0.30103327526012436, "leak": 1.050180085003376, "replacement": 1.030043676495552}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 2.69667387008667, "loss_scale": 4.0, "retries": 0, "seconds": 405.7, "max_memory_gb": 19.72, "pair": 0.28715415857732296, "sign": 0.2411989290267229, "keep": 0.20676787162665278, "leak": 0.31679253838956356, "replacement": 0.5903357043862343}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 2.8632545471191406, "loss_scale": 4.0, "retries": 0, "seconds": 406.7, "max_memory_gb": 19.72, "pair": 0.33401756547391415, "sign": 0.3693358600139618, "keep": 0.06077968259342015, "leak": 0.4976434772834182, "replacement": 0.673486202955246}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 4.018319129943848, "loss_scale": 4.0, "retries": 0, "seconds": 407.8, "max_memory_gb": 19.72, "pair": 0.2367515005171299, "sign": 0.20094581972807646, "keep": 0.1317699938081205, "leak": 1.0167505592107773, "replacement": 0.6572970449924469}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 2.9468278884887695, "loss_scale": 4.0, "retries": 0, "seconds": 408.9, "max_memory_gb": 19.72, "pair": 0.2636346369981766, "sign": 0.3121028635650873, "keep": 0.11939494032412767, "leak": 0.47083638049662113, "replacement": 0.5970926359295845}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 3.9785897731781006, "loss_scale": 4.0, "retries": 0, "seconds": 410.0, "max_memory_gb": 19.72, "pair": 0.25321531295776367, "sign": 0.4667497221380472, "keep": 0.21410190872848034, "leak": 1.2110256627202034, "replacement": 0.8963975682854652}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 2.1747241020202637, "loss_scale": 4.0, "retries": 0, "seconds": 411.0, "max_memory_gb": 19.72, "pair": 0.6302212104201317, "sign": 0.831098847091198, "keep": 0.42855393700301647, "leak": 0.08516247215447947, "replacement": 1.2813382148742676}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 4.061891078948975, "loss_scale": 4.0, "retries": 0, "seconds": 412.1, "max_memory_gb": 19.72, "pair": 0.502179928123951, "sign": 0.6914740232750773, "keep": 0.8989207278937101, "leak": 0.5790905579924583, "replacement": 1.4421499818563461}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 3.1276180744171143, "loss_scale": 4.0, "retries": 0, "seconds": 413.2, "max_memory_gb": 19.72, "pair": 0.33817948400974274, "sign": 0.3382220771163702, "keep": 0.18738701473921537, "leak": 0.632242476567626, "replacement": 0.7590446323156357}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 5.665440559387207, "loss_scale": 4.0, "retries": 0, "seconds": 414.2, "max_memory_gb": 19.72, "pair": 0.09298289520666003, "sign": 0.08444375265389681, "keep": 0.021250645630061626, "leak": 1.371047705411911, "replacement": 0.4885920211672783}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 3.997122287750244, "loss_scale": 4.0, "retries": 0, "seconds": 415.3, "max_memory_gb": 19.72, "pair": 0.4137292727828026, "sign": 0.4976281449198723, "keep": 0.3221890677232295, "leak": 0.45507496129721403, "replacement": 0.9374066218733788}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 3.6693596839904785, "loss_scale": 4.0, "retries": 0, "seconds": 416.4, "max_memory_gb": 19.72, "pair": 0.2886222004890442, "sign": 0.27345651341602206, "keep": 0.08954601205186918, "leak": 0.3891266696155071, "replacement": 0.5674051195383072}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 3.63997483253479, "loss_scale": 4.0, "retries": 0, "seconds": 417.5, "max_memory_gb": 19.72, "pair": 0.3231328297406435, "sign": 0.4703790172934532, "keep": 0.14411916979588568, "leak": 0.5620500752702355, "replacement": 0.7708944231271744}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 2.648216485977173, "loss_scale": 4.0, "retries": 0, "seconds": 418.6, "max_memory_gb": 19.72, "pair": 0.15090292133390903, "sign": 0.16021894849836826, "keep": 0.05335558854858391, "leak": 0.6528479792177677, "replacement": 0.42090219259262085}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 5.459278106689453, "loss_scale": 4.0, "retries": 0, "seconds": 419.8, "max_memory_gb": 19.72, "pair": 0.2979210037738085, "sign": 0.19726653560064733, "keep": 0.07927497243508697, "leak": 1.7652404299005866, "replacement": 0.8775018751621246}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 3.5974464416503906, "loss_scale": 4.0, "retries": 0, "seconds": 420.9, "max_memory_gb": 19.72, "pair": 0.3043876849114895, "sign": 0.3401730917394161, "keep": 0.1221541604027152, "leak": 0.8686951287090778, "replacement": 0.7527250871062279}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 2.915224313735962, "loss_scale": 4.0, "retries": 0, "seconds": 422.0, "max_memory_gb": 19.72, "pair": 0.5519589651376009, "sign": 0.7406134679913521, "keep": 0.7353470213711262, "leak": 0.18236651178449392, "replacement": 1.335530862212181}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 2.9068007469177246, "loss_scale": 4.0, "retries": 0, "seconds": 423.1, "max_memory_gb": 19.72, "pair": 0.1525835506618023, "sign": 0.08097266778349876, "keep": 0.12221786382724531, "leak": 0.32432736456394196, "replacement": 0.33526065200567245}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 3.8573877811431885, "loss_scale": 4.0, "retries": 0, "seconds": 424.2, "max_memory_gb": 19.72, "pair": 0.22788878343999386, "sign": 0.2961190491914749, "keep": 0.1608241805806756, "leak": 0.25898107048124075, "replacement": 0.5211056768894196}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 2.5197455883026123, "loss_scale": 4.0, "retries": 0, "seconds": 425.3, "max_memory_gb": 19.72, "pair": 0.3842058554291725, "sign": 0.5228694435209036, "keep": 0.09028603602200747, "leak": 0.514649626473556, "replacement": 0.8194459900259972}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 4.7538042068481445, "loss_scale": 4.0, "retries": 0, "seconds": 426.3, "max_memory_gb": 19.72, "pair": 0.13217157498002052, "sign": 0.199150743894279, "keep": 0.05034020246239379, "leak": 1.3803072944283485, "replacement": 0.6019938662648201}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 2.608975410461426, "loss_scale": 4.0, "retries": 0, "seconds": 427.4, "max_memory_gb": 19.72, "pair": 0.1799198817461729, "sign": 0.06060277111828327, "keep": 0.060730062425136566, "leak": 0.45155395939946175, "replacement": 0.35347479581832886}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 2.662442684173584, "loss_scale": 4.0, "retries": 0, "seconds": 428.4, "max_memory_gb": 19.72, "pair": 0.36944628320634365, "sign": 0.4822566406801343, "keep": 0.2537212730385363, "leak": 0.3694290332496166, "replacement": 0.8297925069928169}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 3.565047264099121, "loss_scale": 4.0, "retries": 0, "seconds": 429.5, "max_memory_gb": 19.72, "pair": 0.3673617020249367, "sign": 0.3870216179639101, "keep": 0.15877548465505242, "leak": 0.9053639036137611, "replacement": 0.8666012287139893}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 7.040821075439453, "loss_scale": 4.0, "retries": 0, "seconds": 430.6, "max_memory_gb": 19.72, "pair": 0.4308431427925825, "sign": 0.7052611904218793, "keep": 0.8661875929683447, "leak": 0.9959463179111481, "replacement": 1.465554118156433}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 2.8711295127868652, "loss_scale": 4.0, "retries": 0, "seconds": 431.7, "max_memory_gb": 19.72, "pair": 0.323751799762249, "sign": 0.32402449660003185, "keep": 0.21494456054642797, "leak": 0.7194308787584305, "replacement": 0.7730940580368042}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 6.798947811126709, "loss_scale": 4.0, "retries": 0, "seconds": 432.8, "max_memory_gb": 19.72, "pair": 0.4003792479634285, "sign": 0.7301542609930038, "keep": 0.7187699937494472, "leak": 0.8876342736184597, "replacement": 1.3467498645186424}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 4.4124274253845215, "loss_scale": 4.0, "retries": 0, "seconds": 433.9, "max_memory_gb": 19.72, "pair": 0.6929130293428898, "sign": 0.8001720421016216, "keep": 0.30358948558568954, "leak": 0.302887137979269, "replacement": 1.3205155432224274}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 5.386120796203613, "loss_scale": 4.0, "retries": 0, "seconds": 435.0, "max_memory_gb": 19.72, "pair": 0.14882959052920341, "sign": 0.1221684250049293, "keep": 0.028924771468155086, "leak": 1.431875822134316, "replacement": 0.5823451355099678}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 5.113291263580322, "loss_scale": 4.0, "retries": 0, "seconds": 436.1, "max_memory_gb": 19.72, "pair": 0.36244548112154007, "sign": 0.47984363278374076, "keep": 0.403844702988863, "leak": 0.5540981236845255, "replacement": 0.9428141862154007}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 2.937739372253418, "loss_scale": 4.0, "retries": 0, "seconds": 437.1, "max_memory_gb": 19.72, "pair": 0.2834899313747883, "sign": 0.20303475111722946, "keep": 0.05007754423422739, "leak": 0.3877322101034224, "replacement": 0.5069791302084923}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 2.879115581512451, "loss_scale": 4.0, "retries": 0, "seconds": 438.2, "max_memory_gb": 19.72, "pair": 0.6031958796083927, "sign": 0.7850642679259181, "keep": 0.25346880475990474, "leak": 0.6692311554215848, "replacement": 1.2897701859474182}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 21.135059356689453, "loss_scale": 4.0, "retries": 0, "seconds": 439.3, "max_memory_gb": 19.72, "pair": 0.7697683535516262, "sign": 1.0985793881118298, "keep": 2.6854598950594664, "leak": 1.0868192166090012, "replacement": 2.933492809534073}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 3.8665502071380615, "loss_scale": 4.0, "retries": 0, "seconds": 440.4, "max_memory_gb": 19.72, "pair": 0.4128635320812464, "sign": 0.5195674574933946, "keep": 0.5257948529906571, "leak": 0.5658333909232169, "replacement": 1.0770029872655869}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 3.4483842849731445, "loss_scale": 4.0, "retries": 0, "seconds": 441.5, "max_memory_gb": 19.72, "pair": 0.3690231740474701, "sign": 0.4425990432500839, "keep": 0.32501947425771505, "leak": 0.9852254774141329, "replacement": 0.9991388097405434}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 3.098099708557129, "loss_scale": 4.0, "retries": 0, "seconds": 442.6, "max_memory_gb": 19.72, "pair": 0.1378254685550928, "sign": 0.11391142942011356, "keep": 0.1682159643387422, "leak": 1.0879893247038126, "replacement": 0.550886508077383}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 5.594248294830322, "loss_scale": 4.0, "retries": 0, "seconds": 443.7, "max_memory_gb": 19.72, "pair": 0.37159229442477226, "sign": 0.8086197935044765, "keep": 1.0825405571376905, "leak": 0.5054901354014874, "replacement": 1.4435449913144112}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 3.225931406021118, "loss_scale": 4.0, "retries": 0, "seconds": 444.8, "max_memory_gb": 19.72, "pair": 0.48081766068935394, "sign": 0.686751589179039, "keep": 0.2465643851319328, "leak": 0.46324191306484863, "replacement": 1.0632860660552979}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 4.27175235748291, "loss_scale": 4.0, "retries": 0, "seconds": 445.8, "max_memory_gb": 19.72, "pair": 0.17621976137161255, "sign": 0.15628180652856827, "keep": 0.04165822500362992, "leak": 1.5803717039525509, "replacement": 0.670282706618309}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 5.537796974182129, "loss_scale": 4.0, "retries": 0, "seconds": 446.9, "max_memory_gb": 19.72, "pair": 0.4141818629577756, "sign": 0.42378287739120424, "keep": 0.8397329077124596, "leak": 0.643302783370018, "replacement": 1.206765428185463}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 6.017684459686279, "loss_scale": 4.0, "retries": 0, "seconds": 447.9, "max_memory_gb": 19.72, "pair": 0.45094287022948265, "sign": 0.3985782917588949, "keep": 1.4446718986146152, "leak": 0.5231411634013057, "replacement": 1.5033532306551933}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 2.4405124187469482, "loss_scale": 4.0, "retries": 0, "seconds": 449.0, "max_memory_gb": 19.72, "pair": 0.20708700455725193, "sign": 0.15210155956447124, "keep": 0.07876773324096575, "leak": 0.2820746637880802, "replacement": 0.393040306866169}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 6.267953872680664, "loss_scale": 4.0, "retries": 0, "seconds": 450.1, "max_memory_gb": 19.72, "pair": 0.7790032997727394, "sign": 1.1500994563102722, "keep": 1.1949364652391523, "leak": 0.559756051748991, "replacement": 2.0914603024721146}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 2.0297086238861084, "loss_scale": 4.0, "retries": 0, "seconds": 451.1, "max_memory_gb": 19.72, "pair": 0.10754979029297829, "sign": 0.15282022021710873, "keep": 0.028997799614444375, "leak": 0.5421531507745385, "replacement": 0.33399708569049835}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 4.850907802581787, "loss_scale": 4.0, "retries": 0, "seconds": 452.1, "max_memory_gb": 19.72, "pair": 0.2583341132849455, "sign": 0.43281469494104385, "keep": 0.06674297124845907, "leak": 2.042165854945779, "replacement": 1.0186544209718704}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 6.060530662536621, "loss_scale": 4.0, "retries": 0, "seconds": 453.2, "max_memory_gb": 19.72, "pair": 0.20906485989689827, "sign": 0.38614803832024336, "keep": 0.674490055651404, "leak": 1.8752723336219788, "replacement": 1.2082020044326782}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 6.287245273590088, "loss_scale": 4.0, "retries": 0, "seconds": 454.3, "max_memory_gb": 19.72, "pair": 0.15704591386020184, "sign": 0.21114859450608492, "keep": 0.08981904503889382, "leak": 2.3731950521469116, "replacement": 0.9008285030722618}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 4.329636096954346, "loss_scale": 4.0, "retries": 0, "seconds": 455.4, "max_memory_gb": 19.72, "pair": 0.4316410291939974, "sign": 0.42805839981883764, "keep": 0.39672694774344563, "leak": 0.9323835570830852, "replacement": 1.0771295726299286}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 5.941620826721191, "loss_scale": 4.0, "retries": 0, "seconds": 456.5, "max_memory_gb": 19.72, "pair": 0.6785590685904026, "sign": 0.9760228917002678, "keep": 1.050151402130723, "leak": 0.9114925749599934, "replacement": 1.9195193350315094}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 2.501607894897461, "loss_scale": 4.0, "retries": 0, "seconds": 457.6, "max_memory_gb": 19.72, "pair": 0.15519062243402004, "sign": 0.08163472078740597, "keep": 0.07871011877432466, "leak": 0.7702512294054031, "replacement": 0.42792586237192154}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 3.557076930999756, "loss_scale": 4.0, "retries": 0, "seconds": 458.7, "max_memory_gb": 19.72, "pair": 0.2583807520568371, "sign": 0.20892254449427128, "keep": 0.10207714326679707, "leak": 0.7343210373073816, "replacement": 0.5974608510732651}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 2.6774044036865234, "loss_scale": 4.0, "retries": 0, "seconds": 459.8, "max_memory_gb": 19.72, "pair": 0.5775748640298843, "sign": 0.6165966768749058, "keep": 0.4720052648335695, "leak": 0.24050273839384317, "replacement": 1.1820015460252762}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 3.7091004848480225, "loss_scale": 4.0, "retries": 0, "seconds": 460.9, "max_memory_gb": 19.72, "pair": 0.2985789682716131, "sign": 0.313274510204792, "keep": 0.07394744595512748, "leak": 1.7292281687259674, "replacement": 0.9244970083236694}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 2.4784088134765625, "loss_scale": 4.0, "retries": 0, "seconds": 462.0, "max_memory_gb": 19.72, "pair": 0.21218041516840458, "sign": 0.12684163078665733, "keep": 0.014972081524319947, "leak": 0.5695647522807121, "replacement": 0.4254784658551216}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 2.91195011138916, "loss_scale": 4.0, "retries": 0, "seconds": 463.1, "max_memory_gb": 19.72, "pair": 0.23523119650781155, "sign": 0.08093048282898962, "keep": 0.03786637051962316, "leak": 0.47448888421058655, "replacement": 0.4132518470287323}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 2.5704245567321777, "loss_scale": 4.0, "retries": 0, "seconds": 464.2, "max_memory_gb": 19.72, "pair": 0.2662537097930908, "sign": 0.2493770457804203, "keep": 0.23585962317883968, "leak": 0.5473031364381313, "replacement": 0.645697832107544}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.14501428604126, "loss_scale": 4.0, "retries": 0, "seconds": 465.3, "max_memory_gb": 19.72, "pair": 0.16738148964941502, "sign": 0.06643323600292206, "keep": 0.29635887034237385, "leak": 0.9734317949041724, "replacement": 0.5921354964375496}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 2.564751148223877, "loss_scale": 4.0, "retries": 0, "seconds": 466.4, "max_memory_gb": 19.72, "pair": 0.2490966059267521, "sign": 0.2063358873128891, "keep": 0.10569085786119103, "leak": 0.5635689124537748, "replacement": 0.5460021942853928}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 2.5058023929595947, "loss_scale": 4.0, "retries": 0, "seconds": 467.4, "max_memory_gb": 19.72, "pair": 0.23697909340262413, "sign": 0.3025444932281971, "keep": 0.12172452826052904, "leak": 0.7981197386980057, "replacement": 0.6486435383558273}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 1.5441757440567017, "loss_scale": 4.0, "retries": 0, "seconds": 468.5, "max_memory_gb": 19.72, "pair": 0.1814010590314865, "sign": 0.14125493541359901, "keep": 0.010596077132504433, "leak": 0.12347901426255703, "replacement": 0.2881963178515434}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 5.149715900421143, "loss_scale": 4.0, "retries": 0, "seconds": 469.6, "max_memory_gb": 19.72, "pair": 0.446618914604187, "sign": 0.4125865940004587, "keep": 0.4100632122717798, "leak": 0.383930966258049, "replacement": 0.9539265520870686}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 3.097865343093872, "loss_scale": 4.0, "retries": 0, "seconds": 470.7, "max_memory_gb": 19.72, "pair": 0.2078198455274105, "sign": 0.24047251138836145, "keep": 0.15003058779984713, "leak": 0.6573221273720264, "replacement": 0.5674019232392311}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 8.871979713439941, "loss_scale": 4.0, "retries": 0, "seconds": 471.8, "max_memory_gb": 19.72, "pair": 0.3876786734908819, "sign": 0.4742111572995782, "keep": 0.5064978422597051, "leak": 0.7976389899849892, "replacement": 1.0774429589509964}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.8070012331008911, "loss_scale": 4.0, "retries": 0, "seconds": 472.8, "max_memory_gb": 19.72, "pair": 0.17633105255663395, "sign": 0.11260041780769825, "keep": 0.09129611391108483, "leak": 0.29919246956706047, "replacement": 0.3530774377286434}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 6.499906539916992, "loss_scale": 4.0, "retries": 0, "seconds": 473.9, "max_memory_gb": 19.72, "pair": 0.41274574398994446, "sign": 0.4699279488995671, "keep": 0.6989832641556859, "leak": 0.2815366864670068, "replacement": 1.0675855427980423}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 3.8204991817474365, "loss_scale": 4.0, "retries": 0, "seconds": 475.0, "max_memory_gb": 19.72, "pair": 0.22592086344957352, "sign": 0.31876106560230255, "keep": 0.04176268563605845, "leak": 0.4668523841537535, "replacement": 0.522895835340023}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 6.687256336212158, "loss_scale": 4.0, "retries": 0, "seconds": 476.1, "max_memory_gb": 19.72, "pair": 0.6564031355082989, "sign": 0.7289775088429451, "keep": 0.9066798489075154, "leak": 0.6731232907623053, "replacement": 1.6425126381218433}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 2.6000492572784424, "loss_scale": 4.0, "retries": 0, "seconds": 477.1, "max_memory_gb": 19.72, "pair": 0.41149983555078506, "sign": 0.4134440775960684, "keep": 0.07535824691876769, "leak": 1.203478030860424, "replacement": 0.9567705243825912}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 1.9949634075164795, "loss_scale": 4.0, "retries": 0, "seconds": 478.2, "max_memory_gb": 19.72, "pair": 0.18604053556919098, "sign": 0.2903939038515091, "keep": 0.06492037151474506, "leak": 0.5172034669667482, "replacement": 0.49299852550029755}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 3.7973766326904297, "loss_scale": 4.0, "retries": 0, "seconds": 479.4, "max_memory_gb": 19.72, "pair": 0.2561134174466133, "sign": 0.435547667555511, "keep": 0.25093383248895407, "leak": 0.33559808507561684, "replacement": 0.6832537204027176}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 4.002189636230469, "loss_scale": 4.0, "retries": 0, "seconds": 480.5, "max_memory_gb": 19.72, "pair": 0.6234248653054237, "sign": 0.8490796554833651, "keep": 0.9795795168029144, "leak": 0.942960251471959, "replacement": 1.7734944373369217}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 2.750370502471924, "loss_scale": 4.0, "retries": 0, "seconds": 481.5, "max_memory_gb": 19.72, "pair": 0.17707037180662155, "sign": 0.2042043674737215, "keep": 0.270644799224101, "leak": 0.408627268159762, "replacement": 0.5166517794132233}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 3.6898839473724365, "loss_scale": 4.0, "retries": 0, "seconds": 482.6, "max_memory_gb": 19.72, "pair": 0.4580684658139944, "sign": 0.6129045709967613, "keep": 0.5158187682973221, "leak": 0.962833085184684, "replacement": 1.2631383761763573}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 3.547865629196167, "loss_scale": 4.0, "retries": 0, "seconds": 483.6, "max_memory_gb": 19.72, "pair": 0.2035607062280178, "sign": 0.25200365483760834, "keep": 0.3523915004916489, "leak": 0.3770206435583532, "replacement": 0.6000134572386742}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 2.3572134971618652, "loss_scale": 4.0, "retries": 0, "seconds": 484.7, "max_memory_gb": 19.72, "pair": 0.11476832628250122, "sign": 0.18003236129879951, "keep": 0.04826281193527393, "leak": 0.8087599501013756, "replacement": 0.4311058893799782}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 3.572234869003296, "loss_scale": 4.0, "retries": 0, "seconds": 485.8, "max_memory_gb": 19.72, "pair": 0.2541023027151823, "sign": 0.1602820372208953, "keep": 0.07865087874233723, "leak": 0.7530321106314659, "replacement": 0.5618267953395844}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 4.606993675231934, "loss_scale": 4.0, "retries": 0, "seconds": 486.9, "max_memory_gb": 19.72, "pair": 0.12437363527715206, "sign": 0.11552352458238602, "keep": 0.03655203216476366, "leak": 1.0287930183112621, "replacement": 0.4576096571981907}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 5.347643852233887, "loss_scale": 4.0, "retries": 0, "seconds": 487.9, "max_memory_gb": 19.72, "pair": 0.4086206275969744, "sign": 0.5015887934714556, "keep": 0.5778415123932064, "leak": 1.0655668005347252, "replacement": 1.2147274762392044}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 10.611723899841309, "loss_scale": 4.0, "retries": 0, "seconds": 489.0, "max_memory_gb": 19.72, "pair": 0.22659760527312756, "sign": 0.2655646540224552, "keep": 0.0677521865000017, "leak": 1.035278458148241, "replacement": 0.6520756669342518}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 4.0481390953063965, "loss_scale": 4.0, "retries": 0, "seconds": 490.1, "max_memory_gb": 19.72, "pair": 0.31159239634871483, "sign": 0.25557573372498155, "keep": 0.3501956146210432, "leak": 0.5626234859228134, "replacement": 0.7551339194178581}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 4.252364158630371, "loss_scale": 4.0, "retries": 0, "seconds": 491.2, "max_memory_gb": 19.72, "pair": 0.15243532322347164, "sign": 0.33492850698530674, "keep": 0.5763840544968843, "leak": 0.8708468377590179, "replacement": 0.8258033096790314}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 5.53435754776001, "loss_scale": 4.0, "retries": 0, "seconds": 492.2, "max_memory_gb": 19.72, "pair": 0.24986224621534348, "sign": 0.6385449906811118, "keep": 0.04193472908809781, "leak": 1.3402442760998383, "replacement": 0.9251631796360016}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 4.741984844207764, "loss_scale": 4.0, "retries": 0, "seconds": 493.3, "max_memory_gb": 19.72, "pair": 0.12913869693875313, "sign": 0.2217035498470068, "keep": 0.06485586205963045, "leak": 0.5303286090493202, "replacement": 0.40500055253505707}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 4.8438801765441895, "loss_scale": 4.0, "retries": 0, "seconds": 494.4, "max_memory_gb": 19.72, "pair": 0.46118051931262016, "sign": 0.5670826658606529, "keep": 0.2727153589949012, "leak": 0.6918388868361944, "replacement": 1.0540392994880676}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 3.579986572265625, "loss_scale": 4.0, "retries": 0, "seconds": 495.4, "max_memory_gb": 19.72, "pair": 0.24601595848798752, "sign": 0.30412512086331844, "keep": 0.07836016081273556, "leak": 0.6375182915944606, "replacement": 0.5966381579637527}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 3.2541768550872803, "loss_scale": 4.0, "retries": 0, "seconds": 496.5, "max_memory_gb": 19.72, "pair": 0.25619750283658504, "sign": 0.27948649786412716, "keep": 0.03412210289388895, "leak": 0.7218635715544224, "replacement": 0.5934676676988602}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 4.800759792327881, "loss_scale": 4.0, "retries": 0, "seconds": 497.6, "max_memory_gb": 19.72, "pair": 0.2697113025933504, "sign": 0.6178229749202728, "keep": 0.5493513951078057, "leak": 1.0993017554283142, "replacement": 1.128123976290226}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 2.2180368900299072, "loss_scale": 4.0, "retries": 0, "seconds": 498.7, "max_memory_gb": 19.72, "pair": 0.19534287601709366, "sign": 0.23208871856331825, "keep": 0.019478660891763866, "leak": 0.6178117282688618, "replacement": 0.47557949647307396}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 1.581505537033081, "loss_scale": 4.0, "retries": 0, "seconds": 499.7, "max_memory_gb": 19.72, "pair": 0.14143148809671402, "sign": 0.16530387476086617, "keep": 0.013975883950479329, "leak": 0.2912140339612961, "replacement": 0.3038748763501644}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 3.6081318855285645, "loss_scale": 4.0, "retries": 0, "seconds": 500.9, "max_memory_gb": 19.72, "pair": 0.12811819463968277, "sign": 0.20300231780856848, "keep": 0.1013667922525201, "leak": 0.7747912630438805, "replacement": 0.47400055080652237}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 4.334670066833496, "loss_scale": 4.0, "retries": 0, "seconds": 502.0, "max_memory_gb": 19.72, "pair": 0.22566111013293266, "sign": 0.3627898474223912, "keep": 0.21296545211225748, "leak": 0.7809328325092793, "replacement": 0.7087719663977623}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 5.949573040008545, "loss_scale": 4.0, "retries": 0, "seconds": 503.1, "max_memory_gb": 19.72, "pair": 0.2961190640926361, "sign": 0.37343555549159646, "keep": 0.11640769173391163, "leak": 1.075873980182223, "replacement": 0.8100091740489006}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 3.5547878742218018, "loss_scale": 4.0, "retries": 0, "seconds": 504.2, "max_memory_gb": 19.72, "pair": 0.16039356403052807, "sign": 0.13047986663877964, "keep": 0.06603373680263758, "leak": 0.9984856322407722, "replacement": 0.5082717835903168}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 2.2330875396728516, "loss_scale": 4.0, "retries": 0, "seconds": 505.3, "max_memory_gb": 19.72, "pair": 0.14978022314608097, "sign": 0.12395886424928904, "keep": 0.03183137229643762, "leak": 0.3627113066613674, "replacement": 0.31835316494107246}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 5.605560779571533, "loss_scale": 4.0, "retries": 0, "seconds": 506.3, "max_memory_gb": 19.72, "pair": 0.37126377411186695, "sign": 0.17030008509755135, "keep": 0.5282201757654548, "leak": 0.4980791279813275, "replacement": 0.8450436666607857}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 3.4062726497650146, "loss_scale": 4.0, "retries": 0, "seconds": 507.4, "max_memory_gb": 19.72, "pair": 0.15390522591769695, "sign": 0.07581054233014584, "keep": 0.011276799894403666, "leak": 0.34713587554870173, "replacement": 0.28423285856842995}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 5.7440361976623535, "loss_scale": 4.0, "retries": 0, "seconds": 508.5, "max_memory_gb": 19.72, "pair": 0.4057256355881691, "sign": 0.7011942584067583, "keep": 0.6028869394212961, "leak": 0.36159367114305496, "replacement": 1.1481646746397018}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 3.464576005935669, "loss_scale": 4.0, "retries": 0, "seconds": 509.5, "max_memory_gb": 19.72, "pair": 0.2796195615082979, "sign": 0.2104897378012538, "keep": 0.1314864766318351, "leak": 0.15102108102291822, "replacement": 0.4883629381656647}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 14.11728572845459, "loss_scale": 4.0, "retries": 0, "seconds": 510.6, "max_memory_gb": 19.72, "pair": 0.24905969575047493, "sign": 0.20841876976191998, "keep": 0.17354504446848296, "leak": 0.3985305121168494, "replacement": 0.5396742597222328}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 3.0296597480773926, "loss_scale": 4.0, "retries": 0, "seconds": 511.7, "max_memory_gb": 19.72, "pair": 0.28161909990012646, "sign": 0.39116309583187103, "keep": 0.22948779520811513, "leak": 0.0767991702305153, "replacement": 0.6111443415284157}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 5.268613815307617, "loss_scale": 4.0, "retries": 0, "seconds": 512.7, "max_memory_gb": 19.72, "pair": 0.31780919060111046, "sign": 0.4430906530469656, "keep": 0.3245161882368848, "leak": 0.3495930805802345, "replacement": 0.789010863751173}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 2.6503703594207764, "loss_scale": 4.0, "retries": 0, "seconds": 513.8, "max_memory_gb": 19.72, "pair": 0.15765155851840973, "sign": 0.23884929157793522, "keep": 0.17754567554220557, "leak": 1.4603201821446419, "replacement": 0.7309290692210197}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 5.83267068862915, "loss_scale": 4.0, "retries": 0, "seconds": 514.9, "max_memory_gb": 19.72, "pair": 0.2590156029909849, "sign": 0.4018198102712631, "keep": 0.021839488530531526, "leak": 0.7298635765910149, "replacement": 0.6533111482858658}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 3.3157801628112793, "loss_scale": 4.0, "retries": 0, "seconds": 516.0, "max_memory_gb": 19.72, "pair": 0.25402143597602844, "sign": 0.22029756661504507, "keep": 0.012715949211269617, "leak": 0.43278951942920685, "replacement": 0.4787255749106407}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 5.360739707946777, "loss_scale": 4.0, "retries": 0, "seconds": 517.1, "max_memory_gb": 19.72, "pair": 0.5483848433941603, "sign": 0.5189280919730663, "keep": 0.5736147714778781, "leak": 0.37762189243221655, "replacement": 1.1890617534518242}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 3.302107810974121, "loss_scale": 4.0, "retries": 0, "seconds": 518.2, "max_memory_gb": 19.72, "pair": 0.4290534071624279, "sign": 0.5240143574774265, "keep": 0.4574994929134846, "leak": 0.10092723184789065, "replacement": 0.9450421556830406}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 4.499612331390381, "loss_scale": 4.0, "retries": 0, "seconds": 519.2, "max_memory_gb": 19.72, "pair": 0.13456275686621666, "sign": 0.13421859359368682, "keep": 0.05309151625260711, "leak": 0.9112692577764392, "replacement": 0.4560351148247719}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 3.0412917137145996, "loss_scale": 4.0, "retries": 0, "seconds": 520.3, "max_memory_gb": 19.72, "pair": 0.2106461115181446, "sign": 0.40761416405439377, "keep": 0.22342501231469214, "leak": 0.9036995489150286, "replacement": 0.7520905956625938}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 11.792757034301758, "loss_scale": 4.0, "retries": 0, "seconds": 521.4, "max_memory_gb": 19.72, "pair": 0.39628467336297035, "sign": 0.7090787515044212, "keep": 0.5498712696135044, "leak": 0.13390093145426363, "replacement": 1.05923493206501}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 11.558276176452637, "loss_scale": 4.0, "retries": 0, "seconds": 522.4, "max_memory_gb": 19.72, "pair": 0.6467435359954834, "sign": 0.9711972214281559, "keep": 1.2885134229436517, "leak": 0.48995316623404506, "replacement": 1.8990871608257294}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 6.494735240936279, "loss_scale": 4.0, "retries": 0, "seconds": 523.4, "max_memory_gb": 19.72, "pair": 0.3017675569280982, "sign": 0.5948863010853529, "keep": 0.07478971825912595, "leak": 1.1454265862703323, "replacement": 0.9229621887207031}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 6.442273139953613, "loss_scale": 4.0, "retries": 0, "seconds": 524.5, "max_memory_gb": 19.72, "pair": 0.4048187620937824, "sign": 0.6576980724930763, "keep": 0.7623704764991999, "leak": 0.6532116234302521, "replacement": 1.2781559973955154}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 3.233079671859741, "loss_scale": 4.0, "retries": 0, "seconds": 525.5, "max_memory_gb": 19.72, "pair": 0.1946422327309847, "sign": 0.15600526379421353, "keep": 0.06276404182426631, "leak": 0.8448015283793211, "replacement": 0.515227273106575}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 2.2545523643493652, "loss_scale": 4.0, "retries": 0, "seconds": 526.6, "max_memory_gb": 19.72, "pair": 0.22972740046679974, "sign": 0.46582359727472067, "keep": 0.29124270379543304, "leak": 0.6886082855053246, "replacement": 0.7804126292467117}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 4.915779113769531, "loss_scale": 4.0, "retries": 0, "seconds": 527.7, "max_memory_gb": 19.72, "pair": 0.2255275882780552, "sign": 0.3538215048611164, "keep": 0.6980339226720389, "leak": 1.0948143171235643, "replacement": 1.0251588448882103}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 5.657491683959961, "loss_scale": 4.0, "retries": 0, "seconds": 528.7, "max_memory_gb": 19.72, "pair": 0.3037751615047455, "sign": 0.33811836037784815, "keep": 0.03046207409352064, "leak": 0.8356441333889961, "replacement": 0.6969764232635498}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 6.245144844055176, "loss_scale": 4.0, "retries": 0, "seconds": 529.8, "max_memory_gb": 19.72, "pair": 0.1409854982048273, "sign": 0.32155774533748627, "keep": 0.20659518559114076, "leak": 1.9881058372557163, "replacement": 0.9020884521305561}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 4.243091583251953, "loss_scale": 4.0, "retries": 0, "seconds": 530.9, "max_memory_gb": 19.72, "pair": 0.1196037232875824, "sign": 0.09825164452195168, "keep": 0.023617494967766106, "leak": 1.3998880237340927, "replacement": 0.5305103063583374}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 4.934752464294434, "loss_scale": 4.0, "retries": 0, "seconds": 532.0, "max_memory_gb": 19.72, "pair": 0.5604722760617733, "sign": 0.7340696975588799, "keep": 0.3157532470941078, "leak": 0.4346107798628509, "replacement": 1.1940364241600037}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 3.4791908264160156, "loss_scale": 4.0, "retries": 0, "seconds": 533.1, "max_memory_gb": 19.72, "pair": 0.10135242156684399, "sign": 0.22173623088747263, "keep": 0.031005077296867967, "leak": 1.085954574868083, "replacement": 0.49921172857284546}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 5.231159210205078, "loss_scale": 4.0, "retries": 0, "seconds": 534.2, "max_memory_gb": 19.72, "pair": 0.14827040117233992, "sign": 0.29849306773394346, "keep": 0.030136171844787896, "leak": 0.696399487555027, "replacement": 0.4866848811507225}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 2.7781131267547607, "loss_scale": 4.0, "retries": 0, "seconds": 535.2, "max_memory_gb": 19.72, "pair": 0.13423260673880577, "sign": 0.060338811948895454, "keep": 0.06372462119907141, "leak": 0.820578932762146, "replacement": 0.401409063488245}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 16.75375747680664, "loss_scale": 4.0, "retries": 0, "seconds": 536.3, "max_memory_gb": 19.72, "pair": 0.37613560538738966, "sign": 0.7539606373757124, "keep": 1.4533358159824274, "leak": 1.588412020355463, "replacement": 1.8768867142498493}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 4.193229675292969, "loss_scale": 4.0, "retries": 0, "seconds": 537.3, "max_memory_gb": 19.72, "pair": 0.44075118750333786, "sign": 0.41428432799875736, "keep": 0.3037540512159467, "leak": 1.0868973564356565, "replacement": 1.0714947506785393}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 6.709609031677246, "loss_scale": 4.0, "retries": 0, "seconds": 538.4, "max_memory_gb": 19.72, "pair": 0.4445845354348421, "sign": 0.40559370536357164, "keep": 0.3891462739557028, "leak": 0.5970854759216309, "replacement": 0.991225965321064}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 9.59678840637207, "loss_scale": 4.0, "retries": 0, "seconds": 539.4, "max_memory_gb": 19.72, "pair": 0.45679746940732, "sign": 0.6097425394691527, "keep": 0.9165922680986114, "leak": 0.966968696564436, "replacement": 1.461707055568695}
diff --git a/runs/single_llava/replacement/L26_s13/adapter/README.md b/runs/single_llava/replacement/L26_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L26_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L26_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L26_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c95d10135f14b51edb6ae4665ba1963c0e8db44
--- /dev/null
+++ b/runs/single_llava/replacement/L26_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.26.self_attn.o_proj",
+    "model.language_model.layers.26.mlp.down_proj",
+    "model.language_model.layers.26.mlp.up_proj",
+    "model.language_model.layers.26.self_attn.k_proj",
+    "model.language_model.layers.26.mlp.gate_proj",
+    "model.language_model.layers.26.self_attn.q_proj",
+    "model.language_model.layers.26.self_attn.v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L26_s13/config.json b/runs/single_llava/replacement/L26_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..1a20aacc6224733d0c85d207a676a1d858f87301
--- /dev/null
+++ b/runs/single_llava/replacement/L26_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 26,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L26_s13/steps.jsonl b/runs/single_llava/replacement/L26_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..621d020ccb4d11acd08ef050f9d46dfabd4a76b2
--- /dev/null
+++ b/runs/single_llava/replacement/L26_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.20430444180965424, "loss_scale": 4.0, "retries": 0, "seconds": 1.1, "max_memory_gb": 17.99, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.3891367018222809, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 17.99, "pair": 1.0978868901729584, "sign": 0.5366470497101545, "keep": 5.374051658902317, "leak": 5.866295115453113e-06, "replacement": 4.053237676620483}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 0.6085540652275085, "loss_scale": 4.0, "retries": 0, "seconds": 2.9, "max_memory_gb": 18.02, "pair": 1.101164698600769, "sign": 0.5412735566496849, "keep": 10.805828928947449, "leak": 4.072451588399417e-06, "replacement": 6.774716854095459}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.5433093905448914, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 18.02, "pair": 1.096630334854126, "sign": 0.472622730769217, "keep": 7.671235589310527, "leak": 2.4740402295719832e-06, "replacement": 5.168560057878494}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.1622997224330902, "loss_scale": 4.0, "retries": 0, "seconds": 4.8, "max_memory_gb": 18.02, "pair": 1.0986254513263702, "sign": 0.7767439037561417, "keep": 0.1592267919331789, "leak": 1.9124380499224714e-06, "replacement": 1.566611260175705}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.2758665382862091, "loss_scale": 4.0, "retries": 0, "seconds": 5.7, "max_memory_gb": 18.02, "pair": 1.0982667207717896, "sign": 0.7379012405872345, "keep": 2.9353499934077263, "leak": 7.663131160740022e-06, "replacement": 2.934894233942032}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.7011210918426514, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 18.02, "pair": 1.0935260355472565, "sign": 0.5465041967108846, "keep": 6.879095623269677, "leak": 1.669701009632263e-05, "replacement": 4.806330144405365}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.4914099872112274, "loss_scale": 4.0, "retries": 0, "seconds": 7.6, "max_memory_gb": 18.02, "pair": 1.0966963171958923, "sign": 0.5887774080038071, "keep": 7.39281439781189, "leak": 1.726956134007196e-05, "replacement": 5.08749657869339}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.467032253742218, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 18.02, "pair": 1.096414566040039, "sign": 0.6189262038096786, "keep": 5.713345684111118, "leak": 2.0810614230981628e-05, "replacement": 4.262555778026581}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.28524503111839294, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 18.02, "pair": 1.092733770608902, "sign": 0.7807253301143646, "keep": 3.272082567214966, "leak": 4.4442382488796284e-05, "replacement": 3.119148939847946}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.6111869812011719, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 18.02, "pair": 1.0922412276268005, "sign": 0.4544236473739147, "keep": 5.271479671820998, "leak": 6.467181137637112e-05, "replacement": 3.9552089869976044}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.37710312008857727, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 18.02, "pair": 1.0947039723396301, "sign": 0.613068800419569, "keep": 3.0602020928636193, "leak": 5.6439244361605745e-05, "replacement": 2.9313535690307617}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.6137747764587402, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 18.02, "pair": 1.0939912497997284, "sign": 0.513218991458416, "keep": 5.358108721673489, "leak": 6.558913543130984e-05, "replacement": 4.029671460390091}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 0.8592668771743774, "loss_scale": 4.0, "retries": 0, "seconds": 13.1, "max_memory_gb": 18.02, "pair": 1.0893408060073853, "sign": 0.3995503429323435, "keep": 7.800333432853222, "leak": 0.00017007706310323556, "replacement": 5.189325243234634}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.540116548538208, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 18.02, "pair": 1.089871197938919, "sign": 0.6893543526530266, "keep": 5.1912329805782065, "leak": 0.00010421123897685902, "replacement": 4.0301909148693085}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.38818320631980896, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 18.02, "pair": 1.0934352278709412, "sign": 0.7842259407043457, "keep": 2.770485599292442, "leak": 5.636304945255688e-05, "replacement": 2.8708050549030304}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.3692871332168579, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 18.02, "pair": 1.0927081406116486, "sign": 0.7163590490818024, "keep": 3.1962128300219774, "leak": 0.00010518033624862255, "replacement": 3.049020379781723}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 0.7785565853118896, "loss_scale": 4.0, "retries": 0, "seconds": 16.8, "max_memory_gb": 18.05, "pair": 1.0879560112953186, "sign": 0.5427106395363808, "keep": 7.20158501714468, "leak": 0.00010107347770826891, "replacement": 4.960129231214523}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 1.3784759044647217, "loss_scale": 4.0, "retries": 0, "seconds": 17.8, "max_memory_gb": 18.05, "pair": 1.0740564167499542, "sign": 0.5418525077402592, "keep": 10.711880683898926, "leak": 0.0005438591979327612, "replacement": 6.701058864593506}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.3329309821128845, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 18.05, "pair": 1.0905050933361053, "sign": 0.4316210523247719, "keep": 0.024168651551008224, "leak": 0.00015854844241403043, "replacement": 1.3184396028518677}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 0.7307413816452026, "loss_scale": 4.0, "retries": 0, "seconds": 19.6, "max_memory_gb": 18.05, "pair": 1.0801885426044464, "sign": 0.3769558444619179, "keep": 5.557275545783341, "leak": 0.00010704556052587577, "replacement": 4.047331124544144}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 0.7854081988334656, "loss_scale": 4.0, "retries": 0, "seconds": 20.6, "max_memory_gb": 18.05, "pair": 1.0821344256401062, "sign": 0.5756367892026901, "keep": 5.9520544447004795, "leak": 0.0005933481793363171, "replacement": 4.346128463745117}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 0.9974822998046875, "loss_scale": 4.0, "retries": 0, "seconds": 21.5, "max_memory_gb": 18.05, "pair": 1.0699176788330078, "sign": 0.2541386168450117, "keep": 5.318813415244222, "leak": 0.0004854667877225438, "replacement": 3.8565150797367096}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 1.464908480644226, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 18.05, "pair": 1.0701234340667725, "sign": 0.26905862614512444, "keep": 10.056625843048096, "leak": 0.0002988923588418402, "replacement": 6.2330403327941895}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 1.347822904586792, "loss_scale": 4.0, "retries": 0, "seconds": 23.3, "max_memory_gb": 18.05, "pair": 1.0712584257125854, "sign": 0.3972640745341778, "keep": 8.47171350568533, "leak": 0.0007742382222473143, "replacement": 5.505940854549408}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 1.529727578163147, "loss_scale": 4.0, "retries": 0, "seconds": 24.3, "max_memory_gb": 18.05, "pair": 1.062268316745758, "sign": 0.6654579443857074, "keep": 10.55366063117981, "leak": 0.0013249778139652335, "replacement": 6.67215895652771}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.1740739345550537, "loss_scale": 4.0, "retries": 0, "seconds": 25.2, "max_memory_gb": 18.05, "pair": 1.0755049586296082, "sign": 0.6005290448665619, "keep": 8.446240581572056, "leak": 0.0010959602441289462, "replacement": 5.599163800477982}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 0.9227239489555359, "loss_scale": 4.0, "retries": 0, "seconds": 26.2, "max_memory_gb": 18.05, "pair": 1.0742377042770386, "sign": 0.7843079473823309, "keep": 5.656177416443825, "leak": 0.0018291450285525812, "replacement": 4.2949376702308655}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.035441517829895, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 18.05, "pair": 1.0718287229537964, "sign": 0.83172607421875, "keep": 5.2247534096241, "leak": 0.0010738238398744215, "replacement": 4.100336790084839}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.109506607055664, "loss_scale": 4.0, "retries": 0, "seconds": 28.0, "max_memory_gb": 18.05, "pair": 1.061064600944519, "sign": 0.3501255353912711, "keep": 5.547567192465067, "leak": 0.0016917096858508618, "replacement": 4.010333806276321}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.2012622356414795, "loss_scale": 4.0, "retries": 0, "seconds": 28.9, "max_memory_gb": 18.05, "pair": 1.0628867745399475, "sign": 0.8041585255414248, "keep": 5.71531069278717, "leak": 0.0018203973492063596, "replacement": 4.323076546192169}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 1.8445488214492798, "loss_scale": 4.0, "retries": 0, "seconds": 29.9, "max_memory_gb": 18.05, "pair": 1.0497530698776245, "sign": 0.3998531736433506, "keep": 10.753833532333374, "leak": 0.0036380466481205076, "replacement": 6.627505898475647}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.1228865385055542, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 18.05, "pair": 1.0550039410591125, "sign": 0.6284051798284054, "keep": 4.915312610566616, "leak": 0.002916550395614337, "replacement": 3.827592045068741}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.1961826086044312, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 18.05, "pair": 1.0591329634189606, "sign": 0.5325778350234032, "keep": 5.305371189489961, "leak": 0.0023490481544286013, "replacement": 3.9786947071552277}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.2722088098526, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 18.05, "pair": 1.040683537721634, "sign": 0.4351102691143751, "keep": 5.535813257098198, "leak": 0.004829098015761701, "replacement": 4.027352511882782}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.201364278793335, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 18.05, "pair": 1.0535306632518768, "sign": 0.4401222374290228, "keep": 5.178244933485985, "leak": 0.0033872473177325446, "replacement": 3.8635611534118652}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.3100864887237549, "loss_scale": 4.0, "retries": 0, "seconds": 34.5, "max_memory_gb": 18.05, "pair": 1.0406213402748108, "sign": 0.5895094275474548, "keep": 5.025770127773285, "leak": 0.0044513776701364804, "replacement": 3.8493740260601044}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 2.2431750297546387, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 18.05, "pair": 1.0188081860542297, "sign": 0.7828062903136015, "keep": 9.507750749588013, "leak": 0.007217691178084351, "replacement": 6.165891170501709}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 0.8723010420799255, "loss_scale": 4.0, "retries": 0, "seconds": 36.4, "max_memory_gb": 18.05, "pair": 1.0565072298049927, "sign": 0.5641198977828026, "keep": 2.9276847997680306, "leak": 0.005821937209475436, "replacement": 2.8038650453090668}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 1.3996001482009888, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 18.05, "pair": 1.0474109947681427, "sign": 0.5024632774293423, "keep": 5.517545208334923, "leak": 0.002812616099618026, "replacement": 4.058118522167206}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 0.8711141347885132, "loss_scale": 4.0, "retries": 0, "seconds": 38.2, "max_memory_gb": 18.05, "pair": 1.0831864774227142, "sign": 1.0256887078285217, "keep": 2.9472691118717194, "leak": 0.0006936238693706631, "replacement": 3.06983882188797}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.3804035186767578, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 18.05, "pair": 1.0397884547710419, "sign": 0.4057995695620775, "keep": 4.794277461245656, "leak": 0.004225801661959849, "replacement": 3.640883445739746}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 1.4850726127624512, "loss_scale": 4.0, "retries": 0, "seconds": 40.0, "max_memory_gb": 18.05, "pair": 1.029697060585022, "sign": 0.5166045688092709, "keep": 5.006544327363372, "leak": 0.0025057967359316535, "replacement": 3.7918980717658997}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 0.9214417934417725, "loss_scale": 4.0, "retries": 0, "seconds": 41.0, "max_memory_gb": 18.05, "pair": 1.046817198395729, "sign": 0.5818915255367756, "keep": 2.771879270672798, "leak": 0.0073674176674103364, "replacement": 2.725544363260269}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 0.8078120946884155, "loss_scale": 4.0, "retries": 0, "seconds": 41.9, "max_memory_gb": 18.05, "pair": 1.0707470774650574, "sign": 0.8947431594133377, "keep": 2.8716712817549706, "leak": 0.0021363736490604968, "replacement": 2.954488456249237}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 1.63094961643219, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 18.05, "pair": 1.0128856301307678, "sign": 0.7228897996246815, "keep": 4.632593244314194, "leak": 0.015911129786218225, "replacement": 3.6946049332618713}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 1.746590256690979, "loss_scale": 4.0, "retries": 0, "seconds": 43.7, "max_memory_gb": 18.05, "pair": 0.9995108395814896, "sign": 0.6941140033304691, "keep": 4.131288337521255, "leak": 0.011107659950539528, "replacement": 3.414988934993744}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 2.4230761528015137, "loss_scale": 4.0, "retries": 0, "seconds": 44.7, "max_memory_gb": 18.05, "pair": 0.9602736234664917, "sign": 0.20548435859382153, "keep": 7.742325454950333, "leak": 0.023368731126538478, "replacement": 4.940020650625229}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 2.424077033996582, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 18.05, "pair": 0.9701246619224548, "sign": 0.5875395908951759, "keep": 7.285223562270403, "leak": 0.013319738325662911, "replacement": 4.909836322069168}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 2.1924209594726562, "loss_scale": 4.0, "retries": 0, "seconds": 46.5, "max_memory_gb": 18.05, "pair": 1.0176812410354614, "sign": 0.44875358045101166, "keep": 7.390615912154317, "leak": 0.007690041034948081, "replacement": 4.939288467168808}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 0.9772678017616272, "loss_scale": 4.0, "retries": 0, "seconds": 47.4, "max_memory_gb": 18.05, "pair": 1.0328713059425354, "sign": 0.5842990763485432, "keep": 2.7613790296018124, "leak": 0.012062108004101901, "replacement": 2.7087258994579315}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 2.1061410903930664, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 18.05, "pair": 1.0117593258619308, "sign": 0.7106958031654358, "keep": 4.792087912559509, "leak": 0.011116502610093448, "replacement": 3.7659302949905396}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 2.3863112926483154, "loss_scale": 4.0, "retries": 0, "seconds": 49.2, "max_memory_gb": 18.05, "pair": 1.0178586840629578, "sign": 0.7769248634576797, "keep": 7.54023340344429, "leak": 0.02370707152022078, "replacement": 5.182364583015442}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 2.709512948989868, "loss_scale": 4.0, "retries": 0, "seconds": 50.1, "max_memory_gb": 18.05, "pair": 0.9625305235385895, "sign": 0.6473807319998741, "keep": 7.902719095349312, "leak": 0.045129584897949826, "replacement": 5.248862862586975}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 1.7636886835098267, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 18.05, "pair": 1.0023880451917648, "sign": 0.48086827620863914, "keep": 5.05144976824522, "leak": 0.016727160567825194, "replacement": 3.7727289497852325}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 3.58307147026062, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 18.05, "pair": 0.8652148991823196, "sign": 0.19463581405580044, "keep": 9.752703189849854, "leak": 0.03962384816259146, "replacement": 5.848790407180786}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 0.9955385327339172, "loss_scale": 4.0, "retries": 0, "seconds": 52.8, "max_memory_gb": 18.05, "pair": 1.0284764021635056, "sign": 0.7888024039566517, "keep": 2.5291433855891228, "leak": 0.0063563842268195, "replacement": 2.6890383064746857}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 2.173548460006714, "loss_scale": 4.0, "retries": 0, "seconds": 53.7, "max_memory_gb": 18.05, "pair": 0.9866317808628082, "sign": 0.28615523502230644, "keep": 4.915704787475988, "leak": 0.02684660945669748, "replacement": 3.59427347779274}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.0766834020614624, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 18.05, "pair": 1.0501421391963959, "sign": 0.6761246621608734, "keep": 2.860307961702347, "leak": 0.0005461581438555019, "replacement": 2.818494975566864}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 1.7486562728881836, "loss_scale": 4.0, "retries": 0, "seconds": 55.6, "max_memory_gb": 18.05, "pair": 0.9690860360860825, "sign": 0.6354436725378036, "keep": 4.910798886790872, "leak": 0.006846457268693484, "replacement": 3.743918865919113}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 2.858987808227539, "loss_scale": 4.0, "retries": 0, "seconds": 56.5, "max_memory_gb": 18.05, "pair": 0.9289922416210175, "sign": 0.6365292519330978, "keep": 7.178329035639763, "leak": 0.03873905820319479, "replacement": 4.846106112003326}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.1522276401519775, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 18.05, "pair": 1.0245368927717209, "sign": 0.7201215140521526, "keep": 2.7125297924503684, "leak": 0.006130139310243976, "replacement": 2.742395043373108}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 2.763787031173706, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 18.05, "pair": 0.9999464601278305, "sign": 0.6019776612520218, "keep": 7.497527176514268, "leak": 0.014000048425259592, "replacement": 5.053198873996735}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 2.897149085998535, "loss_scale": 4.0, "retries": 0, "seconds": 59.3, "max_memory_gb": 18.08, "pair": 0.9376189112663269, "sign": 0.5051973480731249, "keep": 7.428646881133318, "leak": 0.047370355532621033, "replacement": 4.916383504867554}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 2.724747657775879, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 18.08, "pair": 0.9673717468976974, "sign": 0.5752239041030407, "keep": 7.476969599723816, "leak": 0.041372566294739954, "replacement": 5.003811776638031}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.0245081186294556, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 18.08, "pair": 1.0572275519371033, "sign": 0.6461302936077118, "keep": 1.8377260868437588, "leak": 0.007091064779160661, "replacement": 2.300928622484207}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 3.320253610610962, "loss_scale": 4.0, "retries": 0, "seconds": 62.2, "max_memory_gb": 18.08, "pair": 0.8352111279964447, "sign": 0.3898668773472309, "keep": 7.029904335737228, "leak": 0.10022304976746454, "replacement": 4.570152640342712}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.240111231803894, "loss_scale": 4.0, "retries": 0, "seconds": 63.1, "max_memory_gb": 18.08, "pair": 1.0170746594667435, "sign": 0.5089586824178696, "keep": 2.4180264552123845, "leak": 0.01152229484102918, "replacement": 2.4834477305412292}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 2.1858620643615723, "loss_scale": 4.0, "retries": 0, "seconds": 64.0, "max_memory_gb": 18.08, "pair": 0.9342770129442215, "sign": 0.7530779764056206, "keep": 4.602698337286711, "leak": 0.020660640765242988, "replacement": 3.617330402135849}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 2.4189095497131348, "loss_scale": 4.0, "retries": 0, "seconds": 65.0, "max_memory_gb": 18.08, "pair": 0.969742089509964, "sign": 0.658214658498764, "keep": 4.605402372777462, "leak": 0.05241490508660718, "replacement": 3.6146543622016907}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 3.3241348266601562, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 18.08, "pair": 0.8346130847930908, "sign": 0.18875110894441605, "keep": 6.9236557353287935, "leak": 0.07405248534632847, "replacement": 4.4093297123909}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 3.9796745777130127, "loss_scale": 4.0, "retries": 0, "seconds": 66.8, "max_memory_gb": 18.08, "pair": 0.8025335669517517, "sign": 0.32083315029740334, "keep": 7.9371665716171265, "leak": 0.06906751322549098, "replacement": 4.948800206184387}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.43799570202827454, "loss_scale": 4.0, "retries": 0, "seconds": 67.7, "max_memory_gb": 18.08, "pair": 1.0656234920024872, "sign": 0.8184223920106888, "keep": 0.4433832634240389, "leak": 0.0012067975198988279, "replacement": 1.6968281269073486}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 2.439603328704834, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 18.08, "pair": 0.9020102024078369, "sign": 0.5042743990197778, "keep": 4.615098977461457, "leak": 0.0472239945529509, "replacement": 3.473502904176712}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 1.518140435218811, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 18.08, "pair": 0.9479281604290009, "sign": 0.905163362622261, "keep": 3.3712805658578873, "leak": 0.08345437946081802, "replacement": 3.107013612985611}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 2.6094870567321777, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 18.08, "pair": 0.8698274046182632, "sign": 0.45951980724930763, "keep": 4.273179544135928, "leak": 0.056942495168186724, "replacement": 3.2504126727581024}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 3.3776698112487793, "loss_scale": 4.0, "retries": 0, "seconds": 71.4, "max_memory_gb": 18.08, "pair": 0.8400082737207413, "sign": 0.4984002634882927, "keep": 7.147394239902496, "leak": 0.1522775900339184, "replacement": 4.700974941253662}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 2.496670722961426, "loss_scale": 4.0, "retries": 0, "seconds": 72.3, "max_memory_gb": 18.08, "pair": 0.9384032636880875, "sign": 0.7437423914670944, "keep": 3.6509306393563747, "leak": 0.08092276408569887, "replacement": 3.155970424413681}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 1.6381930112838745, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 18.08, "pair": 0.9569062441587448, "sign": 0.5895940437912941, "keep": 1.8537839208729565, "leak": 0.04177963019901654, "replacement": 2.1890400648117065}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 2.653886318206787, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 18.08, "pair": 0.8319081962108612, "sign": 0.3800410721451044, "keep": 4.11232777312398, "leak": 0.16949163330718875, "replacement": 3.1204655170440674}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 3.765644073486328, "loss_scale": 4.0, "retries": 0, "seconds": 75.0, "max_memory_gb": 18.08, "pair": 0.8204178065061569, "sign": 0.5544220097362995, "keep": 6.1130339950323105, "leak": 0.10632563814215246, "replacement": 4.180727273225784}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 3.9162251949310303, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 18.08, "pair": 0.7758855223655701, "sign": 0.3837550347670913, "keep": 6.2547144293785095, "leak": 0.1002239377562546, "replacement": 4.120176255702972}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 3.6940667629241943, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 18.08, "pair": 0.8027705699205399, "sign": 0.601576179265976, "keep": 5.203101897612214, "leak": 0.15729262540116906, "replacement": 3.744432717561722}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 1.6898682117462158, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 18.08, "pair": 0.9708167761564255, "sign": 0.8537562489509583, "keep": 2.263135534245521, "leak": 0.04350505732145393, "replacement": 2.540138900279999}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 0.8394008278846741, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 18.08, "pair": 1.0267685055732727, "sign": 0.5839048326015472, "keep": 0.15109542664140463, "leak": 0.008945610199589282, "replacement": 1.3965050280094147}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 4.158242225646973, "loss_scale": 4.0, "retries": 0, "seconds": 79.7, "max_memory_gb": 18.08, "pair": 0.9098340719938278, "sign": 0.7316464893519878, "keep": 5.85261806845665, "leak": 0.057536413371963135, "replacement": 4.216350436210632}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 4.153326034545898, "loss_scale": 4.0, "retries": 0, "seconds": 80.6, "max_memory_gb": 18.08, "pair": 0.8172809928655624, "sign": 0.5106425434350967, "keep": 5.73953265324235, "leak": 0.2547650309279561, "replacement": 4.006059944629669}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 2.837991714477539, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 18.08, "pair": 0.8831439316272736, "sign": 0.6262186923995614, "keep": 3.9746374383103102, "leak": 0.09118018732307576, "replacement": 3.2063669562339783}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 0.763258695602417, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 18.08, "pair": 1.0255070179700851, "sign": 0.704873114824295, "keep": 0.22165645414497703, "leak": 0.012396228717989288, "replacement": 1.4918708503246307}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 2.668790817260742, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 18.08, "pair": 0.889084979891777, "sign": 0.6423406675457954, "keep": 3.2954281978309155, "leak": 0.042828277400985826, "replacement": 2.868676483631134}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.3476603031158447, "loss_scale": 4.0, "retries": 0, "seconds": 84.2, "max_memory_gb": 18.08, "pair": 1.0654436945915222, "sign": 1.027376115322113, "keep": 0.638546533882618, "leak": 0.0026308337796763226, "replacement": 1.8990627527236938}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 1.8427850008010864, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 18.08, "pair": 0.9842733144760132, "sign": 0.6584658771753311, "keep": 1.930399565026164, "leak": 0.16982858889969066, "replacement": 2.3211632072925568}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 3.8476293087005615, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 18.08, "pair": 0.7919652312994003, "sign": 0.3130844309926033, "keep": 3.3167138835415244, "leak": 0.19121634756447747, "replacement": 2.6546685099601746}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 4.700286865234375, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 18.08, "pair": 0.7048856168985367, "sign": 0.4035170432180166, "keep": 5.610530331730843, "leak": 0.2269289580306122, "replacement": 3.7686414420604706}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 3.31327223777771, "loss_scale": 4.0, "retries": 0, "seconds": 87.9, "max_memory_gb": 18.08, "pair": 0.7893565818667412, "sign": 0.512047789990902, "keep": 3.8082538675516844, "leak": 0.11893178093305323, "replacement": 2.9792404174804688}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 4.684521675109863, "loss_scale": 4.0, "retries": 0, "seconds": 88.8, "max_memory_gb": 18.08, "pair": 0.6987020820379257, "sign": 0.5140399727970362, "keep": 5.4878112226724625, "leak": 0.4205624920923583, "replacement": 3.804768204689026}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 1.8265985250473022, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 18.08, "pair": 0.8950761407613754, "sign": 0.5481099411845207, "keep": 2.2718060202896595, "leak": 0.12319133066921495, "replacement": 2.33583202958107}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 3.2170937061309814, "loss_scale": 4.0, "retries": 0, "seconds": 90.6, "max_memory_gb": 18.08, "pair": 0.8317094668745995, "sign": 0.5291593745350838, "keep": 3.2789098117500544, "leak": 0.35108046466484666, "replacement": 2.8235141932964325}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.5467730760574341, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 18.08, "pair": 1.029814213514328, "sign": 0.877978578209877, "keep": 0.39969014725647867, "leak": 0.00692059809807688, "replacement": 1.6703787446022034}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 4.5637125968933105, "loss_scale": 4.0, "retries": 0, "seconds": 92.5, "max_memory_gb": 18.08, "pair": 0.8133765459060669, "sign": 0.7186129949986935, "keep": 4.976623363792896, "leak": 0.22459301723574754, "replacement": 3.717142879962921}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 5.044322967529297, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 18.08, "pair": 0.5926061570644379, "sign": 0.17962193256244063, "keep": 4.720948085188866, "leak": 0.2570502166636288, "replacement": 3.1071538031101227}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 4.8406982421875, "loss_scale": 4.0, "retries": 0, "seconds": 94.4, "max_memory_gb": 18.08, "pair": 0.7194699943065643, "sign": 0.3544136546552181, "keep": 5.2392261773347855, "leak": 0.2732827633008128, "replacement": 3.584610641002655}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 3.3870019912719727, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 18.08, "pair": 0.8014905601739883, "sign": 0.4029685743153095, "keep": 3.198997087776661, "leak": 0.18082601600326598, "replacement": 2.6476797461509705}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 2.549837350845337, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 18.08, "pair": 0.9849059730768204, "sign": 0.9908078759908676, "keep": 1.6235106344684027, "leak": 0.18574879138031974, "replacement": 2.3385024964809418}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 4.973724842071533, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 18.08, "pair": 0.7283237427473068, "sign": 0.41351553425192833, "keep": 4.463408425450325, "leak": 0.7101390741881914, "replacement": 3.3443205654621124}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 5.367440700531006, "loss_scale": 4.0, "retries": 0, "seconds": 98.1, "max_memory_gb": 18.08, "pair": 0.7076429352164268, "sign": 0.6104146800935268, "keep": 4.123089104890823, "leak": 0.7160768549219938, "replacement": 3.253414034843445}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.6336920261383057, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 18.08, "pair": 0.9646996110677719, "sign": 0.8309227377176285, "keep": 1.661091735586524, "leak": 0.09592500273720361, "replacement": 2.2346880733966827}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 1.8842566013336182, "loss_scale": 4.0, "retries": 0, "seconds": 99.9, "max_memory_gb": 18.08, "pair": 0.9140075594186783, "sign": 0.7437527105212212, "keep": 1.6798840737901628, "leak": 0.24310984961448412, "replacement": 2.1866034269332886}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 5.396077632904053, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 18.08, "pair": 0.5780918151140213, "sign": 0.4475475372746587, "keep": 4.447541058063507, "leak": 0.7045382735504973, "replacement": 3.201770603656769}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 4.061769962310791, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 18.08, "pair": 0.7404837608337402, "sign": 0.5508752148598433, "keep": 2.887045592069626, "leak": 0.28916502268475597, "replacement": 2.5317353904247284}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 3.8859071731567383, "loss_scale": 4.0, "retries": 0, "seconds": 102.7, "max_memory_gb": 18.08, "pair": 0.6788956820964813, "sign": 0.43675861321389675, "keep": 2.4319638828746974, "leak": 0.18417840590700507, "replacement": 2.1593015789985657}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 4.373366832733154, "loss_scale": 4.0, "retries": 0, "seconds": 103.7, "max_memory_gb": 18.08, "pair": 0.8035828769207001, "sign": 0.48689716681838036, "keep": 2.509360722731799, "leak": 0.13562431081663817, "replacement": 2.3356178402900696}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 4.365772724151611, "loss_scale": 4.0, "retries": 0, "seconds": 104.6, "max_memory_gb": 18.08, "pair": 0.7631739601492882, "sign": 0.5625817254185677, "keep": 2.7870757058262825, "leak": 0.2294898439112174, "replacement": 2.4953751266002655}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 6.143136501312256, "loss_scale": 4.0, "retries": 0, "seconds": 105.5, "max_memory_gb": 18.08, "pair": 0.6263466328382492, "sign": 0.5254000825807452, "keep": 3.6017485857009888, "leak": 0.7212276720965747, "replacement": 2.870227873325348}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 3.9642086029052734, "loss_scale": 4.0, "retries": 0, "seconds": 106.5, "max_memory_gb": 18.08, "pair": 0.7635178491473198, "sign": 0.5822860226035118, "keep": 2.664781227707863, "leak": 0.5574526368573061, "replacement": 2.5264146625995636}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 0.7321425676345825, "loss_scale": 4.0, "retries": 0, "seconds": 107.4, "max_memory_gb": 18.08, "pair": 0.9899483621120453, "sign": 0.7176705375313759, "keep": 0.23589849192649126, "leak": 0.017520053890621057, "replacement": 1.4711129069328308}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 4.202335357666016, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 18.08, "pair": 0.8518420159816742, "sign": 0.4661990664899349, "keep": 2.455071587464772, "leak": 0.3473346377722919, "replacement": 2.3993110358715057}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 2.428288221359253, "loss_scale": 4.0, "retries": 0, "seconds": 109.2, "max_memory_gb": 18.08, "pair": 0.8596402108669281, "sign": 0.5493967793881893, "keep": 1.3328736904077232, "leak": 0.0705399379934113, "replacement": 1.8184104561805725}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 2.6567201614379883, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 18.08, "pair": 0.8930817395448685, "sign": 0.784358024597168, "keep": 0.9572901874780655, "leak": 0.14601306103031675, "replacement": 1.8004091084003448}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 4.191205024719238, "loss_scale": 4.0, "retries": 0, "seconds": 111.1, "max_memory_gb": 18.08, "pair": 0.7348876371979713, "sign": 0.535053146071732, "keep": 2.078266831114888, "leak": 0.3088095281855203, "replacement": 2.1187499463558197}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.0640227794647217, "loss_scale": 4.0, "retries": 0, "seconds": 112.1, "max_memory_gb": 18.08, "pair": 0.899029478430748, "sign": 0.7688510566949844, "keep": 1.394679805263877, "leak": 1.0599436832708307, "replacement": 2.245780795812607}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 4.067654132843018, "loss_scale": 4.0, "retries": 0, "seconds": 113.0, "max_memory_gb": 18.08, "pair": 0.7917928248643875, "sign": 0.7112045902758837, "keep": 1.4813158651813865, "leak": 0.3476016099957633, "replacement": 1.9749534726142883}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.584042549133301, "loss_scale": 4.0, "retries": 0, "seconds": 113.9, "max_memory_gb": 18.08, "pair": 0.9277344048023224, "sign": 0.8050377871841192, "keep": 1.8030044566839933, "leak": 0.03491408720583422, "replacement": 2.2404839992523193}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 2.3504672050476074, "loss_scale": 4.0, "retries": 0, "seconds": 114.9, "max_memory_gb": 18.08, "pair": 0.9859337955713272, "sign": 1.0159407258033752, "keep": 0.8169888583943248, "leak": 0.004523476229223888, "replacement": 1.9035294651985168}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 2.5856828689575195, "loss_scale": 4.0, "retries": 0, "seconds": 115.7, "max_memory_gb": 18.08, "pair": 0.8990656286478043, "sign": 0.8312741816043854, "keep": 0.8699501380324364, "leak": 0.10111286336905323, "replacement": 1.7749559879302979}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 4.317740440368652, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 18.08, "pair": 0.7346967235207558, "sign": 0.6205835537984967, "keep": 2.0205952525138855, "leak": 0.09297925236774063, "replacement": 2.0785309076309204}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 2.4614310264587402, "loss_scale": 4.0, "retries": 0, "seconds": 117.6, "max_memory_gb": 18.08, "pair": 0.7549597397446632, "sign": 0.3984924890100956, "keep": 0.6484679067507386, "leak": 0.2641608938574791, "replacement": 1.3444801568984985}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 3.992464542388916, "loss_scale": 4.0, "retries": 0, "seconds": 118.5, "max_memory_gb": 18.08, "pair": 0.5984445214271545, "sign": 0.3479398740455508, "keep": 1.047091354150325, "leak": 1.028389147715643, "replacement": 1.5530573725700378}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 6.487461090087891, "loss_scale": 4.0, "retries": 0, "seconds": 119.4, "max_memory_gb": 18.08, "pair": 0.5579602941870689, "sign": 0.2923852978274226, "keep": 1.6688509583473206, "leak": 1.07411352917552, "replacement": 1.8071068227291107}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 4.110263347625732, "loss_scale": 4.0, "retries": 0, "seconds": 120.3, "max_memory_gb": 18.08, "pair": 0.7097188718616962, "sign": 0.8371566999703646, "keep": 1.3996139094233513, "leak": 0.6087268903502263, "replacement": 1.980285882949829}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 2.0542426109313965, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 18.08, "pair": 0.82693300396204, "sign": 0.7176511436700821, "keep": 0.8160677254199982, "leak": 0.20700006851438957, "replacement": 1.6455424427986145}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 5.10151481628418, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 18.08, "pair": 0.6591338440775871, "sign": 0.165133036673069, "keep": 1.7605376788415015, "leak": 0.1730756964534521, "replacement": 1.665238082408905}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 4.702840805053711, "loss_scale": 4.0, "retries": 0, "seconds": 123.1, "max_memory_gb": 18.08, "pair": 0.5668906681239605, "sign": 0.4260461814701557, "keep": 1.4079479575157166, "leak": 1.4443617937031377, "replacement": 1.8449782133102417}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.9778816103935242, "loss_scale": 4.0, "retries": 0, "seconds": 124.0, "max_memory_gb": 18.08, "pair": 0.9487025439739227, "sign": 0.731594517827034, "keep": 0.36840449552983046, "leak": 0.06369967386854114, "replacement": 1.5146269798278809}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 4.006744861602783, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 18.08, "pair": 0.5465986020863056, "sign": 0.4640005100518465, "keep": 1.0182814002037048, "leak": 2.0263012669638556, "replacement": 1.7943148612976074}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 3.051851511001587, "loss_scale": 4.0, "retries": 0, "seconds": 125.8, "max_memory_gb": 18.08, "pair": 0.8060698211193085, "sign": 0.51526147313416, "keep": 0.5200893115252256, "leak": 0.19626431692449842, "replacement": 1.3728112876415253}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 2.8722119331359863, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 18.08, "pair": 0.7165940776467323, "sign": 0.48485036566853523, "keep": 0.5696734022349119, "leak": 1.6460574697557604, "replacement": 1.6553703248500824}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 3.0179412364959717, "loss_scale": 4.0, "retries": 0, "seconds": 127.7, "max_memory_gb": 18.08, "pair": 0.709594763815403, "sign": 0.3412346728146076, "keep": 0.6171598909422755, "leak": 0.051167350728064775, "replacement": 1.2015838921070099}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 3.8207738399505615, "loss_scale": 4.0, "retries": 0, "seconds": 128.7, "max_memory_gb": 18.08, "pair": 0.654446192085743, "sign": 0.6960017532110214, "keep": 1.0378619581460953, "leak": 0.6531786585692316, "replacement": 1.684672698378563}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 3.5181453227996826, "loss_scale": 4.0, "retries": 0, "seconds": 129.6, "max_memory_gb": 18.08, "pair": 0.7041589207947254, "sign": 0.6201281100511551, "keep": 0.3872460015118122, "leak": 0.7799402214586735, "replacement": 1.4028310179710388}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 2.776585102081299, "loss_scale": 4.0, "retries": 0, "seconds": 130.5, "max_memory_gb": 18.08, "pair": 0.7877612709999084, "sign": 0.6067065745592117, "keep": 0.3952900152653456, "leak": 1.0374326508608647, "replacement": 1.5481177270412445}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 2.3655588626861572, "loss_scale": 4.0, "retries": 0, "seconds": 131.4, "max_memory_gb": 18.08, "pair": 0.6702181845903397, "sign": 0.4264768874272704, "keep": 0.37961566960439086, "leak": 0.258891696241335, "replacement": 1.1379873752593994}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.6285443305969238, "loss_scale": 4.0, "retries": 0, "seconds": 132.4, "max_memory_gb": 18.08, "pair": 0.9915387630462646, "sign": 1.0272632837295532, "keep": 0.25624779239296913, "leak": 0.05728749171976233, "replacement": 1.6476161777973175}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 3.4284114837646484, "loss_scale": 4.0, "retries": 0, "seconds": 133.3, "max_memory_gb": 18.08, "pair": 0.6113454923033714, "sign": 0.421254999935627, "keep": 0.7433980368077755, "leak": 0.8314193424448604, "replacement": 1.4015268087387085}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 2.7631585597991943, "loss_scale": 4.0, "retries": 0, "seconds": 134.3, "max_memory_gb": 18.08, "pair": 0.5531102828681469, "sign": 0.30918548069894314, "keep": 0.2916687927208841, "leak": 0.847285283729434, "replacement": 1.0653586983680725}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 2.013188600540161, "loss_scale": 4.0, "retries": 0, "seconds": 135.2, "max_memory_gb": 18.08, "pair": 0.6509607806801796, "sign": 0.5565085904672742, "keep": 0.3578678108751774, "leak": 0.4028538738784846, "replacement": 1.2088624686002731}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 2.4700794219970703, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 18.08, "pair": 0.6327286958694458, "sign": 0.5280623510479927, "keep": 0.5039985184557736, "leak": 0.6305784332798794, "replacement": 1.306403785943985}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 1.6265292167663574, "loss_scale": 4.0, "retries": 0, "seconds": 137.0, "max_memory_gb": 18.08, "pair": 0.7191789224743843, "sign": 0.35383545607328415, "keep": 0.11342487228102982, "leak": 0.17186038568615913, "replacement": 0.9957741945981979}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 1.4509973526000977, "loss_scale": 4.0, "retries": 0, "seconds": 138.0, "max_memory_gb": 18.08, "pair": 0.770623117685318, "sign": 0.5330395847558975, "keep": 0.2321803974919021, "leak": 0.08470560808200389, "replacement": 1.1744095087051392}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 3.0096068382263184, "loss_scale": 4.0, "retries": 0, "seconds": 138.9, "max_memory_gb": 18.08, "pair": 0.5487310811877251, "sign": 0.3715525735169649, "keep": 0.5820462629199028, "leak": 1.5113351311447332, "replacement": 1.4033643305301666}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 2.217822313308716, "loss_scale": 4.0, "retries": 0, "seconds": 139.8, "max_memory_gb": 18.08, "pair": 0.7566523104906082, "sign": 0.4921076074242592, "keep": 0.3492208067327738, "leak": 0.1480431705713272, "replacement": 1.2143272757530212}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 3.0697240829467773, "loss_scale": 4.0, "retries": 0, "seconds": 140.8, "max_memory_gb": 18.08, "pair": 0.4324900545179844, "sign": 0.26212419010698795, "keep": 0.2213565744459629, "leak": 1.8348330622538924, "replacement": 1.132938727736473}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 2.45536470413208, "loss_scale": 4.0, "retries": 0, "seconds": 141.7, "max_memory_gb": 18.08, "pair": 0.5756844207644463, "sign": 0.31068716384470463, "keep": 0.13088050857186317, "leak": 1.2130386251956224, "replacement": 1.0997278690338135}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 2.9733834266662598, "loss_scale": 4.0, "retries": 0, "seconds": 142.7, "max_memory_gb": 18.08, "pair": 0.5125684402883053, "sign": 0.26133652962744236, "keep": 0.8465240225195885, "leak": 0.8544018279062584, "replacement": 1.2800991982221603}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 2.9729490280151367, "loss_scale": 4.0, "retries": 0, "seconds": 143.6, "max_memory_gb": 18.08, "pair": 0.7196492254734039, "sign": 0.7046431079506874, "keep": 1.4344109892845154, "leak": 0.8626390190329403, "replacement": 2.004835993051529}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 1.590311884880066, "loss_scale": 4.0, "retries": 0, "seconds": 144.6, "max_memory_gb": 18.08, "pair": 0.7255420237779617, "sign": 0.5722531788051128, "keep": 0.3090603125747293, "leak": 0.3438465050821833, "replacement": 1.2521604597568512}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 3.0077972412109375, "loss_scale": 4.0, "retries": 0, "seconds": 145.5, "max_memory_gb": 18.08, "pair": 0.9653626680374146, "sign": 1.317508652806282, "keep": 0.33727971464395523, "leak": 0.14192648205789737, "replacement": 1.8282385170459747}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.3420504331588745, "loss_scale": 4.0, "retries": 0, "seconds": 146.4, "max_memory_gb": 18.08, "pair": 0.7764231897890568, "sign": 0.7241501808166504, "keep": 0.4791885204613209, "leak": 0.2617115232424112, "replacement": 1.4435204565525055}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 3.275177478790283, "loss_scale": 4.0, "retries": 0, "seconds": 147.3, "max_memory_gb": 18.08, "pair": 0.49418575689196587, "sign": 0.22691901586949825, "keep": 0.11386529635637999, "leak": 1.8049442805349827, "replacement": 1.115813985466957}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 3.4432151317596436, "loss_scale": 4.0, "retries": 0, "seconds": 148.2, "max_memory_gb": 18.08, "pair": 0.48222094774246216, "sign": 0.3684785373043269, "keep": 0.23790495470166206, "leak": 1.5251950542781287, "replacement": 1.1667114943265915}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 1.659379005432129, "loss_scale": 4.0, "retries": 0, "seconds": 149.2, "max_memory_gb": 18.08, "pair": 0.7232127748429775, "sign": 0.6242681555449963, "keep": 0.3797446363605559, "leak": 0.46448548373155063, "replacement": 1.3413405418395996}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 1.3760091066360474, "loss_scale": 4.0, "retries": 0, "seconds": 150.1, "max_memory_gb": 18.08, "pair": 0.7235964573919773, "sign": 0.6404927298426628, "keep": 0.14850584790110588, "leak": 0.27078502462245524, "replacement": 1.1857920363545418}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 3.4674203395843506, "loss_scale": 4.0, "retries": 0, "seconds": 150.9, "max_memory_gb": 18.08, "pair": 0.513031005859375, "sign": 0.36839551106095314, "keep": 0.1909419558942318, "leak": 0.5352305478299968, "replacement": 0.926507368683815}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 2.0743463039398193, "loss_scale": 4.0, "retries": 0, "seconds": 151.9, "max_memory_gb": 18.08, "pair": 0.7896331623196602, "sign": 0.8487347047775984, "keep": 0.5369588434696198, "leak": 0.21987565064046066, "replacement": 1.5374488532543182}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 2.175380229949951, "loss_scale": 4.0, "retries": 0, "seconds": 152.8, "max_memory_gb": 18.08, "pair": 0.7559615522623062, "sign": 0.6806550733745098, "keep": 0.29797304421663284, "leak": 0.8680587932467461, "replacement": 1.462290346622467}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 1.6062283515930176, "loss_scale": 4.0, "retries": 0, "seconds": 153.8, "max_memory_gb": 18.08, "pair": 0.8013586848974228, "sign": 0.6381008625030518, "keep": 0.3530006995424628, "leak": 0.29441367140680086, "replacement": 1.3705129027366638}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 2.2343673706054688, "loss_scale": 4.0, "retries": 0, "seconds": 154.7, "max_memory_gb": 18.08, "pair": 0.4557283967733383, "sign": 0.4024450993165374, "keep": 0.51379534881562, "leak": 1.0328469006344676, "replacement": 1.1720603555440903}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 2.9736616611480713, "loss_scale": 4.0, "retries": 0, "seconds": 155.6, "max_memory_gb": 18.08, "pair": 0.6507727205753326, "sign": 0.5099172070622444, "keep": 0.1772908614948392, "leak": 0.8555817157030106, "replacement": 1.2082721441984177}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 4.827003479003906, "loss_scale": 4.0, "retries": 0, "seconds": 156.6, "max_memory_gb": 18.08, "pair": 0.399955578148365, "sign": 0.4673080937936902, "keep": 0.7288225963711739, "leak": 2.2989524726290256, "replacement": 1.572759062051773}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 2.9662435054779053, "loss_scale": 4.0, "retries": 0, "seconds": 157.5, "max_memory_gb": 18.08, "pair": 0.37605113722383976, "sign": 0.1638796296902001, "keep": 0.0560084180906415, "leak": 1.4586918875575066, "replacement": 0.8506681174039841}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 2.699092388153076, "loss_scale": 4.0, "retries": 0, "seconds": 158.4, "max_memory_gb": 18.08, "pair": 0.5524421334266663, "sign": 0.4187507126480341, "keep": 0.09785054088570178, "leak": 1.3229650328867137, "replacement": 1.1414839625358582}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 4.188480377197266, "loss_scale": 4.0, "retries": 0, "seconds": 159.4, "max_memory_gb": 18.08, "pair": 0.3088381625711918, "sign": 0.12699328362941742, "keep": 0.1893379231914878, "leak": 2.587179121095687, "replacement": 1.1137985587120056}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 4.460096836090088, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 18.08, "pair": 0.4166543632745743, "sign": 0.38732171431183815, "keep": 0.3085769657045603, "leak": 1.2129732333123684, "replacement": 1.0678469985723495}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 1.3771029710769653, "loss_scale": 4.0, "retries": 0, "seconds": 161.2, "max_memory_gb": 18.08, "pair": 0.5603987649083138, "sign": 0.34840553998947144, "keep": 0.3029540544375777, "leak": 0.6004599630832672, "replacement": 1.0361935496330261}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 2.1941115856170654, "loss_scale": 4.0, "retries": 0, "seconds": 162.1, "max_memory_gb": 18.08, "pair": 0.5947296805679798, "sign": 0.6028969027101994, "keep": 0.3340512327849865, "leak": 0.820572423748672, "replacement": 1.2683468759059906}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 5.28812313079834, "loss_scale": 4.0, "retries": 0, "seconds": 163.0, "max_memory_gb": 18.08, "pair": 0.26536752469837666, "sign": 0.2006472498178482, "keep": 0.06169295497238636, "leak": 2.834512174129486, "replacement": 1.1051656603813171}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 1.97707200050354, "loss_scale": 4.0, "retries": 0, "seconds": 163.9, "max_memory_gb": 18.08, "pair": 0.48611390590667725, "sign": 0.3483570273965597, "keep": 0.273112874943763, "leak": 0.5674203072703676, "replacement": 0.9387039393186569}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 2.2421658039093018, "loss_scale": 4.0, "retries": 0, "seconds": 164.9, "max_memory_gb": 18.08, "pair": 0.7503446266055107, "sign": 0.6295682564377785, "keep": 0.37950684363022447, "leak": 0.8984555666102096, "replacement": 1.4794960618019104}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 1.6340910196304321, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 18.08, "pair": 0.7512690797448158, "sign": 0.7768184766173363, "keep": 0.6167791737243533, "leak": 0.4696789395900396, "replacement": 1.5654876083135605}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 2.6598918437957764, "loss_scale": 4.0, "retries": 0, "seconds": 166.6, "max_memory_gb": 18.08, "pair": 0.5721235498785973, "sign": 0.47830310743302107, "keep": 0.5424576168879867, "leak": 0.48124360269866884, "replacement": 1.2028148472309113}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 3.426691770553589, "loss_scale": 4.0, "retries": 0, "seconds": 167.6, "max_memory_gb": 18.08, "pair": 0.2913704141974449, "sign": 0.16687698010355234, "keep": 0.09269363060593605, "leak": 1.4985954947769642, "replacement": 0.7958045825362206}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 2.0679657459259033, "loss_scale": 4.0, "retries": 0, "seconds": 168.5, "max_memory_gb": 18.08, "pair": 0.7505518421530724, "sign": 0.6237285546958447, "keep": 0.24449395900592208, "leak": 0.5076017914398108, "replacement": 1.3115635812282562}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 2.7292840480804443, "loss_scale": 4.0, "retries": 0, "seconds": 169.4, "max_memory_gb": 18.08, "pair": 0.42487063631415367, "sign": 0.1910522854886949, "keep": 0.02805443573743105, "leak": 0.6976497769355774, "replacement": 0.7088364362716675}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 4.478233814239502, "loss_scale": 4.0, "retries": 0, "seconds": 170.4, "max_memory_gb": 18.08, "pair": 0.582878615707159, "sign": 0.35708205262199044, "keep": 0.11215912364423275, "leak": 1.5653785644099116, "replacement": 1.208843857049942}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 4.299640655517578, "loss_scale": 4.0, "retries": 0, "seconds": 171.4, "max_memory_gb": 18.08, "pair": 0.4024541862308979, "sign": 0.29805505741387606, "keep": 0.06915723159909248, "leak": 2.477516246493906, "replacement": 1.2054393887519836}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 2.2873713970184326, "loss_scale": 4.0, "retries": 0, "seconds": 172.3, "max_memory_gb": 18.08, "pair": 0.4471853896975517, "sign": 0.2556554954499006, "keep": 0.09029399626888335, "leak": 0.6479155197739601, "replacement": 0.7821390181779861}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 3.4643123149871826, "loss_scale": 4.0, "retries": 0, "seconds": 173.3, "max_memory_gb": 18.08, "pair": 0.44928595423698425, "sign": 0.10977627895772457, "keep": 0.2160252733156085, "leak": 0.6453221894335002, "replacement": 0.7735172882676125}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.8916382789611816, "loss_scale": 4.0, "retries": 0, "seconds": 174.2, "max_memory_gb": 18.08, "pair": 0.8051733374595642, "sign": 0.6617717891931534, "keep": 0.27336651959922165, "leak": 0.17675187622080557, "replacement": 1.3169304430484772}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 2.5786969661712646, "loss_scale": 4.0, "retries": 0, "seconds": 175.1, "max_memory_gb": 18.08, "pair": 0.43765558302402496, "sign": 0.2673593107610941, "keep": 0.057012093253433704, "leak": 1.0370718780905008, "replacement": 0.8591092526912689}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 4.4278564453125, "loss_scale": 4.0, "retries": 0, "seconds": 176.0, "max_memory_gb": 18.08, "pair": 0.20855509489774704, "sign": 0.14689651876688004, "keep": 0.05378189217299223, "leak": 1.5077736154198647, "replacement": 0.6858376935124397}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 1.8854897022247314, "loss_scale": 4.0, "retries": 0, "seconds": 176.9, "max_memory_gb": 18.08, "pair": 0.705691609531641, "sign": 0.6351833995431662, "keep": 0.5885611232370138, "leak": 0.406647976487875, "replacement": 1.4192258417606354}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 3.2720165252685547, "loss_scale": 4.0, "retries": 0, "seconds": 177.8, "max_memory_gb": 18.08, "pair": 0.5829058885574341, "sign": 0.5057358127087355, "keep": 0.4085181087721139, "leak": 0.8669363642593453, "replacement": 1.2567669302225113}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 2.8821611404418945, "loss_scale": 4.0, "retries": 0, "seconds": 178.7, "max_memory_gb": 18.08, "pair": 0.3641744703054428, "sign": 0.19427653029561043, "keep": 0.06590088468510658, "leak": 1.190652072429657, "replacement": 0.7919261902570724}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 2.025034189224243, "loss_scale": 4.0, "retries": 0, "seconds": 179.7, "max_memory_gb": 18.08, "pair": 0.7322860285639763, "sign": 0.7702117096632719, "keep": 0.5639582928270102, "leak": 0.46828424429986626, "replacement": 1.5164420902729034}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 4.073199272155762, "loss_scale": 4.0, "retries": 0, "seconds": 180.6, "max_memory_gb": 18.08, "pair": 0.539742037653923, "sign": 0.3707070117816329, "keep": 0.182241914793849, "leak": 1.4003447225113632, "replacement": 1.1663026809692383}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 5.582300662994385, "loss_scale": 4.0, "retries": 0, "seconds": 181.6, "max_memory_gb": 18.08, "pair": 0.2683965377509594, "sign": 0.13223921228200197, "keep": 0.07086957711726427, "leak": 1.8803181648254395, "replacement": 0.8400304615497589}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 3.847616195678711, "loss_scale": 4.0, "retries": 0, "seconds": 182.5, "max_memory_gb": 18.08, "pair": 0.5777133014053106, "sign": 0.5714874695986509, "keep": 0.24155119620263577, "leak": 1.127469516126439, "replacement": 1.2661000490188599}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 2.372626543045044, "loss_scale": 4.0, "retries": 0, "seconds": 183.4, "max_memory_gb": 18.08, "pair": 0.6086818166077137, "sign": 0.6044331453740597, "keep": 0.34265212155878544, "leak": 0.9322806359268725, "replacement": 1.315294623374939}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 2.7837717533111572, "loss_scale": 4.0, "retries": 0, "seconds": 184.4, "max_memory_gb": 18.08, "pair": 0.45256977155804634, "sign": 0.2716711675748229, "keep": 0.2405614983290434, "leak": 0.7817587219178677, "replacement": 0.9041257798671722}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 1.2727148532867432, "loss_scale": 4.0, "retries": 0, "seconds": 185.3, "max_memory_gb": 18.08, "pair": 0.7439449653029442, "sign": 0.728996092453599, "keep": 0.4950315705500543, "leak": 0.21687997862318298, "replacement": 1.4101787507534027}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 2.374234199523926, "loss_scale": 4.0, "retries": 0, "seconds": 186.3, "max_memory_gb": 18.08, "pair": 0.600078646093607, "sign": 0.6495641050860286, "keep": 0.3187252013012767, "leak": 0.5742500713095069, "replacement": 1.2277858331799507}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 1.693649411201477, "loss_scale": 4.0, "retries": 0, "seconds": 187.2, "max_memory_gb": 18.08, "pair": 0.5702704340219498, "sign": 0.46164892613887787, "keep": 0.007552159950137138, "leak": 0.2901479525899049, "replacement": 0.8774079531431198}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 1.49213707447052, "loss_scale": 4.0, "retries": 0, "seconds": 188.1, "max_memory_gb": 18.08, "pair": 0.5229066722095013, "sign": 0.5508892703801394, "keep": 0.32064833492040634, "leak": 0.5173746941727586, "replacement": 1.0880191326141357}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 3.620610475540161, "loss_scale": 4.0, "retries": 0, "seconds": 189.0, "max_memory_gb": 18.08, "pair": 0.520597331225872, "sign": 0.5835717432200909, "keep": 0.2886372748762369, "leak": 0.45436379685997963, "replacement": 1.070292741060257}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 1.9109373092651367, "loss_scale": 4.0, "retries": 0, "seconds": 190.0, "max_memory_gb": 18.08, "pair": 0.4607658330351114, "sign": 0.48159753158688545, "keep": 0.19049449788872153, "leak": 0.7365526277571917, "replacement": 0.9809500277042389}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 2.9367105960845947, "loss_scale": 4.0, "retries": 0, "seconds": 190.9, "max_memory_gb": 18.08, "pair": 0.44744032621383667, "sign": 0.5300706252455711, "keep": 0.2854798778425902, "leak": 0.9663378968834877, "replacement": 1.0968000441789627}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 1.423449993133545, "loss_scale": 4.0, "retries": 0, "seconds": 191.9, "max_memory_gb": 18.08, "pair": 0.9550713747739792, "sign": 1.0421335250139236, "keep": 1.4542058389633894, "leak": 0.12077789934284056, "replacement": 2.23343563079834}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 3.9474401473999023, "loss_scale": 4.0, "retries": 0, "seconds": 192.8, "max_memory_gb": 18.08, "pair": 0.29743194207549095, "sign": 0.1216006875038147, "keep": 0.055954559706151485, "leak": 1.1374157667160034, "replacement": 0.6705635040998459}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 1.6995882987976074, "loss_scale": 4.0, "retries": 0, "seconds": 193.8, "max_memory_gb": 18.08, "pair": 0.629410944879055, "sign": 0.6477580796927214, "keep": 0.4217044934630394, "leak": 0.5000440669828095, "replacement": 1.2891533002257347}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 1.8098100423812866, "loss_scale": 4.0, "retries": 0, "seconds": 194.7, "max_memory_gb": 18.08, "pair": 0.6101316437125206, "sign": 0.5582827590405941, "keep": 0.4504770892672241, "leak": 0.6986852190457284, "replacement": 1.2891828417778015}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 3.7219936847686768, "loss_scale": 4.0, "retries": 0, "seconds": 195.7, "max_memory_gb": 18.08, "pair": 0.5575390607118607, "sign": 0.5448109591379762, "keep": 0.31314181350171566, "leak": 0.9366889637531131, "replacement": 1.2206876277923584}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 2.0570180416107178, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 18.08, "pair": 0.6427992433309555, "sign": 0.7777926363050938, "keep": 0.5785131976008415, "leak": 0.2145657439250499, "replacement": 1.3745936304330826}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 4.836806297302246, "loss_scale": 4.0, "retries": 0, "seconds": 197.6, "max_memory_gb": 18.08, "pair": 0.487608821131289, "sign": 0.48893390223383904, "keep": 0.6022336026653647, "leak": 1.7512398634571582, "replacement": 1.4710025489330292}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 4.148850917816162, "loss_scale": 4.0, "retries": 0, "seconds": 198.5, "max_memory_gb": 18.08, "pair": 0.6972639709711075, "sign": 0.7150236293673515, "keep": 1.236627399979625, "leak": 0.1682182371860108, "replacement": 1.7151440978050232}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 1.0646371841430664, "loss_scale": 4.0, "retries": 0, "seconds": 199.5, "max_memory_gb": 18.08, "pair": 0.8350172638893127, "sign": 0.9579727500677109, "keep": 0.8204993318067864, "leak": 0.33899354931782, "replacement": 1.8090016543865204}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.4748291969299316, "loss_scale": 4.0, "retries": 0, "seconds": 200.4, "max_memory_gb": 18.08, "pair": 0.5689356625080109, "sign": 0.4036157876253128, "keep": 0.5083305109292269, "leak": 0.2517087000887841, "replacement": 1.0878360122442245}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 3.864272356033325, "loss_scale": 4.0, "retries": 0, "seconds": 201.4, "max_memory_gb": 18.08, "pair": 0.36110998690128326, "sign": 0.27997349575161934, "keep": 0.14195216400548816, "leak": 1.5449568331241608, "replacement": 0.9583120197057724}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 1.5832250118255615, "loss_scale": 4.0, "retries": 0, "seconds": 202.3, "max_memory_gb": 18.08, "pair": 0.903190478682518, "sign": 1.080828532576561, "keep": 0.3291864430066198, "leak": 0.17774661821860605, "replacement": 1.6526346653699875}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 1.6036057472229004, "loss_scale": 4.0, "retries": 0, "seconds": 203.2, "max_memory_gb": 18.08, "pair": 0.7280808687210083, "sign": 0.6265250667929649, "keep": 0.23012464586645365, "leak": 0.15542115294374526, "replacement": 1.195261001586914}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 3.4631857872009277, "loss_scale": 4.0, "retries": 0, "seconds": 204.2, "max_memory_gb": 18.08, "pair": 0.5223371312022209, "sign": 0.5452052280306816, "keep": 0.3070547712268308, "leak": 0.45387979503721, "replacement": 1.0619370639324188}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 3.1403489112854004, "loss_scale": 4.0, "retries": 0, "seconds": 205.1, "max_memory_gb": 18.08, "pair": 0.40063929185271263, "sign": 0.24780713394284248, "keep": 0.2687400709837675, "leak": 0.368415706732776, "replacement": 0.7510168254375458}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 2.2242796421051025, "loss_scale": 4.0, "retries": 0, "seconds": 206.0, "max_memory_gb": 18.08, "pair": 0.4066610783338547, "sign": 0.23022069595754147, "keep": 0.39144818764179945, "leak": 1.2094666231423616, "replacement": 1.0198622047901154}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 1.2066768407821655, "loss_scale": 4.0, "retries": 0, "seconds": 207.0, "max_memory_gb": 18.08, "pair": 0.7277858555316925, "sign": 0.718110978603363, "keep": 0.3928854032419622, "leak": 0.19300468340225052, "replacement": 1.3315352350473404}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 1.9599071741104126, "loss_scale": 4.0, "retries": 0, "seconds": 207.9, "max_memory_gb": 18.08, "pair": 0.5524804778397083, "sign": 0.5443065203726292, "keep": 0.23583389399573207, "leak": 0.6480741974664852, "replacement": 1.1045692637562752}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 2.240966320037842, "loss_scale": 4.0, "retries": 0, "seconds": 208.9, "max_memory_gb": 18.08, "pair": 0.5664461515843868, "sign": 0.5534228160977364, "keep": 0.2247538110241294, "leak": 0.7070034942695571, "replacement": 1.1322852820158005}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 4.109792709350586, "loss_scale": 4.0, "retries": 0, "seconds": 209.8, "max_memory_gb": 18.08, "pair": 0.46884194761514664, "sign": 0.4233116339892149, "keep": 0.3242097180336714, "leak": 1.0684245694428682, "replacement": 1.1097087860107422}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 2.6316580772399902, "loss_scale": 4.0, "retries": 0, "seconds": 210.7, "max_memory_gb": 18.08, "pair": 0.35441644862294197, "sign": 0.20927329827100039, "keep": 0.052956419240217656, "leak": 0.6124777272343636, "replacement": 0.6386507451534271}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 2.538942813873291, "loss_scale": 4.0, "retries": 0, "seconds": 211.6, "max_memory_gb": 18.08, "pair": 0.5532651245594025, "sign": 0.5288887768983841, "keep": 0.1395166739821434, "leak": 0.6513794437050819, "replacement": 1.0503127425909042}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 1.5370200872421265, "loss_scale": 4.0, "retries": 0, "seconds": 212.6, "max_memory_gb": 18.08, "pair": 0.6645739153027534, "sign": 0.7294804598204792, "keep": 0.521552037447691, "leak": 0.35864600725471973, "replacement": 1.3797516524791718}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 1.1270129680633545, "loss_scale": 4.0, "retries": 0, "seconds": 213.4, "max_memory_gb": 18.08, "pair": 0.7748409509658813, "sign": 0.7386412844061852, "keep": 0.6563299727276899, "leak": 0.1857188674621284, "replacement": 1.5187563002109528}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 2.4438250064849854, "loss_scale": 4.0, "retries": 0, "seconds": 214.3, "max_memory_gb": 18.08, "pair": 0.7471863478422165, "sign": 0.8708249628543854, "keep": 0.4522301392862573, "leak": 0.4469673135317862, "replacement": 1.5204557031393051}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 5.060100078582764, "loss_scale": 4.0, "retries": 0, "seconds": 215.3, "max_memory_gb": 18.08, "pair": 0.4467077739536762, "sign": 0.5073374677449465, "keep": 0.08120389236137271, "leak": 0.7980314828455448, "replacement": 0.9404863119125366}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 5.031671524047852, "loss_scale": 4.0, "retries": 0, "seconds": 216.2, "max_memory_gb": 18.08, "pair": 0.4713378772139549, "sign": 0.22697460372000933, "keep": 0.5377534851431847, "leak": 1.1477367728948593, "replacement": 1.1406361311674118}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.270239353179932, "loss_scale": 4.0, "retries": 0, "seconds": 217.1, "max_memory_gb": 18.08, "pair": 0.49575834535062313, "sign": 0.5078220292925835, "keep": 0.40058062016032636, "leak": 1.4475371800363064, "replacement": 1.3118439614772797}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 2.454220771789551, "loss_scale": 4.0, "retries": 0, "seconds": 218.1, "max_memory_gb": 18.08, "pair": 0.6336701288819313, "sign": 0.6800586879253387, "keep": 0.5344708105549216, "leak": 0.4200660018250346, "replacement": 1.3459513485431671}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 1.6635205745697021, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 18.08, "pair": 0.6494389250874519, "sign": 0.5427944511175156, "keep": 0.17008710745722055, "leak": 0.5475089730025502, "replacement": 1.1427569538354874}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 3.2311208248138428, "loss_scale": 4.0, "retries": 0, "seconds": 219.9, "max_memory_gb": 18.08, "pair": 0.5586676895618439, "sign": 0.48740733973681927, "keep": 0.27332742349244654, "leak": 0.6260952731536236, "replacement": 1.095558911561966}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 2.587306022644043, "loss_scale": 4.0, "retries": 0, "seconds": 220.8, "max_memory_gb": 18.08, "pair": 0.4060305058956146, "sign": 0.2773940050974488, "keep": 0.10473812650889158, "leak": 0.6281173583120108, "replacement": 0.7541259154677391}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 2.521531343460083, "loss_scale": 4.0, "retries": 0, "seconds": 221.7, "max_memory_gb": 18.08, "pair": 0.551158607006073, "sign": 0.6954127717763186, "keep": 0.4654457252472639, "leak": 1.0418279683217406, "replacement": 1.3920448124408722}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 2.4687535762786865, "loss_scale": 4.0, "retries": 0, "seconds": 222.7, "max_memory_gb": 18.08, "pair": 0.46497977524995804, "sign": 0.3979591503739357, "keep": 0.18584865890443325, "leak": 0.6085411012172699, "replacement": 0.909018948674202}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 1.125914216041565, "loss_scale": 4.0, "retries": 0, "seconds": 223.6, "max_memory_gb": 18.08, "pair": 0.7322099059820175, "sign": 0.7900319844484329, "keep": 0.37727742455899715, "leak": 0.32437743805348873, "replacement": 1.3969589620828629}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 0.9403464794158936, "loss_scale": 4.0, "retries": 0, "seconds": 224.5, "max_memory_gb": 18.08, "pair": 0.6237516440451145, "sign": 0.6759335640817881, "keep": 0.3469265113817528, "leak": 0.3152265790849924, "replacement": 1.2139883413910866}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 1.8882336616516113, "loss_scale": 4.0, "retries": 0, "seconds": 225.5, "max_memory_gb": 18.08, "pair": 0.33281007781624794, "sign": 0.40941048227250576, "keep": 0.26769024156965315, "leak": 1.0003917068243027, "replacement": 0.92145836353302}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 2.2756595611572266, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 18.08, "pair": 0.7161732092499733, "sign": 0.782154225744307, "keep": 0.6646691272035241, "leak": 0.17552493885159492, "replacement": 1.4834661856293678}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 3.331455707550049, "loss_scale": 4.0, "retries": 0, "seconds": 227.3, "max_memory_gb": 18.08, "pair": 0.8600178435444832, "sign": 1.0031241551041603, "keep": 0.6841911030933261, "leak": 0.45892572704178747, "replacement": 1.8184069693088531}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 3.254484176635742, "loss_scale": 4.0, "retries": 0, "seconds": 228.3, "max_memory_gb": 18.08, "pair": 0.3690341766923666, "sign": 0.24600239703431726, "keep": 0.20392260677181184, "leak": 0.9327925369143486, "replacement": 0.8271948397159576}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 1.19435715675354, "loss_scale": 4.0, "retries": 0, "seconds": 229.2, "max_memory_gb": 18.08, "pair": 0.47025810182094574, "sign": 0.4596340209245682, "keep": 0.24087262008106336, "leak": 0.5883690975606441, "replacement": 0.9676037132740021}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 2.9436912536621094, "loss_scale": 4.0, "retries": 0, "seconds": 230.1, "max_memory_gb": 18.08, "pair": 0.3617044538259506, "sign": 0.21241123043000698, "keep": 0.14817220345139503, "leak": 0.7160584451630712, "replacement": 0.7210107743740082}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 4.314134120941162, "loss_scale": 4.0, "retries": 0, "seconds": 231.1, "max_memory_gb": 18.08, "pair": 0.7223023250699043, "sign": 0.9801253080368042, "keep": 0.13657724997028708, "leak": 0.4334710710681975, "replacement": 1.3890213668346405}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 5.684929847717285, "loss_scale": 4.0, "retries": 0, "seconds": 232.0, "max_memory_gb": 18.08, "pair": 0.3291196636855602, "sign": 0.2994745224714279, "keep": 0.3363733512815088, "leak": 1.6034828759729862, "replacement": 1.0479143261909485}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 1.6236516237258911, "loss_scale": 4.0, "retries": 0, "seconds": 232.9, "max_memory_gb": 18.08, "pair": 0.49302102625370026, "sign": 0.5167978554964066, "keep": 0.27289414033293724, "leak": 0.7302781267208047, "replacement": 1.0704365372657776}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 4.644170761108398, "loss_scale": 4.0, "retries": 0, "seconds": 233.9, "max_memory_gb": 18.08, "pair": 0.30529414117336273, "sign": 0.263826054520905, "keep": 0.038142894161865115, "leak": 1.4957448951900005, "replacement": 0.8302148431539536}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 3.092916488647461, "loss_scale": 4.0, "retries": 0, "seconds": 234.8, "max_memory_gb": 18.08, "pair": 0.38970835134387016, "sign": 0.34697909094393253, "keep": 0.18166973581537604, "leak": 0.8925036510336213, "replacement": 0.8771586716175079}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 2.6073851585388184, "loss_scale": 4.0, "retries": 0, "seconds": 235.7, "max_memory_gb": 18.08, "pair": 0.41599779576063156, "sign": 0.5705531695857644, "keep": 0.5357884354889393, "leak": 0.6072062738239765, "replacement": 1.1209701895713806}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 1.2093288898468018, "loss_scale": 4.0, "retries": 0, "seconds": 236.6, "max_memory_gb": 18.08, "pair": 0.671209029853344, "sign": 0.799443207681179, "keep": 0.2829270213842392, "leak": 0.5122370719909668, "replacement": 1.3404533863067627}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 1.7935081720352173, "loss_scale": 4.0, "retries": 0, "seconds": 237.6, "max_memory_gb": 18.08, "pair": 0.5872110985219479, "sign": 0.6567334476858377, "keep": 0.24187398282811046, "leak": 0.5067037269473076, "replacement": 1.1631907224655151}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 3.5555365085601807, "loss_scale": 4.0, "retries": 0, "seconds": 238.5, "max_memory_gb": 18.08, "pair": 0.5178112126886845, "sign": 0.5365023873746395, "keep": 0.11979802919086069, "leak": 1.1636161767019075, "replacement": 1.1368654817342758}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 3.196810245513916, "loss_scale": 4.0, "retries": 0, "seconds": 239.4, "max_memory_gb": 18.08, "pair": 0.35031197033822536, "sign": 0.23406265024095774, "keep": 0.14629418030381203, "leak": 1.1191440969705582, "replacement": 0.8202764093875885}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 1.0116990804672241, "loss_scale": 4.0, "retries": 0, "seconds": 240.4, "max_memory_gb": 18.08, "pair": 0.3165798895061016, "sign": 0.2712961845099926, "keep": 0.028201931971125305, "leak": 0.9826561212539673, "replacement": 0.7119929641485214}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 2.016329288482666, "loss_scale": 4.0, "retries": 0, "seconds": 241.3, "max_memory_gb": 18.08, "pair": 0.4417066238820553, "sign": 0.43638743460178375, "keep": 0.3151319529861212, "leak": 0.5836766771972179, "replacement": 0.9633855074644089}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 2.2300288677215576, "loss_scale": 4.0, "retries": 0, "seconds": 242.2, "max_memory_gb": 18.08, "pair": 0.39898451417684555, "sign": 0.34482428804039955, "keep": 0.22029371559619904, "leak": 0.9275712276212289, "replacement": 0.9134363010525703}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 2.267507553100586, "loss_scale": 4.0, "retries": 0, "seconds": 243.1, "max_memory_gb": 18.08, "pair": 0.4607449471950531, "sign": 0.5589561797678471, "keep": 0.534155439469032, "leak": 1.1985567845404148, "replacement": 1.306939959526062}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 3.804412841796875, "loss_scale": 4.0, "retries": 0, "seconds": 244.1, "max_memory_gb": 18.08, "pair": 0.42003526724874973, "sign": 0.4135902039706707, "keep": 0.5191387739032507, "leak": 1.1185575565323234, "replacement": 1.166039153933525}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 2.407670021057129, "loss_scale": 4.0, "retries": 0, "seconds": 245.0, "max_memory_gb": 18.08, "pair": 0.33841482177376747, "sign": 0.21830083429813385, "keep": 0.7339097479707561, "leak": 0.9858702532947063, "replacement": 1.0609877333045006}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 2.5501506328582764, "loss_scale": 4.0, "retries": 0, "seconds": 245.9, "max_memory_gb": 18.08, "pair": 0.43485091999173164, "sign": 0.4297843463718891, "keep": 0.16681587451603264, "leak": 1.100502088200301, "replacement": 1.0082765519618988}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 4.45094108581543, "loss_scale": 4.0, "retries": 0, "seconds": 246.9, "max_memory_gb": 18.08, "pair": 0.4380870955064893, "sign": 0.4306012592278421, "keep": 0.921816045884043, "leak": 0.4080002760019852, "replacement": 1.216295838356018}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 2.6467576026916504, "loss_scale": 4.0, "retries": 0, "seconds": 247.8, "max_memory_gb": 18.08, "pair": 0.5037122517824173, "sign": 0.49529130570590496, "keep": 0.24791823653504252, "leak": 0.7351885082316585, "replacement": 1.0591141730546951}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.8109174966812134, "loss_scale": 4.0, "retries": 0, "seconds": 248.7, "max_memory_gb": 18.08, "pair": 0.6982181519269943, "sign": 0.8000236600637436, "keep": 0.4462362111080438, "leak": 0.4526678384436309, "replacement": 1.4345150738954544}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.214918613433838, "loss_scale": 4.0, "retries": 0, "seconds": 249.6, "max_memory_gb": 18.08, "pair": 0.4915138930082321, "sign": 0.633696036413312, "keep": 1.2469463190063834, "leak": 0.6784284854365978, "replacement": 1.6014421582221985}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 1.9551509618759155, "loss_scale": 4.0, "retries": 0, "seconds": 250.6, "max_memory_gb": 18.08, "pair": 0.35795640386641026, "sign": 0.3914042115211487, "keep": 0.27045252081006765, "leak": 0.871877608820796, "replacement": 0.9068541973829269}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 3.910907745361328, "loss_scale": 4.0, "retries": 0, "seconds": 251.5, "max_memory_gb": 18.08, "pair": 0.3655080758035183, "sign": 0.5400682203471661, "keep": 0.4559717667289078, "leak": 0.6576636675745249, "replacement": 1.0279439687728882}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 4.530470371246338, "loss_scale": 4.0, "retries": 0, "seconds": 252.4, "max_memory_gb": 18.08, "pair": 0.25110509619116783, "sign": 0.1087445579469204, "keep": 0.028026263462379575, "leak": 1.5170534923672676, "replacement": 0.6987538710236549}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 2.1067585945129395, "loss_scale": 4.0, "retries": 0, "seconds": 253.4, "max_memory_gb": 18.08, "pair": 0.3157966695725918, "sign": 0.1978483460843563, "keep": 0.02290821252972819, "leak": 0.6582064926624298, "replacement": 0.5907265767455101}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 2.5793049335479736, "loss_scale": 4.0, "retries": 0, "seconds": 254.3, "max_memory_gb": 18.08, "pair": 0.7218192294239998, "sign": 0.8083149790763855, "keep": 0.31415209802798927, "leak": 0.20634020498255268, "replacement": 1.334637776017189}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 4.770883560180664, "loss_scale": 4.0, "retries": 0, "seconds": 255.3, "max_memory_gb": 18.08, "pair": 0.4120276179164648, "sign": 0.4682813389226794, "keep": 1.1764410138130188, "leak": 1.2158982483670115, "replacement": 1.5383633822202682}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 3.204246759414673, "loss_scale": 4.0, "retries": 0, "seconds": 256.3, "max_memory_gb": 18.08, "pair": 0.2543197926133871, "sign": 0.15129761770367622, "keep": 0.02426175697473809, "leak": 0.8059767782688141, "replacement": 0.5435936748981476}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 3.2173733711242676, "loss_scale": 4.0, "retries": 0, "seconds": 257.3, "max_memory_gb": 18.08, "pair": 0.358048003166914, "sign": 0.3438010476529598, "keep": 0.012452610186301172, "leak": 1.162206381559372, "replacement": 0.8267264068126678}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 1.8497077226638794, "loss_scale": 4.0, "retries": 0, "seconds": 258.3, "max_memory_gb": 18.08, "pair": 0.4697243385016918, "sign": 0.5035604778677225, "keep": 0.27257914061192423, "leak": 0.3567701764695812, "replacement": 0.946986734867096}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 3.2361440658569336, "loss_scale": 4.0, "retries": 0, "seconds": 259.3, "max_memory_gb": 18.08, "pair": 0.5181155577301979, "sign": 0.33734405040740967, "keep": 0.2062778744730167, "leak": 0.9847471863031387, "replacement": 1.036113291978836}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 3.408675193786621, "loss_scale": 4.0, "retries": 0, "seconds": 260.3, "max_memory_gb": 18.08, "pair": 0.23552753776311874, "sign": 0.22709783632308245, "keep": 0.03437359561212361, "leak": 1.097979724407196, "replacement": 0.6407581716775894}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 2.327996253967285, "loss_scale": 4.0, "retries": 0, "seconds": 261.3, "max_memory_gb": 18.08, "pair": 0.37122809514403343, "sign": 0.31225467193871737, "keep": 0.5741243019001558, "leak": 0.4840412635821849, "replacement": 0.9354279115796089}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 3.4629290103912354, "loss_scale": 4.0, "retries": 0, "seconds": 262.3, "max_memory_gb": 18.08, "pair": 0.38642196729779243, "sign": 0.406834009103477, "keep": 0.3097042039735243, "leak": 0.8650905042886734, "replacement": 0.9609636962413788}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 5.398530006408691, "loss_scale": 4.0, "retries": 0, "seconds": 263.4, "max_memory_gb": 18.08, "pair": 0.15543363988399506, "sign": 0.10567117016762495, "keep": 0.13223234564065933, "leak": 2.293088436126709, "replacement": 0.8476575016975403}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 3.9960975646972656, "loss_scale": 4.0, "retries": 0, "seconds": 264.4, "max_memory_gb": 18.08, "pair": 0.41199231520295143, "sign": 0.38047715835273266, "keep": 0.1936760563403368, "leak": 0.5347070840070955, "replacement": 0.8327456936240196}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 4.604925632476807, "loss_scale": 4.0, "retries": 0, "seconds": 265.3, "max_memory_gb": 18.08, "pair": 0.5681411847472191, "sign": 0.8074884563684464, "keep": 0.16169631015509367, "leak": 1.0033891177736223, "replacement": 1.3035808801651}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 2.6502771377563477, "loss_scale": 4.0, "retries": 0, "seconds": 266.3, "max_memory_gb": 18.08, "pair": 0.48799077421426773, "sign": 0.9124516844749451, "keep": 0.04360132559668273, "leak": 0.8326791487634182, "replacement": 1.1741870194673538}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 3.4815304279327393, "loss_scale": 4.0, "retries": 0, "seconds": 267.2, "max_memory_gb": 18.08, "pair": 0.4775056354701519, "sign": 0.5017815716564655, "keep": 0.5374555010348558, "leak": 0.6188569441437721, "replacement": 1.1518383547663689}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 4.536765098571777, "loss_scale": 4.0, "retries": 0, "seconds": 268.2, "max_memory_gb": 18.08, "pair": 0.32984794955700636, "sign": 0.5928267645649612, "keep": 0.029442546423524618, "leak": 1.5389199182391167, "replacement": 1.0257125943899155}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 2.8950488567352295, "loss_scale": 4.0, "retries": 0, "seconds": 269.1, "max_memory_gb": 18.08, "pair": 0.37582360208034515, "sign": 0.2965117562562227, "keep": 0.2470886204391718, "leak": 0.5308938398957253, "replacement": 0.7803472280502319}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 2.707019090652466, "loss_scale": 4.0, "retries": 0, "seconds": 270.1, "max_memory_gb": 18.08, "pair": 0.5235666744410992, "sign": 0.5959261392708868, "keep": 0.5273637030040845, "leak": 0.4469909081235528, "replacement": 1.1969593316316605}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 2.8427138328552246, "loss_scale": 4.0, "retries": 0, "seconds": 271.0, "max_memory_gb": 18.08, "pair": 0.1961941458284855, "sign": 0.08985457848757505, "keep": 0.02710012352326885, "leak": 0.8539467826485634, "replacement": 0.46815820783376694}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 4.461032867431641, "loss_scale": 4.0, "retries": 0, "seconds": 272.0, "max_memory_gb": 18.08, "pair": 0.2601407654583454, "sign": 0.23298504576086998, "keep": 0.138665261445567, "leak": 1.0282446816563606, "replacement": 0.7030270770192146}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 2.3615899085998535, "loss_scale": 4.0, "retries": 0, "seconds": 272.9, "max_memory_gb": 18.08, "pair": 0.3695642128586769, "sign": 0.29519206564873457, "keep": 0.15672941866796464, "leak": 0.6323079094290733, "replacement": 0.7536019533872604}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 3.188145160675049, "loss_scale": 4.0, "retries": 0, "seconds": 273.8, "max_memory_gb": 18.08, "pair": 0.43634573742747307, "sign": 0.20214168261736631, "keep": 0.46714621782302856, "leak": 0.4381414204835892, "replacement": 0.880525067448616}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 2.972748279571533, "loss_scale": 4.0, "retries": 0, "seconds": 274.7, "max_memory_gb": 18.08, "pair": 0.5614778846502304, "sign": 0.7833452597260475, "keep": 0.7093743234872818, "leak": 0.7478318139910698, "replacement": 1.494795560836792}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 2.4324142932891846, "loss_scale": 4.0, "retries": 0, "seconds": 275.6, "max_memory_gb": 18.08, "pair": 0.5561957713216543, "sign": 0.6889715809375048, "keep": 0.3791044575627893, "leak": 0.7639049502067792, "replacement": 1.281210020184517}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 2.0904629230499268, "loss_scale": 4.0, "retries": 0, "seconds": 276.5, "max_memory_gb": 18.08, "pair": 0.541460283100605, "sign": 0.80075553804636, "keep": 0.10518423520261422, "leak": 0.6518566086888313, "replacement": 1.1573943346738815}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 4.664424896240234, "loss_scale": 4.0, "retries": 0, "seconds": 277.4, "max_memory_gb": 18.08, "pair": 0.6505732461810112, "sign": 0.8089917525649071, "keep": 0.9363469244563021, "leak": 0.22699960693717003, "replacement": 1.5799925327301025}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 5.031586170196533, "loss_scale": 4.0, "retries": 0, "seconds": 278.4, "max_memory_gb": 18.08, "pair": 0.2327330745756626, "sign": 0.11217869352549314, "keep": 0.04679398471489549, "leak": 1.3484309054911137, "replacement": 0.6493271365761757}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 2.697988748550415, "loss_scale": 4.0, "retries": 0, "seconds": 279.3, "max_memory_gb": 18.08, "pair": 0.7266874089837074, "sign": 0.8494881857186556, "keep": 0.42659541219472885, "leak": 0.6781187472315651, "replacement": 1.5342588424682617}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 2.7373592853546143, "loss_scale": 4.0, "retries": 0, "seconds": 280.2, "max_memory_gb": 18.08, "pair": 0.4225878231227398, "sign": 0.4386797994375229, "keep": 0.2384403720498085, "leak": 0.62842839024961, "replacement": 0.9182550311088562}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 4.211363315582275, "loss_scale": 4.0, "retries": 0, "seconds": 281.1, "max_memory_gb": 18.08, "pair": 0.20064521580934525, "sign": 0.07912204274907708, "keep": 0.03598137176595628, "leak": 1.2863486856222153, "replacement": 0.579784095287323}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 3.3107943534851074, "loss_scale": 4.0, "retries": 0, "seconds": 282.0, "max_memory_gb": 18.08, "pair": 0.29420114681124687, "sign": 0.21190223190933466, "keep": 0.029047950345557183, "leak": 1.7102317661046982, "replacement": 0.8422341644763947}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 3.3341522216796875, "loss_scale": 4.0, "retries": 0, "seconds": 283.0, "max_memory_gb": 18.08, "pair": 0.3261876255273819, "sign": 0.5239511355757713, "keep": 0.06926248781383038, "leak": 0.9168844595551491, "replacement": 0.852015532553196}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 3.9732720851898193, "loss_scale": 4.0, "retries": 0, "seconds": 283.9, "max_memory_gb": 18.08, "pair": 0.26774775981903076, "sign": 0.1626054896041751, "keep": 0.10951435990864411, "leak": 1.3559404723346233, "replacement": 0.7427928000688553}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 3.0564613342285156, "loss_scale": 4.0, "retries": 0, "seconds": 284.9, "max_memory_gb": 18.08, "pair": 0.25025613233447075, "sign": 0.06211879011243582, "keep": 0.09949963074177504, "leak": 0.7654861398041248, "replacement": 0.5224368870258331}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 3.708889961242676, "loss_scale": 4.0, "retries": 0, "seconds": 285.8, "max_memory_gb": 18.08, "pair": 0.41519196331501007, "sign": 0.35971488151699305, "keep": 0.12119421316310763, "leak": 0.866641201544553, "replacement": 0.8723067790269852}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 4.135947227478027, "loss_scale": 4.0, "retries": 0, "seconds": 286.7, "max_memory_gb": 18.08, "pair": 0.40781282261013985, "sign": 0.615517932921648, "keep": 0.07986630007508211, "leak": 0.9168266840279102, "replacement": 0.984711617231369}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 2.398874521255493, "loss_scale": 4.0, "retries": 0, "seconds": 287.6, "max_memory_gb": 18.08, "pair": 0.6204469427466393, "sign": 0.6612554469611496, "keep": 0.368687326554209, "leak": 0.276883730373811, "replacement": 1.2046392410993576}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 2.0528366565704346, "loss_scale": 4.0, "retries": 0, "seconds": 288.6, "max_memory_gb": 18.08, "pair": 0.322994701564312, "sign": 0.15295394975692034, "keep": 0.02796029584715143, "leak": 0.6079298555850983, "replacement": 0.5654342621564865}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 1.3402352333068848, "loss_scale": 4.0, "retries": 0, "seconds": 289.5, "max_memory_gb": 18.08, "pair": 0.43850649148225784, "sign": 0.47807520627975464, "keep": 0.2113942367141135, "leak": 0.7643611375242472, "replacement": 0.9743314683437347}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.219494342803955, "loss_scale": 4.0, "retries": 0, "seconds": 290.4, "max_memory_gb": 18.08, "pair": 0.38950444385409355, "sign": 0.36211065109819174, "keep": 0.06471527746180072, "leak": 0.6543587120977463, "replacement": 0.7665070742368698}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 7.469383716583252, "loss_scale": 4.0, "retries": 0, "seconds": 291.4, "max_memory_gb": 18.08, "pair": 0.42492748983204365, "sign": 0.5806236630305648, "keep": 0.4385485094971955, "leak": 1.8018290773034096, "replacement": 1.3849709033966064}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 3.2653820514678955, "loss_scale": 4.0, "retries": 0, "seconds": 292.3, "max_memory_gb": 18.08, "pair": 0.465384591370821, "sign": 0.5138253099285066, "keep": 0.2707230739761144, "leak": 0.777657636499498, "replacement": 1.052073210477829}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 5.10919189453125, "loss_scale": 4.0, "retries": 0, "seconds": 293.3, "max_memory_gb": 18.08, "pair": 0.27607310470193624, "sign": 0.2512968434020877, "keep": 0.08071509562432766, "leak": 1.795509662013501, "replacement": 0.8909565135836601}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 3.494797468185425, "loss_scale": 4.0, "retries": 0, "seconds": 294.2, "max_memory_gb": 18.08, "pair": 0.6167144514620304, "sign": 0.7248102556914091, "keep": 0.8178812563419342, "leak": 0.24986435333266854, "replacement": 1.4505263268947601}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 4.437374114990234, "loss_scale": 4.0, "retries": 0, "seconds": 295.1, "max_memory_gb": 18.08, "pair": 0.42988932877779007, "sign": 0.45783738791942596, "keep": 0.36639382503926754, "leak": 1.274181897751987, "replacement": 1.1605504006147385}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 3.5301389694213867, "loss_scale": 4.0, "retries": 0, "seconds": 296.0, "max_memory_gb": 18.08, "pair": 0.36547211930155754, "sign": 0.295413002371788, "keep": 0.0723845586180687, "leak": 0.8042245991528034, "replacement": 0.7504270523786545}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 2.93112850189209, "loss_scale": 4.0, "retries": 0, "seconds": 296.9, "max_memory_gb": 18.08, "pair": 0.46337785571813583, "sign": 0.3812631107866764, "keep": 0.20904521737247705, "leak": 0.48614220321178436, "replacement": 0.8800675868988037}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 1.985084056854248, "loss_scale": 4.0, "retries": 0, "seconds": 297.8, "max_memory_gb": 18.08, "pair": 0.4382968507707119, "sign": 0.4067229554057121, "keep": 0.2640319325728342, "leak": 0.6752847488969564, "replacement": 0.9424954801797867}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 2.618001699447632, "loss_scale": 4.0, "retries": 0, "seconds": 298.7, "max_memory_gb": 18.08, "pair": 0.3436155840754509, "sign": 0.30032240878790617, "keep": 0.1505437433370389, "leak": 0.7883813064545393, "replacement": 0.7661439627408981}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 4.457301616668701, "loss_scale": 4.0, "retries": 0, "seconds": 299.6, "max_memory_gb": 18.08, "pair": 0.40661855414509773, "sign": 0.4832497350871563, "keep": 0.03780599357560277, "leak": 0.5749481916427612, "replacement": 0.810883417725563}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 4.4931488037109375, "loss_scale": 4.0, "retries": 0, "seconds": 300.6, "max_memory_gb": 18.08, "pair": 0.3373619578778744, "sign": 0.2615233901888132, "keep": 0.09682853799313307, "leak": 1.69082909822464, "replacement": 0.9392451643943787}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 2.404031991958618, "loss_scale": 4.0, "retries": 0, "seconds": 301.5, "max_memory_gb": 18.08, "pair": 0.19723619148135185, "sign": 0.09469729755073786, "keep": 0.06420693977270275, "leak": 0.7583903223276138, "replacement": 0.46628589928150177}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 2.8135459423065186, "loss_scale": 4.0, "retries": 0, "seconds": 302.4, "max_memory_gb": 18.08, "pair": 0.5774549692869186, "sign": 0.6197347103152424, "keep": 0.5437538588885218, "leak": 0.4209550683153793, "replacement": 1.2644379958510399}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 2.9876694679260254, "loss_scale": 4.0, "retries": 0, "seconds": 303.4, "max_memory_gb": 18.08, "pair": 0.44210605323314667, "sign": 0.5812158361077309, "keep": 0.1234437447565142, "leak": 0.4310388630256057, "replacement": 0.9021955728530884}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 4.299084663391113, "loss_scale": 4.0, "retries": 0, "seconds": 304.3, "max_memory_gb": 18.08, "pair": 0.3058435805141926, "sign": 0.13905784441158175, "keep": 0.41204932564869523, "leak": 0.8083839461833122, "replacement": 0.7834931463003159}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 3.6081981658935547, "loss_scale": 4.0, "retries": 0, "seconds": 305.2, "max_memory_gb": 18.08, "pair": 0.5311901196837425, "sign": 0.6288807839155197, "keep": 0.1736100635607727, "leak": 0.6982066070486326, "replacement": 1.1069871932268143}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 2.321484088897705, "loss_scale": 4.0, "retries": 0, "seconds": 306.2, "max_memory_gb": 18.08, "pair": 0.4210871644318104, "sign": 0.4160562288016081, "keep": 0.19238817936275154, "leak": 0.5386204505339265, "replacement": 0.859964482486248}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 4.681082725524902, "loss_scale": 4.0, "retries": 0, "seconds": 307.1, "max_memory_gb": 18.08, "pair": 0.47662883158773184, "sign": 0.5550434035249054, "keep": 0.295295638818061, "leak": 0.8782029557041824, "replacement": 1.1213491261005402}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 1.9702725410461426, "loss_scale": 4.0, "retries": 0, "seconds": 308.0, "max_memory_gb": 18.08, "pair": 0.6999601125717163, "sign": 0.6633706241846085, "keep": 0.5009776294464245, "leak": 0.3215188429458067, "replacement": 1.3625139594078064}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 3.6380348205566406, "loss_scale": 4.0, "retries": 0, "seconds": 309.0, "max_memory_gb": 18.08, "pair": 0.3697920795530081, "sign": 0.39088828675448895, "keep": 0.27845475065987557, "leak": 0.8324530348181725, "replacement": 0.9125768542289734}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 2.3234376907348633, "loss_scale": 4.0, "retries": 0, "seconds": 309.9, "max_memory_gb": 18.08, "pair": 0.6753835529088974, "sign": 0.801789715886116, "keep": 0.48783727386035025, "leak": 0.46202036256727297, "replacement": 1.4357021898031235}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 3.424238920211792, "loss_scale": 4.0, "retries": 0, "seconds": 310.8, "max_memory_gb": 18.08, "pair": 0.5697039887309074, "sign": 0.5486472174525261, "keep": 0.7512458562850952, "leak": 0.8237226232886314, "replacement": 1.425581157207489}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 5.4393310546875, "loss_scale": 4.0, "retries": 0, "seconds": 311.8, "max_memory_gb": 18.08, "pair": 0.591011181473732, "sign": 0.6175779029726982, "keep": 0.9606270340736955, "leak": 0.284520355402492, "replacement": 1.4512437731027603}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 2.881770133972168, "loss_scale": 4.0, "retries": 0, "seconds": 312.7, "max_memory_gb": 18.08, "pair": 0.3259099815040827, "sign": 0.18637068197131157, "keep": 0.07031314284540713, "leak": 1.0573639534413815, "replacement": 0.7185928672552109}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 3.1012847423553467, "loss_scale": 4.0, "retries": 0, "seconds": 313.7, "max_memory_gb": 18.08, "pair": 0.6101152449846268, "sign": 1.1332536824047565, "keep": 0.4996865945868194, "leak": 0.3651532471412793, "replacement": 1.51787368953228}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 1.8029812574386597, "loss_scale": 4.0, "retries": 0, "seconds": 314.6, "max_memory_gb": 18.08, "pair": 0.5093433484435081, "sign": 0.5389196183532476, "keep": 0.30802331445738673, "leak": 0.4446026485529728, "replacement": 1.0439654886722565}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 2.5872864723205566, "loss_scale": 4.0, "retries": 0, "seconds": 315.6, "max_memory_gb": 18.08, "pair": 0.5301861949265003, "sign": 0.7732694372534752, "keep": 0.2607246986590326, "leak": 0.8063636347651482, "replacement": 1.248774215579033}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 1.7173644304275513, "loss_scale": 4.0, "retries": 0, "seconds": 316.5, "max_memory_gb": 18.08, "pair": 0.3497206103056669, "sign": 0.35532872937619686, "keep": 0.5239936138968915, "leak": 0.4048640578985214, "replacement": 0.890597764402628}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 3.4838786125183105, "loss_scale": 4.0, "retries": 0, "seconds": 317.4, "max_memory_gb": 18.08, "pair": 0.3633274845778942, "sign": 0.3870646534487605, "keep": 0.4392149578779936, "leak": 0.775965865701437, "replacement": 0.9704587608575821}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 3.039809226989746, "loss_scale": 4.0, "retries": 0, "seconds": 318.4, "max_memory_gb": 18.08, "pair": 0.5627719648182392, "sign": 0.6410414595156908, "keep": 0.4583709330763668, "leak": 0.7206724407151341, "replacement": 1.2926462292671204}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 2.981631278991699, "loss_scale": 4.0, "retries": 0, "seconds": 319.3, "max_memory_gb": 18.08, "pair": 0.46700921282172203, "sign": 0.5605859979987144, "keep": 0.028418836533091962, "leak": 0.41152026876807213, "replacement": 0.8643916994333267}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 3.689260244369507, "loss_scale": 4.0, "retries": 0, "seconds": 320.2, "max_memory_gb": 18.08, "pair": 0.24117715656757355, "sign": 0.1598772369325161, "keep": 0.023043703462462872, "leak": 1.0886202678084373, "replacement": 0.6047926843166351}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 4.193528652191162, "loss_scale": 4.0, "retries": 0, "seconds": 321.1, "max_memory_gb": 18.08, "pair": 0.37962386198341846, "sign": 0.40375247970223427, "keep": 0.24319263314828277, "leak": 1.0885224221274257, "replacement": 0.9752270206809044}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 2.935180187225342, "loss_scale": 4.0, "retries": 0, "seconds": 322.1, "max_memory_gb": 18.08, "pair": 0.44457017071545124, "sign": 0.6642969027161598, "keep": 0.6739502958953381, "leak": 0.6503052860498428, "replacement": 1.2762700915336609}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 3.4185497760772705, "loss_scale": 4.0, "retries": 0, "seconds": 323.0, "max_memory_gb": 18.08, "pair": 0.44142265245318413, "sign": 0.4063863744959235, "keep": 0.3226112271659076, "leak": 0.8086983959947247, "replacement": 1.0080960541963577}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 3.0398783683776855, "loss_scale": 4.0, "retries": 0, "seconds": 323.9, "max_memory_gb": 18.08, "pair": 0.23254527896642685, "sign": 0.1933598779141903, "keep": 0.02883982309140265, "leak": 0.9894321784377098, "replacement": 0.5910031795501709}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 3.5973875522613525, "loss_scale": 4.0, "retries": 0, "seconds": 324.9, "max_memory_gb": 18.08, "pair": 0.40698670595884323, "sign": 0.41644382290542126, "keep": 0.1804107860662043, "leak": 1.1191493072838057, "replacement": 0.9852012991905212}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 2.201347827911377, "loss_scale": 4.0, "retries": 0, "seconds": 325.8, "max_memory_gb": 18.08, "pair": 0.49170146882534027, "sign": 0.7147891893982887, "keep": 0.2296921070665121, "leak": 0.1989890438853763, "replacement": 1.0136893689632416}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 2.244734287261963, "loss_scale": 4.0, "retries": 0, "seconds": 326.7, "max_memory_gb": 18.08, "pair": 0.2804508348926902, "sign": 0.17215073760598898, "keep": 0.12321388232521713, "leak": 0.18458622880280018, "replacement": 0.4742796942591667}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 2.287039041519165, "loss_scale": 4.0, "retries": 0, "seconds": 327.6, "max_memory_gb": 18.08, "pair": 0.494182039052248, "sign": 0.6517810802906752, "keep": 0.6351176453754306, "leak": 0.4385372903198004, "replacement": 1.2472657598555088}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 4.690596103668213, "loss_scale": 4.0, "retries": 0, "seconds": 328.5, "max_memory_gb": 18.08, "pair": 0.39221521466970444, "sign": 0.5387430526316166, "keep": 0.6016860008239746, "leak": 0.8639165535569191, "replacement": 1.1784089356660843}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 2.663703680038452, "loss_scale": 4.0, "retries": 0, "seconds": 329.5, "max_memory_gb": 18.08, "pair": 0.5045347847044468, "sign": 0.4929627776145935, "keep": 0.2980683230853174, "leak": 0.303488096304136, "replacement": 0.9759223461151123}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 8.358214378356934, "loss_scale": 4.0, "retries": 0, "seconds": 330.4, "max_memory_gb": 18.08, "pair": 0.4754188945516944, "sign": 0.5711669584270567, "keep": 0.29297115735244006, "leak": 3.102265982190147, "replacement": 1.6830544471740723}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 2.290658473968506, "loss_scale": 4.0, "retries": 0, "seconds": 331.4, "max_memory_gb": 18.08, "pair": 0.1991661749780178, "sign": 0.19618035852909088, "keep": 0.05450035072863102, "leak": 1.0580469444394112, "replacement": 0.5890182554721832}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 9.92783260345459, "loss_scale": 4.0, "retries": 0, "seconds": 332.3, "max_memory_gb": 18.08, "pair": 0.20173227228224277, "sign": 0.161435003625229, "keep": 0.1585627612657845, "leak": 3.3366920053958893, "replacement": 1.195904165506363}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 2.971555233001709, "loss_scale": 4.0, "retries": 0, "seconds": 333.3, "max_memory_gb": 18.08, "pair": 0.5914207436144352, "sign": 0.6904966402798891, "keep": 0.4604556446429342, "leak": 0.24536258890293539, "replacement": 1.2282375171780586}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 1.9959280490875244, "loss_scale": 4.0, "retries": 0, "seconds": 334.2, "max_memory_gb": 18.08, "pair": 0.39658353105187416, "sign": 0.6036375127732754, "keep": 0.06825841590762138, "leak": 0.7013894913598051, "replacement": 0.9078788757324219}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 3.7235817909240723, "loss_scale": 4.0, "retries": 0, "seconds": 335.2, "max_memory_gb": 18.08, "pair": 0.29320305958390236, "sign": 0.2341163158416748, "keep": 0.10767732525710016, "leak": 1.0913471095263958, "replacement": 0.7369366735219955}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 3.5347275733947754, "loss_scale": 4.0, "retries": 0, "seconds": 336.1, "max_memory_gb": 18.08, "pair": 0.3356140637770295, "sign": 0.5198297314345837, "keep": 0.03569350120960735, "leak": 1.2110265884548426, "replacement": 0.9161323234438896}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 6.167210578918457, "loss_scale": 4.0, "retries": 0, "seconds": 337.0, "max_memory_gb": 18.08, "pair": 0.489579401910305, "sign": 0.4870933061465621, "keep": 1.799082414014265, "leak": 0.34249793738126755, "replacement": 1.7182917669415474}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 7.249122142791748, "loss_scale": 4.0, "retries": 0, "seconds": 337.9, "max_memory_gb": 18.08, "pair": 0.5810094922780991, "sign": 1.2557679116725922, "keep": 0.05219879187643528, "leak": 1.0078256018459797, "replacement": 1.4869492799043655}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 5.51386833190918, "loss_scale": 4.0, "retries": 0, "seconds": 338.9, "max_memory_gb": 18.08, "pair": 0.16182421147823334, "sign": 0.07596762734465301, "keep": 0.03973666159436107, "leak": 2.00555756688118, "replacement": 0.7210657447576523}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 2.2704148292541504, "loss_scale": 4.0, "retries": 0, "seconds": 339.8, "max_memory_gb": 18.08, "pair": 0.3784872591495514, "sign": 0.40237935446202755, "keep": 0.2902899738983251, "leak": 0.300841910764575, "replacement": 0.800032414495945}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 9.683090209960938, "loss_scale": 4.0, "retries": 0, "seconds": 340.7, "max_memory_gb": 18.08, "pair": 0.3109831288456917, "sign": 0.3625490013509989, "keep": 0.05091381585225463, "leak": 2.551402486860752, "replacement": 1.1555651426315308}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 2.119439125061035, "loss_scale": 4.0, "retries": 0, "seconds": 341.7, "max_memory_gb": 18.08, "pair": 0.3349495083093643, "sign": 0.39013119600713253, "keep": 0.2869785928633064, "leak": 0.7367834168253466, "replacement": 0.8577002584934235}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 2.4243710041046143, "loss_scale": 4.0, "retries": 0, "seconds": 342.6, "max_memory_gb": 18.08, "pair": 0.21880227513611317, "sign": 0.13758974242955446, "keep": 0.06569248781306669, "leak": 0.539168001152575, "replacement": 0.4552353620529175}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 4.38275671005249, "loss_scale": 4.0, "retries": 0, "seconds": 343.5, "max_memory_gb": 18.08, "pair": 0.3695645462721586, "sign": 0.38457243517041206, "keep": 0.0631881644949317, "leak": 1.441079180687666, "replacement": 0.9537146389484406}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 4.711656093597412, "loss_scale": 4.0, "retries": 0, "seconds": 344.4, "max_memory_gb": 18.08, "pair": 0.2608973868191242, "sign": 0.30191568471491337, "keep": 1.0299185067415237, "leak": 1.1701425611972809, "replacement": 1.2193501144647598}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 3.666351795196533, "loss_scale": 4.0, "retries": 0, "seconds": 345.4, "max_memory_gb": 18.08, "pair": 0.17661073431372643, "sign": 0.08167917001992464, "keep": 0.1742120881099254, "leak": 1.6854468882083893, "replacement": 0.7259181141853333}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 5.43403959274292, "loss_scale": 4.0, "retries": 0, "seconds": 346.3, "max_memory_gb": 18.08, "pair": 0.2688294621184468, "sign": 0.26127954106777906, "keep": 0.09399521618615836, "leak": 1.001197224482894, "replacement": 0.6967661529779434}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 1.865486979484558, "loss_scale": 4.0, "retries": 0, "seconds": 347.2, "max_memory_gb": 18.08, "pair": 0.6255212910473347, "sign": 0.8336171545088291, "keep": 0.0997084395494312, "leak": 0.5114860867615789, "replacement": 1.220055639743805}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 2.4959716796875, "loss_scale": 4.0, "retries": 0, "seconds": 348.1, "max_memory_gb": 18.08, "pair": 0.31852939911186695, "sign": 0.306403873488307, "keep": 0.4533647458301857, "leak": 0.8039641883224249, "replacement": 0.8994047567248344}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 4.585592269897461, "loss_scale": 4.0, "retries": 0, "seconds": 349.1, "max_memory_gb": 18.08, "pair": 0.4185749813914299, "sign": 0.5068778656423092, "keep": 0.2974753245944157, "leak": 1.434949914459139, "replacement": 1.1794890612363815}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.020156145095825, "loss_scale": 4.0, "retries": 0, "seconds": 350.0, "max_memory_gb": 18.08, "pair": 0.3036869764328003, "sign": 0.26228552404791117, "keep": 0.22206653153989464, "leak": 0.4628812428563833, "replacement": 0.6615832969546318}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 1.6644548177719116, "loss_scale": 4.0, "retries": 0, "seconds": 350.9, "max_memory_gb": 18.08, "pair": 0.3401396684348583, "sign": 0.38043025881052017, "keep": 0.0754158126655966, "leak": 0.5215795170515776, "replacement": 0.6984575763344765}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 3.6770553588867188, "loss_scale": 4.0, "retries": 0, "seconds": 351.9, "max_memory_gb": 18.08, "pair": 0.30327521078288555, "sign": 0.22848647646605968, "keep": 0.12374059902504086, "leak": 0.7536538876593113, "replacement": 0.6678022146224976}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 3.449312925338745, "loss_scale": 4.0, "retries": 0, "seconds": 352.8, "max_memory_gb": 18.08, "pair": 0.26630761101841927, "sign": 0.3066223356872797, "keep": 0.100110596511513, "leak": 0.9043955057859421, "replacement": 0.6957729607820511}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 4.223988056182861, "loss_scale": 4.0, "retries": 0, "seconds": 353.7, "max_memory_gb": 18.08, "pair": 0.23680645134299994, "sign": 0.5169567307457328, "keep": 0.3035530224442482, "leak": 1.5374472886323929, "replacement": 1.0314231365919113}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 2.323984146118164, "loss_scale": 4.0, "retries": 0, "seconds": 354.6, "max_memory_gb": 18.08, "pair": 0.7169655859470367, "sign": 0.8757376745343208, "keep": 0.41951595433056355, "leak": 0.08906510507222265, "replacement": 1.3868586868047714}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 3.9820139408111572, "loss_scale": 4.0, "retries": 0, "seconds": 355.6, "max_memory_gb": 18.08, "pair": 0.5054416880011559, "sign": 0.7050309330224991, "keep": 0.8393998825922608, "leak": 0.8070384934544563, "replacement": 1.4794168025255203}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 2.2431118488311768, "loss_scale": 4.0, "retries": 0, "seconds": 356.5, "max_memory_gb": 18.08, "pair": 0.4107961244881153, "sign": 0.3684382252395153, "keep": 0.18339453544467688, "leak": 0.23381188698112965, "replacement": 0.7451654821634293}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 4.888091564178467, "loss_scale": 4.0, "retries": 0, "seconds": 357.4, "max_memory_gb": 18.08, "pair": 0.1346188634634018, "sign": 0.0965824294835329, "keep": 0.018189781345427036, "leak": 1.1910105794668198, "replacement": 0.48975761234760284}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 2.7677762508392334, "loss_scale": 4.0, "retries": 0, "seconds": 358.3, "max_memory_gb": 18.08, "pair": 0.44974836334586143, "sign": 0.5078473798930645, "keep": 0.33405689714709297, "leak": 0.4189668266335502, "replacement": 0.975442185997963}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 3.079542636871338, "loss_scale": 4.0, "retries": 0, "seconds": 359.3, "max_memory_gb": 18.08, "pair": 0.29341600835323334, "sign": 0.3120562667027116, "keep": 0.10255949583370239, "leak": 0.7767181098461151, "replacement": 0.6949034482240677}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 2.9182329177856445, "loss_scale": 4.0, "retries": 0, "seconds": 360.2, "max_memory_gb": 18.08, "pair": 0.3865323029458523, "sign": 0.48430959694087505, "keep": 0.17866669595241547, "leak": 0.5676197465509176, "replacement": 0.8599253594875336}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 3.151327610015869, "loss_scale": 4.0, "retries": 0, "seconds": 361.2, "max_memory_gb": 18.08, "pair": 0.21322042867541313, "sign": 0.17251134850084782, "keep": 0.04833338619209826, "leak": 0.8540325462818146, "replacement": 0.537150926887989}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 4.995655059814453, "loss_scale": 4.0, "retries": 0, "seconds": 362.1, "max_memory_gb": 18.08, "pair": 0.26270682737231255, "sign": 0.20373822981491685, "keep": 0.07839416805654764, "leak": 1.5406223759055138, "replacement": 0.7889286130666733}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 2.7435901165008545, "loss_scale": 4.0, "retries": 0, "seconds": 363.1, "max_memory_gb": 18.08, "pair": 0.39695082046091557, "sign": 0.34046535566449165, "keep": 0.10884836362674832, "leak": 0.6437334045767784, "replacement": 0.782541036605835}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 2.2862792015075684, "loss_scale": 4.0, "retries": 0, "seconds": 364.0, "max_memory_gb": 18.08, "pair": 0.5837595760822296, "sign": 0.7437341436743736, "keep": 0.7138517051935196, "leak": 0.15001436963211745, "replacement": 1.3500561267137527}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 1.5723469257354736, "loss_scale": 4.0, "retries": 0, "seconds": 365.0, "max_memory_gb": 18.08, "pair": 0.15243016183376312, "sign": 0.09803896583616734, "keep": 0.10472687866422348, "leak": 0.3636314831674099, "replacement": 0.3447209596633911}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 2.393369197845459, "loss_scale": 4.0, "retries": 0, "seconds": 365.9, "max_memory_gb": 18.08, "pair": 0.23569574765861034, "sign": 0.3031965345144272, "keep": 0.11084553180262446, "leak": 0.36373284093861, "replacement": 0.5336499959230423}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 2.7058610916137695, "loss_scale": 4.0, "retries": 0, "seconds": 366.9, "max_memory_gb": 18.08, "pair": 0.41069090366363525, "sign": 0.612968816421926, "keep": 0.1154449398163706, "leak": 0.5922539168968797, "replacement": 0.9229612499475479}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 3.9629058837890625, "loss_scale": 4.0, "retries": 0, "seconds": 367.8, "max_memory_gb": 18.08, "pair": 0.18546409625560045, "sign": 0.22207533288747072, "keep": 0.02797868865309283, "leak": 1.588120736181736, "replacement": 0.7075213193893433}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 3.379042148590088, "loss_scale": 4.0, "retries": 0, "seconds": 368.7, "max_memory_gb": 18.08, "pair": 0.1568252593278885, "sign": 0.059719452634453773, "keep": 0.049520754255354404, "leak": 0.5782049894332886, "replacement": 0.35599660128355026}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 2.3938746452331543, "loss_scale": 4.0, "retries": 0, "seconds": 369.6, "max_memory_gb": 18.08, "pair": 0.3498564790934324, "sign": 0.4908067537471652, "keep": 0.2663225932046771, "leak": 0.6609049513936043, "replacement": 0.8936473578214645}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 2.7911570072174072, "loss_scale": 4.0, "retries": 0, "seconds": 370.6, "max_memory_gb": 18.08, "pair": 0.445510221645236, "sign": 0.41660570725798607, "keep": 0.18323566461913288, "leak": 0.7120270585874096, "replacement": 0.9234376549720764}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 6.384407043457031, "loss_scale": 4.0, "retries": 0, "seconds": 371.5, "max_memory_gb": 18.08, "pair": 0.438447542488575, "sign": 0.6668839515186846, "keep": 0.7458092018496245, "leak": 0.9659031257033348, "replacement": 1.3862698823213577}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 3.8713948726654053, "loss_scale": 4.0, "retries": 0, "seconds": 372.5, "max_memory_gb": 18.08, "pair": 0.3980771005153656, "sign": 0.349989814683795, "keep": 0.3189445436000824, "leak": 0.3253236673772335, "replacement": 0.813875213265419}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 8.262919425964355, "loss_scale": 4.0, "retries": 0, "seconds": 373.5, "max_memory_gb": 18.08, "pair": 0.3867694213986397, "sign": 0.7451956421136856, "keep": 0.7851813234155998, "leak": 1.3955900222063065, "replacement": 1.500855453312397}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 2.2839667797088623, "loss_scale": 4.0, "retries": 0, "seconds": 374.4, "max_memory_gb": 18.08, "pair": 0.6765171512961388, "sign": 0.8045151140540838, "keep": 0.39010098576545715, "leak": 0.16105705588961428, "replacement": 1.3140894621610641}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 4.405023097991943, "loss_scale": 4.0, "retries": 0, "seconds": 375.3, "max_memory_gb": 18.08, "pair": 0.20312317088246346, "sign": 0.14395536202937365, "keep": 0.028977167909033597, "leak": 1.1902240365743637, "replacement": 0.5871454402804375}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 2.8355493545532227, "loss_scale": 4.0, "retries": 0, "seconds": 376.3, "max_memory_gb": 18.08, "pair": 0.31229871697723866, "sign": 0.42617512540891767, "keep": 0.2905389405786991, "leak": 0.4310592822730541, "replacement": 0.7784205675125122}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 1.853843331336975, "loss_scale": 4.0, "retries": 0, "seconds": 377.2, "max_memory_gb": 18.08, "pair": 0.25173822045326233, "sign": 0.20954060275107622, "keep": 0.06451780151110142, "leak": 0.6922246403992176, "replacement": 0.5618235692381859}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 2.6554362773895264, "loss_scale": 4.0, "retries": 0, "seconds": 378.2, "max_memory_gb": 18.08, "pair": 0.6342670097947121, "sign": 0.804159564897418, "keep": 0.27138133835978806, "leak": 0.7438944175373763, "replacement": 1.35801113396883}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 17.775522232055664, "loss_scale": 4.0, "retries": 0, "seconds": 379.1, "max_memory_gb": 18.08, "pair": 0.9352272907271981, "sign": 1.2252743169665337, "keep": 2.6152142074424773, "leak": 1.5837093219161034, "replacement": 3.2513988614082336}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 3.410555362701416, "loss_scale": 4.0, "retries": 0, "seconds": 380.1, "max_memory_gb": 18.08, "pair": 0.48920173570513725, "sign": 0.5722925867885351, "keep": 0.5896857972256839, "leak": 0.6937266639433801, "replacement": 1.2436226308345795}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 2.6819381713867188, "loss_scale": 4.0, "retries": 0, "seconds": 381.0, "max_memory_gb": 18.08, "pair": 0.43845561891794205, "sign": 0.4705764874815941, "keep": 0.3500070547743235, "leak": 0.6375762549578212, "replacement": 1.008141428232193}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 2.654045820236206, "loss_scale": 4.0, "retries": 0, "seconds": 381.9, "max_memory_gb": 18.08, "pair": 0.17969130165874958, "sign": 0.15659538470208645, "keep": 0.2772814199561253, "leak": 0.7421214617788792, "replacement": 0.5821600705385208}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 4.574853420257568, "loss_scale": 4.0, "retries": 0, "seconds": 382.9, "max_memory_gb": 18.08, "pair": 0.49751440435647964, "sign": 0.8842830136418343, "keep": 1.0322322327119764, "leak": 0.533053613267839, "replacement": 1.5890354067087173}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 4.127829551696777, "loss_scale": 4.0, "retries": 0, "seconds": 383.8, "max_memory_gb": 18.08, "pair": 0.5016747713088989, "sign": 0.7209944650530815, "keep": 0.26874354397295974, "leak": 0.6205587327713147, "replacement": 1.1516834646463394}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 4.822322845458984, "loss_scale": 4.0, "retries": 0, "seconds": 384.8, "max_memory_gb": 18.08, "pair": 0.23194021359086037, "sign": 0.17738472297787666, "keep": 0.05963022634387016, "leak": 1.3457027881231625, "replacement": 0.6868733838200569}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 4.319307804107666, "loss_scale": 4.0, "retries": 0, "seconds": 385.7, "max_memory_gb": 18.08, "pair": 0.37966490630060434, "sign": 0.3781444518826902, "keep": 0.7065673787146807, "leak": 0.7616748437285423, "replacement": 1.1124394834041595}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 4.751025676727295, "loss_scale": 4.0, "retries": 0, "seconds": 386.6, "max_memory_gb": 18.08, "pair": 0.4503657780587673, "sign": 0.3995972126722336, "keep": 1.2581526912981644, "leak": 0.8775522857904434, "replacement": 1.4986288100481033}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 3.658057451248169, "loss_scale": 4.0, "retries": 0, "seconds": 387.5, "max_memory_gb": 18.08, "pair": 0.2194742187857628, "sign": 0.18585634278133512, "keep": 0.0765651359106414, "leak": 0.5865568765439093, "replacement": 0.4973241910338402}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 4.685391426086426, "loss_scale": 4.0, "retries": 0, "seconds": 388.4, "max_memory_gb": 18.08, "pair": 0.7775724083185196, "sign": 1.1106607466936111, "keep": 1.2241067457944155, "leak": 0.5616924315690994, "replacement": 2.08537919819355}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 2.281252861022949, "loss_scale": 4.0, "retries": 0, "seconds": 389.3, "max_memory_gb": 18.08, "pair": 0.11376507580280304, "sign": 0.15604650881141424, "keep": 0.018682404537685215, "leak": 0.9335903003811836, "replacement": 0.43452711403369904}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 4.944572448730469, "loss_scale": 4.0, "retries": 0, "seconds": 390.2, "max_memory_gb": 18.08, "pair": 0.3189143165946007, "sign": 0.48685264587402344, "keep": 0.07077006809413433, "leak": 1.8578350320458412, "replacement": 1.0621844232082367}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 4.53099250793457, "loss_scale": 4.0, "retries": 0, "seconds": 391.2, "max_memory_gb": 18.08, "pair": 0.24512936361134052, "sign": 0.3942023329436779, "keep": 0.7658665196504444, "leak": 1.592036746442318, "replacement": 1.2231729552149773}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 6.538027286529541, "loss_scale": 4.0, "retries": 0, "seconds": 392.1, "max_memory_gb": 18.08, "pair": 0.1598698738962412, "sign": 0.2148965634405613, "keep": 0.12525124475359917, "leak": 2.519506558775902, "replacement": 0.9598203897476196}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 5.189336776733398, "loss_scale": 4.0, "retries": 0, "seconds": 393.0, "max_memory_gb": 18.08, "pair": 0.4772565886378288, "sign": 0.4462892934679985, "keep": 0.39906080160290003, "leak": 0.9085492207668722, "replacement": 1.1270689219236374}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 3.241523265838623, "loss_scale": 4.0, "retries": 0, "seconds": 393.9, "max_memory_gb": 18.08, "pair": 0.659219142049551, "sign": 0.9158564619719982, "keep": 0.8606753163039684, "leak": 0.7938976604491472, "replacement": 1.745959460735321}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 2.8748323917388916, "loss_scale": 4.0, "retries": 0, "seconds": 394.9, "max_memory_gb": 18.08, "pair": 0.13899901509284973, "sign": 0.08683517295867205, "keep": 0.11115048825740814, "leak": 1.04218715056777, "replacement": 0.49853862822055817}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 3.911066770553589, "loss_scale": 4.0, "retries": 0, "seconds": 395.8, "max_memory_gb": 18.08, "pair": 0.2760431654751301, "sign": 0.2155747776851058, "keep": 0.14908804465085268, "leak": 0.9521568268537521, "replacement": 0.6964137926697731}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 2.4333600997924805, "loss_scale": 4.0, "retries": 0, "seconds": 396.8, "max_memory_gb": 18.08, "pair": 0.5965766459703445, "sign": 0.6567390314303339, "keep": 0.49217979796230793, "leak": 0.23953384347259998, "replacement": 1.2309195399284363}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 3.9449832439422607, "loss_scale": 4.0, "retries": 0, "seconds": 397.7, "max_memory_gb": 18.08, "pair": 0.30185187235474586, "sign": 0.32801760360598564, "keep": 0.0562437332700938, "leak": 1.625192865729332, "replacement": 0.9002807438373566}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 4.01553201675415, "loss_scale": 4.0, "retries": 0, "seconds": 398.7, "max_memory_gb": 18.08, "pair": 0.1755197811871767, "sign": 0.11361707374453545, "keep": 0.01808855979470536, "leak": 1.0623029246926308, "replacement": 0.5069483369588852}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 3.4777190685272217, "loss_scale": 4.0, "retries": 0, "seconds": 399.6, "max_memory_gb": 18.08, "pair": 0.21480571664869785, "sign": 0.08100312110036612, "keep": 0.029460991674568504, "leak": 0.8621475733816624, "replacement": 0.4855746775865555}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 2.8507778644561768, "loss_scale": 4.0, "retries": 0, "seconds": 400.5, "max_memory_gb": 18.08, "pair": 0.26461399532854557, "sign": 0.2465452142059803, "keep": 0.2963593590538949, "leak": 0.68990283831954, "replacement": 0.7085419818758965}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 6.148624420166016, "loss_scale": 4.0, "retries": 0, "seconds": 401.4, "max_memory_gb": 18.08, "pair": 0.20078876335173845, "sign": 0.07253010058775544, "keep": 0.24918307972257026, "leak": 0.9765904466621578, "replacement": 0.6057929694652557}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 2.6309938430786133, "loss_scale": 4.0, "retries": 0, "seconds": 402.4, "max_memory_gb": 18.08, "pair": 0.29739097878336906, "sign": 0.21504327468574047, "keep": 0.10615215508732945, "leak": 0.6999432309530675, "replacement": 0.6329744905233383}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 2.425848960876465, "loss_scale": 4.0, "retries": 0, "seconds": 403.3, "max_memory_gb": 18.08, "pair": 0.2996643241494894, "sign": 0.33861350268125534, "keep": 0.12899766629561782, "leak": 0.9308102279901505, "replacement": 0.7661724537611008}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 1.919450044631958, "loss_scale": 4.0, "retries": 0, "seconds": 404.2, "max_memory_gb": 18.08, "pair": 0.2237764187157154, "sign": 0.1558891460299492, "keep": 0.008051214623264968, "leak": 0.21193391020642594, "replacement": 0.35873008519411087}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 3.708198070526123, "loss_scale": 4.0, "retries": 0, "seconds": 405.1, "max_memory_gb": 18.08, "pair": 0.4027054253965616, "sign": 0.39918647333979607, "keep": 0.34122174396179616, "leak": 0.3928058445453644, "replacement": 0.8711109533905983}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 2.4663312435150146, "loss_scale": 4.0, "retries": 0, "seconds": 406.0, "max_memory_gb": 18.08, "pair": 0.255934439599514, "sign": 0.26013223454356194, "keep": 0.23786449991166592, "leak": 0.6452861689031124, "replacement": 0.6662543639540672}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 8.368413925170898, "loss_scale": 4.0, "retries": 0, "seconds": 407.0, "max_memory_gb": 18.08, "pair": 0.33119964227080345, "sign": 0.4259047140367329, "keep": 0.41615747893229127, "leak": 1.4899115078151226, "replacement": 1.1247086226940155}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 1.6422686576843262, "loss_scale": 4.0, "retries": 0, "seconds": 407.9, "max_memory_gb": 18.08, "pair": 0.16200745478272438, "sign": 0.10413246741518378, "keep": 0.1031100309046451, "leak": 0.3938904218375683, "replacement": 0.364101305603981}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 5.795284748077393, "loss_scale": 4.0, "retries": 0, "seconds": 408.8, "max_memory_gb": 18.08, "pair": 0.38154299929738045, "sign": 0.47472639195621014, "keep": 0.88513508066535, "leak": 0.38644884154200554, "replacement": 1.1580859497189522}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 2.725877285003662, "loss_scale": 4.0, "retries": 0, "seconds": 409.7, "max_memory_gb": 18.08, "pair": 0.30063309893012047, "sign": 0.38708698004484177, "keep": 0.03371259325649589, "leak": 0.6528699535410851, "replacement": 0.6742503941059113}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 5.183367729187012, "loss_scale": 4.0, "retries": 0, "seconds": 410.6, "max_memory_gb": 18.08, "pair": 0.633500013500452, "sign": 0.6973043028265238, "keep": 0.8001367441465845, "leak": 0.8244537338614464, "replacement": 1.588333934545517}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 2.471644401550293, "loss_scale": 4.0, "retries": 0, "seconds": 411.5, "max_memory_gb": 18.08, "pair": 0.43052076920866966, "sign": 0.4282463788986206, "keep": 0.07947923615574837, "leak": 1.1441831514239311, "replacement": 0.9704294055700302}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 1.5712608098983765, "loss_scale": 4.0, "retries": 0, "seconds": 412.5, "max_memory_gb": 18.08, "pair": 0.256082646548748, "sign": 0.3270312547683716, "keep": 0.07129687204724178, "leak": 0.5685983784496784, "replacement": 0.5973963141441345}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 3.154703378677368, "loss_scale": 4.0, "retries": 0, "seconds": 413.4, "max_memory_gb": 18.08, "pair": 0.30289178900420666, "sign": 0.45306683145463467, "keep": 0.28489184379577637, "leak": 0.5512168891727924, "replacement": 0.8096753507852554}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 3.5452938079833984, "loss_scale": 4.0, "retries": 0, "seconds": 414.3, "max_memory_gb": 18.08, "pair": 0.63951819203794, "sign": 0.8415095303207636, "keep": 0.9112003538757563, "leak": 0.861060220161562, "replacement": 1.7311382293701172}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 3.3102684020996094, "loss_scale": 4.0, "retries": 0, "seconds": 415.2, "max_memory_gb": 18.08, "pair": 0.2441832348704338, "sign": 0.2363654039800167, "keep": 0.2898779575480148, "leak": 0.4282358760247007, "replacement": 0.614363893866539}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 2.399545669555664, "loss_scale": 4.0, "retries": 0, "seconds": 416.2, "max_memory_gb": 18.08, "pair": 0.5118056796491146, "sign": 0.6739907488226891, "keep": 0.5753533169045113, "leak": 0.7049092135403043, "replacement": 1.3127050250768661}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 2.942915201187134, "loss_scale": 4.0, "retries": 0, "seconds": 417.1, "max_memory_gb": 18.08, "pair": 0.30334046110510826, "sign": 0.3269470240920782, "keep": 0.5888762993272394, "leak": 0.4409560011699796, "replacement": 0.8714911341667175}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 1.7622253894805908, "loss_scale": 4.0, "retries": 0, "seconds": 418.0, "max_memory_gb": 18.08, "pair": 0.14702091738581657, "sign": 0.18621649965643883, "keep": 0.04675634784507565, "leak": 0.7781930714845657, "replacement": 0.45805560797452927}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 4.1948561668396, "loss_scale": 4.0, "retries": 0, "seconds": 418.9, "max_memory_gb": 18.08, "pair": 0.2549957912415266, "sign": 0.15325158555060625, "keep": 0.13028750737430528, "leak": 0.9411343596875668, "replacement": 0.6320489421486855}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 3.966717004776001, "loss_scale": 4.0, "retries": 0, "seconds": 419.8, "max_memory_gb": 18.08, "pair": 0.17049002647399902, "sign": 0.12203934416174889, "keep": 0.03325764823239297, "leak": 0.8182025030255318, "replacement": 0.45268917083740234}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 4.556535720825195, "loss_scale": 4.0, "retries": 0, "seconds": 420.7, "max_memory_gb": 18.08, "pair": 0.4518941082060337, "sign": 0.5203271703794599, "keep": 0.6724703907966614, "leak": 0.642968273255974, "replacement": 1.2090349569916725}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 5.595543384552002, "loss_scale": 4.0, "retries": 0, "seconds": 421.7, "max_memory_gb": 18.08, "pair": 0.28647447749972343, "sign": 0.25961728394031525, "keep": 0.03520606248639524, "leak": 0.7991303079761565, "replacement": 0.6336687356233597}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 3.7584073543548584, "loss_scale": 4.0, "retries": 0, "seconds": 422.6, "max_memory_gb": 18.08, "pair": 0.36892225593328476, "sign": 0.2804031753912568, "keep": 0.4278631657361984, "leak": 0.6089069709996693, "replacement": 0.8752822205424309}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 4.366347789764404, "loss_scale": 4.0, "retries": 0, "seconds": 423.5, "max_memory_gb": 18.08, "pair": 0.18193834647536278, "sign": 0.3716252464801073, "keep": 0.6299469452351332, "leak": 0.825810968875885, "replacement": 0.8891771659255028}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 6.811075210571289, "loss_scale": 4.0, "retries": 0, "seconds": 424.4, "max_memory_gb": 18.08, "pair": 0.37539685517549515, "sign": 0.7607540301978588, "keep": 0.052338757552206516, "leak": 1.505865152925253, "replacement": 1.158409558236599}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 4.039304256439209, "loss_scale": 4.0, "retries": 0, "seconds": 425.4, "max_memory_gb": 18.08, "pair": 0.19534394331276417, "sign": 0.29559342935681343, "keep": 0.09772197133861482, "leak": 0.6579147204756737, "replacement": 0.5564803183078766}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 3.3313260078430176, "loss_scale": 4.0, "retries": 0, "seconds": 426.3, "max_memory_gb": 18.08, "pair": 0.4880008064210415, "sign": 0.5596703663468361, "keep": 0.21423108968883753, "leak": 0.5201506428420544, "replacement": 1.0049891918897629}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 3.4506006240844727, "loss_scale": 4.0, "retries": 0, "seconds": 427.2, "max_memory_gb": 18.08, "pair": 0.34938786178827286, "sign": 0.3777813743799925, "keep": 0.09698752081021667, "leak": 0.303102437639609, "replacement": 0.6625479459762573}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 3.156679630279541, "loss_scale": 4.0, "retries": 0, "seconds": 428.2, "max_memory_gb": 18.08, "pair": 0.27117770723998547, "sign": 0.3024659249931574, "keep": 0.032058387296274304, "leak": 0.9745492935180664, "replacement": 0.6820771992206573}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 3.0417215824127197, "loss_scale": 4.0, "retries": 0, "seconds": 429.1, "max_memory_gb": 18.08, "pair": 0.2759182173758745, "sign": 0.5619950331747532, "keep": 0.4476276934146881, "leak": 1.0513237155973911, "replacement": 1.043560542166233}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 2.973207473754883, "loss_scale": 4.0, "retries": 0, "seconds": 430.1, "max_memory_gb": 18.08, "pair": 0.23822784796357155, "sign": 0.2514737620949745, "keep": 0.024995765706989914, "leak": 0.9546757154166698, "replacement": 0.615131564438343}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 2.1420178413391113, "loss_scale": 4.0, "retries": 0, "seconds": 431.0, "max_memory_gb": 18.08, "pair": 0.16909561678767204, "sign": 0.16798538528382778, "keep": 0.011293778254184872, "leak": 0.542933389544487, "replacement": 0.3944685533642769}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 4.22949743270874, "loss_scale": 4.0, "retries": 0, "seconds": 431.9, "max_memory_gb": 18.08, "pair": 0.15133840963244438, "sign": 0.2350423615425825, "keep": 0.11065781192155555, "leak": 0.9197001904249191, "replacement": 0.5541135519742966}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 3.448878526687622, "loss_scale": 4.0, "retries": 0, "seconds": 432.9, "max_memory_gb": 18.08, "pair": 0.22757434099912643, "sign": 0.3557428433559835, "keep": 0.16028786404058337, "leak": 0.952589601278305, "replacement": 0.723737083375454}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 8.522185325622559, "loss_scale": 4.0, "retries": 0, "seconds": 433.8, "max_memory_gb": 18.08, "pair": 0.3421843834221363, "sign": 0.46664000023156404, "keep": 0.1036907663801685, "leak": 2.019288895651698, "replacement": 1.1321719586849213}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 3.675720453262329, "loss_scale": 4.0, "retries": 0, "seconds": 434.7, "max_memory_gb": 18.08, "pair": 0.21732813119888306, "sign": 0.14772124961018562, "keep": 0.06953271897509694, "leak": 0.9531148187816143, "replacement": 0.5642338246107101}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 3.236302375793457, "loss_scale": 4.0, "retries": 0, "seconds": 435.6, "max_memory_gb": 18.08, "pair": 0.20846301689743996, "sign": 0.16448112297803164, "keep": 0.017701123957522213, "leak": 0.5138694467023015, "replacement": 0.42802150547504425}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 5.041062355041504, "loss_scale": 4.0, "retries": 0, "seconds": 436.5, "max_memory_gb": 18.08, "pair": 0.3415811378508806, "sign": 0.16138963773846626, "keep": 0.3747392155928537, "leak": 0.5279521148186177, "replacement": 0.7416336089372635}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 3.177201509475708, "loss_scale": 4.0, "retries": 0, "seconds": 437.4, "max_memory_gb": 18.08, "pair": 0.13539146818220615, "sign": 0.07028682203963399, "keep": 0.007265992811881006, "leak": 0.5372612290084362, "replacement": 0.30848317220807076}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 4.24685001373291, "loss_scale": 4.0, "retries": 0, "seconds": 438.4, "max_memory_gb": 18.08, "pair": 0.4467851109802723, "sign": 0.6981537193059921, "keep": 0.5694125732406974, "leak": 0.31210736371576786, "replacement": 1.1585950404405594}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 3.14223051071167, "loss_scale": 4.0, "retries": 0, "seconds": 439.3, "max_memory_gb": 18.08, "pair": 0.3098724689334631, "sign": 0.251931743696332, "keep": 0.13211931567639112, "leak": 0.2230314053595066, "replacement": 0.557655856013298}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 7.470364093780518, "loss_scale": 4.0, "retries": 0, "seconds": 440.2, "max_memory_gb": 18.08, "pair": 0.3552567344158888, "sign": 0.20899290218949318, "keep": 0.15173133893404156, "leak": 0.41511182440444827, "replacement": 0.6393968239426613}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 2.2741599082946777, "loss_scale": 4.0, "retries": 0, "seconds": 441.1, "max_memory_gb": 18.08, "pair": 0.32491172663867474, "sign": 0.42108941450715065, "keep": 0.20172661822289228, "leak": 0.19953098520636559, "replacement": 0.686202485114336}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 2.8812623023986816, "loss_scale": 4.0, "retries": 0, "seconds": 442.0, "max_memory_gb": 18.08, "pair": 0.33004469983279705, "sign": 0.4383890014141798, "keep": 0.22805285998038016, "leak": 0.3904553949832916, "replacement": 0.7608794495463371}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 2.9701082706451416, "loss_scale": 4.0, "retries": 0, "seconds": 443.0, "max_memory_gb": 18.08, "pair": 0.2063883114606142, "sign": 0.30147759057581425, "keep": 0.310295038158074, "leak": 1.1497808024287224, "replacement": 0.7997198477387428}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 4.992951393127441, "loss_scale": 4.0, "retries": 0, "seconds": 443.9, "max_memory_gb": 18.08, "pair": 0.31269653514027596, "sign": 0.42943847738206387, "keep": 0.032642632781062275, "leak": 0.7911537252366543, "replacement": 0.7415254935622215}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 4.718103408813477, "loss_scale": 4.0, "retries": 0, "seconds": 444.8, "max_memory_gb": 18.08, "pair": 0.29089732095599174, "sign": 0.2514370824210346, "keep": 0.009708836558274925, "leak": 0.6165420659817755, "replacement": 0.5756057798862457}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 4.673788547515869, "loss_scale": 4.0, "retries": 0, "seconds": 445.7, "max_memory_gb": 18.08, "pair": 0.630951464176178, "sign": 0.5762723200023174, "keep": 0.5799512532539666, "leak": 0.3887184797786176, "replacement": 1.3062428832054138}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 2.359884262084961, "loss_scale": 4.0, "retries": 0, "seconds": 446.6, "max_memory_gb": 18.08, "pair": 0.46782934851944447, "sign": 0.537935696542263, "keep": 0.4369950592517853, "leak": 0.20785254333168268, "replacement": 1.0072578564286232}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 5.218808174133301, "loss_scale": 4.0, "retries": 0, "seconds": 447.6, "max_memory_gb": 18.08, "pair": 0.16684463620185852, "sign": 0.17008430254645646, "keep": 0.045840306440368295, "leak": 1.1236945185810328, "replacement": 0.5557305589318275}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 2.2985177040100098, "loss_scale": 4.0, "retries": 0, "seconds": 448.5, "max_memory_gb": 18.08, "pair": 0.2842071335762739, "sign": 0.45133667066693306, "keep": 0.2373119192197919, "leak": 0.8451637388207018, "replacement": 0.8398223668336868}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 5.480173110961914, "loss_scale": 4.0, "retries": 0, "seconds": 449.4, "max_memory_gb": 18.08, "pair": 0.5208753123879433, "sign": 0.742872953414917, "keep": 0.4139093179255724, "leak": 0.12908925162628293, "replacement": 1.1315387338399887}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 7.99068546295166, "loss_scale": 4.0, "retries": 0, "seconds": 450.3, "max_memory_gb": 18.08, "pair": 0.7063684165477753, "sign": 0.9505104534327984, "keep": 1.1002773009240627, "leak": 0.36831123009324074, "replacement": 1.8238401263952255}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 4.397697448730469, "loss_scale": 4.0, "retries": 0, "seconds": 451.2, "max_memory_gb": 18.08, "pair": 0.3946104608476162, "sign": 0.7264763191342354, "keep": 0.16919381776824594, "leak": 1.188930980861187, "replacement": 1.1396782770752907}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 4.891580581665039, "loss_scale": 4.0, "retries": 0, "seconds": 452.1, "max_memory_gb": 18.08, "pair": 0.4810077026486397, "sign": 0.6461763959378004, "keep": 0.6249734270386398, "leak": 0.5842142999172211, "replacement": 1.2626361846923828}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 2.8145430088043213, "loss_scale": 4.0, "retries": 0, "seconds": 453.0, "max_memory_gb": 18.08, "pair": 0.229572594165802, "sign": 0.169164820574224, "keep": 0.05793742905370891, "leak": 0.8930335231125355, "replacement": 0.566382110118866}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 1.8918206691741943, "loss_scale": 4.0, "retries": 0, "seconds": 454.0, "max_memory_gb": 18.08, "pair": 0.285328371450305, "sign": 0.4941195803694427, "keep": 0.3749130545184016, "leak": 0.7346268920227885, "replacement": 0.9035014063119888}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 4.408839702606201, "loss_scale": 4.0, "retries": 0, "seconds": 454.9, "max_memory_gb": 18.08, "pair": 0.26357460487633944, "sign": 0.3794477619230747, "keep": 0.7311511950974818, "leak": 1.2577149893622845, "replacement": 1.1333028115332127}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 5.609830379486084, "loss_scale": 4.0, "retries": 0, "seconds": 455.8, "max_memory_gb": 18.08, "pair": 0.2929923105984926, "sign": 0.37114670500159264, "keep": 0.038456897949799895, "leak": 0.9635923355817795, "replacement": 0.7386922091245651}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 4.358281135559082, "loss_scale": 4.0, "retries": 0, "seconds": 456.7, "max_memory_gb": 18.08, "pair": 0.17939608171582222, "sign": 0.3153982125222683, "keep": 0.19181646563811228, "leak": 1.8500696495175362, "replacement": 0.8955208249390125}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 3.35223650932312, "loss_scale": 4.0, "retries": 0, "seconds": 457.6, "max_memory_gb": 18.08, "pair": 0.13582704961299896, "sign": 0.12633382435888052, "keep": 0.01159695308888331, "leak": 1.2405155822634697, "replacement": 0.5149213299155235}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 5.706480503082275, "loss_scale": 4.0, "retries": 0, "seconds": 458.6, "max_memory_gb": 18.08, "pair": 0.6381998807191849, "sign": 0.8426526784896851, "keep": 0.41756727546453476, "leak": 0.42112782411277294, "replacement": 1.373591810464859}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 5.233283996582031, "loss_scale": 4.0, "retries": 0, "seconds": 459.5, "max_memory_gb": 18.08, "pair": 0.12004156038165092, "sign": 0.2172435987740755, "keep": 0.029197236988693476, "leak": 1.7429047301411629, "replacement": 0.6789881661534309}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 3.872549533843994, "loss_scale": 4.0, "retries": 0, "seconds": 460.4, "max_memory_gb": 18.08, "pair": 0.21864508651196957, "sign": 0.299017496407032, "keep": 0.05374109395779669, "leak": 0.5367111042141914, "replacement": 0.5292021706700325}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 3.1299264430999756, "loss_scale": 4.0, "retries": 0, "seconds": 461.3, "max_memory_gb": 18.08, "pair": 0.14152215421199799, "sign": 0.06102348119020462, "keep": 0.07794205844402313, "leak": 1.0021911561489105, "replacement": 0.4615527093410492}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 7.670608997344971, "loss_scale": 4.0, "retries": 0, "seconds": 462.2, "max_memory_gb": 18.08, "pair": 0.3423683438450098, "sign": 0.6101210806518793, "keep": 1.077306798499194, "leak": 1.2174926474690437, "replacement": 1.4904554337263107}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 4.19900369644165, "loss_scale": 4.0, "retries": 0, "seconds": 463.1, "max_memory_gb": 18.08, "pair": 0.45545584335923195, "sign": 0.41034118365496397, "keep": 0.2306529760826379, "leak": 0.9027910549193621, "replacement": 1.0016507059335709}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 3.25956130027771, "loss_scale": 4.0, "retries": 0, "seconds": 464.1, "max_memory_gb": 18.08, "pair": 0.3986855950206518, "sign": 0.3762376047670841, "keep": 0.30934673245064914, "leak": 0.29521962627768517, "replacement": 0.815282691270113}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 5.419785499572754, "loss_scale": 4.0, "retries": 0, "seconds": 465.0, "max_memory_gb": 18.08, "pair": 0.44098885729908943, "sign": 0.5758831016719341, "keep": 0.8054608106031083, "leak": 0.7180341500788927, "replacement": 1.3111693523824215}
diff --git a/runs/single_llava/replacement/L30_s13/adapter/README.md b/runs/single_llava/replacement/L30_s13/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae587920ca55972d94e38f783837f71be14f1dda
--- /dev/null
+++ b/runs/single_llava/replacement/L30_s13/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: /data/OPSD/assets/models/llava-1.5-7b-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/data/OPSD/assets/models/llava-1.5-7b-hf
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.20.0
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L30_s13/adapter/adapter_config.json b/runs/single_llava/replacement/L30_s13/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d4f5037a836be7d1d6d249f69022d20a88af3144
--- /dev/null
+++ b/runs/single_llava/replacement/L30_s13/adapter/adapter_config.json
@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.20.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.language_model.layers.30.self_attn.k_proj",
+    "model.language_model.layers.30.mlp.down_proj",
+    "model.language_model.layers.30.mlp.up_proj",
+    "model.language_model.layers.30.self_attn.v_proj",
+    "model.language_model.layers.30.self_attn.o_proj",
+    "model.language_model.layers.30.self_attn.q_proj",
+    "model.language_model.layers.30.mlp.gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L30_s13/config.json b/runs/single_llava/replacement/L30_s13/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..ae231da376cf2f52026981a03e847bcc250f14ef
--- /dev/null
+++ b/runs/single_llava/replacement/L30_s13/config.json
@@ -0,0 +1,24 @@
+{
+  "behavior": "replacement",
+  "seed": 13,
+  "model": "/data/OPSD/assets/models/llava-1.5-7b-hf",
+  "backbone": "llava",
+  "train": "/data/OPSD/assets/v3/replacement.jsonl",
+  "items": 1000,
+  "steps": 500,
+  "epochs": 2,
+  "steps_per_epoch": 250,
+  "groups_per_step": 4,
+  "lr": 2e-05,
+  "scheduler": "constant",
+  "warmup_steps": 0,
+  "final_fraction": 1.0,
+  "weight_decay": 0.0,
+  "clip": 1.0,
+  "initial_loss_scale": 4.0,
+  "adapter": {
+    "layer": 30,
+    "rank": 16,
+    "trainable_parameters": 1249280
+  }
+}
\ No newline at end of file
diff --git a/runs/single_llava/replacement/L30_s13/steps.jsonl b/runs/single_llava/replacement/L30_s13/steps.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2a94ecd0ed74c0e3403e32423acb784ce4067148
--- /dev/null
+++ b/runs/single_llava/replacement/L30_s13/steps.jsonl
@@ -0,0 +1,500 @@
+{"step": 1, "steps": 500, "lr": 2e-05, "grad_norm": 0.3429843783378601, "loss_scale": 4.0, "retries": 0, "seconds": 1.0, "max_memory_gb": 16.38, "pair": 1.0986123085021973, "sign": 1.1921392381191254, "keep": 3.3128167763352394, "leak": 0.0, "replacement": 3.3510903418064117}
+{"step": 2, "steps": 500, "lr": 2e-05, "grad_norm": 0.4436773657798767, "loss_scale": 4.0, "retries": 0, "seconds": 1.8, "max_memory_gb": 16.39, "pair": 1.0982657670974731, "sign": 0.53667813539505, "keep": 5.37558821355924, "leak": 7.74035250117322e-06, "replacement": 4.054400980472565}
+{"step": 3, "steps": 500, "lr": 2e-05, "grad_norm": 1.012261986732483, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 16.4, "pair": 1.100785344839096, "sign": 0.5417508380487561, "keep": 10.80703091621399, "leak": 3.92140566418675e-06, "replacement": 6.7751771211624146}
+{"step": 4, "steps": 500, "lr": 2e-05, "grad_norm": 0.7422756552696228, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 16.41, "pair": 1.09737029671669, "sign": 0.4724778402596712, "keep": 7.670053149573505, "leak": 3.964868324146664e-05, "replacement": 5.168645620346069}
+{"step": 5, "steps": 500, "lr": 2e-05, "grad_norm": 0.26236289739608765, "loss_scale": 4.0, "retries": 0, "seconds": 4.4, "max_memory_gb": 16.41, "pair": 1.0968212485313416, "sign": 0.7758724689483643, "keep": 0.1586710885167122, "leak": 1.6961476783095009e-06, "replacement": 1.5640934705734253}
+{"step": 6, "steps": 500, "lr": 2e-05, "grad_norm": 0.4073719084262848, "loss_scale": 4.0, "retries": 0, "seconds": 5.3, "max_memory_gb": 16.41, "pair": 1.0978506207466125, "sign": 0.7377481609582901, "keep": 2.9326701052486897, "leak": 2.126055356033163e-05, "replacement": 2.9330650866031647}
+{"step": 7, "steps": 500, "lr": 2e-05, "grad_norm": 0.8270846605300903, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 16.41, "pair": 1.0939617455005646, "sign": 0.5469365585595369, "keep": 6.881011146120727, "leak": 2.0176196358079324e-05, "replacement": 4.8079405426979065}
+{"step": 8, "steps": 500, "lr": 2e-05, "grad_norm": 0.9408368468284607, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 16.41, "pair": 1.0951908230781555, "sign": 0.5885861590504646, "keep": 7.390064761042595, "leak": 4.687002103764826e-06, "replacement": 5.084517419338226}
+{"step": 9, "steps": 500, "lr": 2e-05, "grad_norm": 0.592767059803009, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 16.41, "pair": 1.0984699428081512, "sign": 0.6193567225709558, "keep": 5.712316103279591, "leak": 2.0102002736388513e-05, "replacement": 4.264311373233795}
+{"step": 10, "steps": 500, "lr": 2e-05, "grad_norm": 0.4060381054878235, "loss_scale": 4.0, "retries": 0, "seconds": 8.7, "max_memory_gb": 16.41, "pair": 1.0951454043388367, "sign": 0.7821202576160431, "keep": 3.271262899041176, "leak": 1.7898400302307493e-05, "replacement": 3.1218414306640625}
+{"step": 11, "steps": 500, "lr": 2e-05, "grad_norm": 0.821196436882019, "loss_scale": 4.0, "retries": 0, "seconds": 9.6, "max_memory_gb": 16.41, "pair": 1.0960260331630707, "sign": 0.45448407903313637, "keep": 5.267575999721885, "leak": 3.800328795477981e-05, "replacement": 3.9570655822753906}
+{"step": 12, "steps": 500, "lr": 2e-05, "grad_norm": 0.4728087782859802, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 16.41, "pair": 1.0951949656009674, "sign": 0.6127127483487129, "keep": 3.056115133687854, "leak": 4.9967074232881714e-05, "replacement": 2.929621398448944}
+{"step": 13, "steps": 500, "lr": 2e-05, "grad_norm": 0.8161009550094604, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 16.41, "pair": 1.0954241156578064, "sign": 0.5126253589987755, "keep": 5.3559667486697435, "leak": 6.734106864314526e-05, "replacement": 4.029737055301666}
+{"step": 14, "steps": 500, "lr": 2e-05, "grad_norm": 1.1725826263427734, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 16.41, "pair": 1.094210535287857, "sign": 0.400267181918025, "keep": 7.789187416434288, "leak": 6.516099529108033e-05, "replacement": 5.188954144716263}
+{"step": 15, "steps": 500, "lr": 2e-05, "grad_norm": 0.8149782419204712, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 16.41, "pair": 1.0917468965053558, "sign": 0.6906115114688873, "keep": 5.185595519607887, "leak": 0.00013448012032313272, "replacement": 4.029884070158005}
+{"step": 16, "steps": 500, "lr": 2e-05, "grad_norm": 0.5712746381759644, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 16.41, "pair": 1.0937826335430145, "sign": 0.7851769849658012, "keep": 2.7700998107902706, "leak": 7.754341731924796e-05, "replacement": 2.8714404702186584}
+{"step": 17, "steps": 500, "lr": 2e-05, "grad_norm": 0.5774487853050232, "loss_scale": 4.0, "retries": 0, "seconds": 14.7, "max_memory_gb": 16.41, "pair": 1.0920360088348389, "sign": 0.7161311507225037, "keep": 3.1917530503124, "leak": 0.00010866702268685913, "replacement": 3.04600527882576}
+{"step": 18, "steps": 500, "lr": 2e-05, "grad_norm": 1.5675479173660278, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 16.43, "pair": 1.0896675884723663, "sign": 0.5414039939641953, "keep": 7.190407142043114, "leak": 9.894790150610788e-05, "replacement": 4.955597847700119}
+{"step": 19, "steps": 500, "lr": 2e-05, "grad_norm": 2.267014265060425, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 16.43, "pair": 1.0852871537208557, "sign": 0.5427628234028816, "keep": 10.679415464401245, "leak": 0.0005727196185603134, "replacement": 6.696519374847412}
+{"step": 20, "steps": 500, "lr": 2e-05, "grad_norm": 0.5365700721740723, "loss_scale": 4.0, "retries": 0, "seconds": 17.3, "max_memory_gb": 16.43, "pair": 1.0838044881820679, "sign": 0.430129311978817, "keep": 0.023923699278384447, "leak": 0.00034807542397174984, "replacement": 1.3109180331230164}
+{"step": 21, "steps": 500, "lr": 2e-05, "grad_norm": 1.1727731227874756, "loss_scale": 4.0, "retries": 0, "seconds": 18.1, "max_memory_gb": 16.43, "pair": 1.0853038430213928, "sign": 0.3780562411993742, "keep": 5.5351408501155674, "leak": 0.0002881675827666186, "replacement": 4.041974484920502}
+{"step": 22, "steps": 500, "lr": 2e-05, "grad_norm": 1.1507128477096558, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 16.43, "pair": 1.0859830975532532, "sign": 0.5753970555961132, "keep": 5.938040215522051, "leak": 0.0003515311473165639, "replacement": 4.342789590358734}
+{"step": 23, "steps": 500, "lr": 2e-05, "grad_norm": 1.3786067962646484, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 16.43, "pair": 1.0783152282238007, "sign": 0.2544085681438446, "keep": 5.286244811490178, "leak": 0.0007979920992511325, "replacement": 3.8488413393497467}
+{"step": 24, "steps": 500, "lr": 2e-05, "grad_norm": 2.343310594558716, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 16.43, "pair": 1.0799534618854523, "sign": 0.2697269320487976, "keep": 10.022592782974243, "leak": 0.00036188875674270093, "replacement": 6.226203680038452}
+{"step": 25, "steps": 500, "lr": 2e-05, "grad_norm": 2.121293783187866, "loss_scale": 4.0, "retries": 0, "seconds": 21.6, "max_memory_gb": 16.43, "pair": 1.084597498178482, "sign": 0.39891596138477325, "keep": 8.436541698873043, "leak": 0.0006919998268211103, "replacement": 5.5024994015693665}
+{"step": 26, "steps": 500, "lr": 2e-05, "grad_norm": 2.5593831539154053, "loss_scale": 4.0, "retries": 0, "seconds": 22.5, "max_memory_gb": 16.43, "pair": 1.0738567113876343, "sign": 0.6624221866950393, "keep": 10.506862044334412, "leak": 0.0013827558418597619, "replacement": 6.658844470977783}
+{"step": 27, "steps": 500, "lr": 2e-05, "grad_norm": 1.861872673034668, "loss_scale": 4.0, "retries": 0, "seconds": 23.4, "max_memory_gb": 16.43, "pair": 1.085176169872284, "sign": 0.6014249213039875, "keep": 8.418876774609089, "leak": 0.0003858997515635565, "replacement": 5.595423549413681}
+{"step": 28, "steps": 500, "lr": 2e-05, "grad_norm": 1.4697927236557007, "loss_scale": 4.0, "retries": 0, "seconds": 24.2, "max_memory_gb": 16.43, "pair": 1.0822704434394836, "sign": 0.7830134350806475, "keep": 5.628555208444595, "leak": 0.002157469691155711, "replacement": 4.288593947887421}
+{"step": 29, "steps": 500, "lr": 2e-05, "grad_norm": 1.2549256086349487, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 16.43, "pair": 1.0796819031238556, "sign": 0.8353481628000736, "keep": 5.20834041479975, "leak": 0.0009814336053750594, "replacement": 4.101771533489227}
+{"step": 30, "steps": 500, "lr": 2e-05, "grad_norm": 1.7704213857650757, "loss_scale": 4.0, "retries": 0, "seconds": 26.0, "max_memory_gb": 16.43, "pair": 1.062389761209488, "sign": 0.3487458247691393, "keep": 5.515985967591405, "leak": 0.0017817557408312812, "replacement": 3.995201051235199}
+{"step": 31, "steps": 500, "lr": 2e-05, "grad_norm": 1.660224437713623, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 16.43, "pair": 1.0812162458896637, "sign": 0.8073747456073761, "keep": 5.67985999584198, "leak": 0.0005693130453403583, "replacement": 4.324975907802582}
+{"step": 32, "steps": 500, "lr": 2e-05, "grad_norm": 2.9931981563568115, "loss_scale": 4.0, "retries": 0, "seconds": 27.7, "max_memory_gb": 16.43, "pair": 1.0683910250663757, "sign": 0.40035247802734375, "keep": 10.694684505462646, "leak": 0.004118029173696414, "replacement": 6.616939067840576}
+{"step": 33, "steps": 500, "lr": 2e-05, "grad_norm": 1.8088352680206299, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 16.43, "pair": 1.0768182575702667, "sign": 0.6290418431162834, "keep": 4.8969846656546, "leak": 0.0014793273621762637, "replacement": 3.8402012288570404}
+{"step": 34, "steps": 500, "lr": 2e-05, "grad_norm": 1.983383059501648, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 16.43, "pair": 1.065633088350296, "sign": 0.5324632003903389, "keep": 5.268202092498541, "leak": 0.0025814954497036524, "replacement": 3.9666111171245575}
+{"step": 35, "steps": 500, "lr": 2e-05, "grad_norm": 1.8969272375106812, "loss_scale": 4.0, "retries": 0, "seconds": 30.3, "max_memory_gb": 16.43, "pair": 1.0615380704402924, "sign": 0.43556130304932594, "keep": 5.489829547703266, "leak": 0.002143896846973803, "replacement": 4.024769634008408}
+{"step": 36, "steps": 500, "lr": 2e-05, "grad_norm": 1.7926617860794067, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 16.43, "pair": 1.0576203763484955, "sign": 0.437847463414073, "keep": 5.1383847668766975, "leak": 0.0014582853436877485, "replacement": 3.846100926399231}
+{"step": 37, "steps": 500, "lr": 2e-05, "grad_norm": 1.9022445678710938, "loss_scale": 4.0, "retries": 0, "seconds": 32.0, "max_memory_gb": 16.43, "pair": 1.0619879364967346, "sign": 0.5923036374151707, "keep": 4.9532270058989525, "leak": 0.002850579306042045, "replacement": 3.8354659378528595}
+{"step": 38, "steps": 500, "lr": 2e-05, "grad_norm": 3.4735796451568604, "loss_scale": 4.0, "retries": 0, "seconds": 32.9, "max_memory_gb": 16.43, "pair": 1.054140865802765, "sign": 0.7805880643427372, "keep": 9.428683280944824, "leak": 0.007999472960364074, "replacement": 6.160776495933533}
+{"step": 39, "steps": 500, "lr": 2e-05, "grad_norm": 1.484021544456482, "loss_scale": 4.0, "retries": 0, "seconds": 33.8, "max_memory_gb": 16.43, "pair": 1.0625324249267578, "sign": 0.5631309859454632, "keep": 2.899188106879592, "leak": 0.0022623269906034693, "replacement": 2.794257551431656}
+{"step": 40, "steps": 500, "lr": 2e-05, "grad_norm": 2.1676957607269287, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 16.43, "pair": 1.062167763710022, "sign": 0.5050944425165653, "keep": 5.459860548377037, "leak": 0.0017397544070263393, "replacement": 4.045080304145813}
+{"step": 41, "steps": 500, "lr": 2e-05, "grad_norm": 1.3377463817596436, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 16.43, "pair": 1.0849041640758514, "sign": 1.0278205573558807, "keep": 2.9161222130060196, "leak": 0.001204199820165286, "replacement": 3.057176560163498}
+{"step": 42, "steps": 500, "lr": 2e-05, "grad_norm": 1.9727965593338013, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 16.43, "pair": 1.038846641778946, "sign": 0.4029136523604393, "keep": 4.734323691576719, "leak": 0.002720647433307022, "replacement": 3.608145445585251}
+{"step": 43, "steps": 500, "lr": 2e-05, "grad_norm": 2.1596431732177734, "loss_scale": 4.0, "retries": 0, "seconds": 37.2, "max_memory_gb": 16.43, "pair": 1.055245041847229, "sign": 0.5210654772818089, "keep": 4.918579608201981, "leak": 0.0023455975751858205, "replacement": 3.7756539583206177}
+{"step": 44, "steps": 500, "lr": 2e-05, "grad_norm": 1.6431639194488525, "loss_scale": 4.0, "retries": 0, "seconds": 38.1, "max_memory_gb": 16.43, "pair": 1.0504529178142548, "sign": 0.5793143659830093, "keep": 2.731611644849181, "leak": 0.005896723858313635, "replacement": 2.707390248775482}
+{"step": 45, "steps": 500, "lr": 2e-05, "grad_norm": 1.1945501565933228, "loss_scale": 4.0, "retries": 0, "seconds": 38.9, "max_memory_gb": 16.43, "pair": 1.0753966867923737, "sign": 0.8936696425080299, "keep": 2.8354460448026657, "leak": 0.0008804031786553423, "replacement": 2.9401746094226837}
+{"step": 46, "steps": 500, "lr": 2e-05, "grad_norm": 2.518996238708496, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 16.43, "pair": 1.050042986869812, "sign": 0.7245800979435444, "keep": 4.60342401266098, "leak": 0.010711130609706743, "replacement": 3.7167228758335114}
+{"step": 47, "steps": 500, "lr": 2e-05, "grad_norm": 2.7387092113494873, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 16.43, "pair": 1.028320848941803, "sign": 0.6938414163887501, "keep": 4.068283984437585, "leak": 0.0054384632530855015, "replacement": 3.4107432067394257}
+{"step": 48, "steps": 500, "lr": 2e-05, "grad_norm": 3.733739137649536, "loss_scale": 4.0, "retries": 0, "seconds": 41.5, "max_memory_gb": 16.43, "pair": 1.0066991448402405, "sign": 0.20861215237528086, "keep": 7.627080678939819, "leak": 0.010680867824703455, "replacement": 4.927215725183487}
+{"step": 49, "steps": 500, "lr": 2e-05, "grad_norm": 3.574000597000122, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 16.43, "pair": 1.0382517874240875, "sign": 0.6006822064518929, "keep": 7.126338344067335, "leak": 0.005815208282001549, "replacement": 4.903215795755386}
+{"step": 50, "steps": 500, "lr": 2e-05, "grad_norm": 3.7111597061157227, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 16.43, "pair": 1.0375053584575653, "sign": 0.461838036775589, "keep": 7.245344944298267, "leak": 0.00847145402804017, "replacement": 4.893214702606201}
+{"step": 51, "steps": 500, "lr": 2e-05, "grad_norm": 1.761353850364685, "loss_scale": 4.0, "retries": 0, "seconds": 44.1, "max_memory_gb": 16.43, "pair": 1.0337559580802917, "sign": 0.5812417455017567, "keep": 2.680713805370033, "leak": 0.015045952146465424, "replacement": 2.668495237827301}
+{"step": 52, "steps": 500, "lr": 2e-05, "grad_norm": 3.3885769844055176, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 16.43, "pair": 1.042187049984932, "sign": 0.7087653279304504, "keep": 4.668413996696472, "leak": 0.0047280029830290005, "replacement": 3.7319586277008057}
+{"step": 53, "steps": 500, "lr": 2e-05, "grad_norm": 3.6587674617767334, "loss_scale": 4.0, "retries": 0, "seconds": 45.8, "max_memory_gb": 16.43, "pair": 1.049127221107483, "sign": 0.7901832014322281, "keep": 7.390839695930481, "leak": 0.024036169143073494, "replacement": 5.1456476747989655}
+{"step": 54, "steps": 500, "lr": 2e-05, "grad_norm": 4.140815734863281, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 16.43, "pair": 1.0373325794935226, "sign": 0.6644663885235786, "keep": 7.821972459554672, "leak": 0.021404003418865614, "replacement": 5.285902976989746}
+{"step": 55, "steps": 500, "lr": 2e-05, "grad_norm": 2.6822707653045654, "loss_scale": 4.0, "retries": 0, "seconds": 47.5, "max_memory_gb": 16.43, "pair": 1.014402449131012, "sign": 0.47845447808504105, "keep": 4.9685486406087875, "leak": 0.021495473221875727, "replacement": 3.743277996778488}
+{"step": 56, "steps": 500, "lr": 2e-05, "grad_norm": 5.288413047790527, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 16.43, "pair": 1.0081507563591003, "sign": 0.21783548779785633, "keep": 9.557452917098999, "leak": 0.009802538115764037, "replacement": 5.898245573043823}
+{"step": 57, "steps": 500, "lr": 2e-05, "grad_norm": 1.4901350736618042, "loss_scale": 4.0, "retries": 0, "seconds": 49.2, "max_memory_gb": 16.43, "pair": 1.0340093225240707, "sign": 0.7917463593184948, "keep": 2.464253645390272, "leak": 0.004838863664190285, "replacement": 2.6632189750671387}
+{"step": 58, "steps": 500, "lr": 2e-05, "grad_norm": 3.081835985183716, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 16.43, "pair": 0.9877592027187347, "sign": 0.2857147641479969, "keep": 4.799713067826815, "leak": 0.03050208743661642, "replacement": 3.53809854388237}
+{"step": 59, "steps": 500, "lr": 2e-05, "grad_norm": 1.7102758884429932, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 16.43, "pair": 1.0473313927650452, "sign": 0.6727709919214249, "keep": 2.7954099141061306, "leak": 0.0015439522111668769, "replacement": 2.78180792927742}
+{"step": 60, "steps": 500, "lr": 2e-05, "grad_norm": 2.9283642768859863, "loss_scale": 4.0, "retries": 0, "seconds": 51.7, "max_memory_gb": 16.43, "pair": 1.0154751241207123, "sign": 0.6432036012411118, "keep": 4.787855628877878, "leak": 0.008958589460235089, "replacement": 3.73324453830719}
+{"step": 61, "steps": 500, "lr": 2e-05, "grad_norm": 4.494391918182373, "loss_scale": 4.0, "retries": 0, "seconds": 52.6, "max_memory_gb": 16.43, "pair": 1.0160079151391983, "sign": 0.6435596533119678, "keep": 6.910176485776901, "leak": 0.016239848114594224, "replacement": 4.796935975551605}
+{"step": 62, "steps": 500, "lr": 2e-05, "grad_norm": 1.8154653310775757, "loss_scale": 4.0, "retries": 0, "seconds": 53.5, "max_memory_gb": 16.43, "pair": 1.0301435738801956, "sign": 0.7172935828566551, "keep": 2.645821970887482, "leak": 0.0030055475167500845, "replacement": 2.712452620267868}
+{"step": 63, "steps": 500, "lr": 2e-05, "grad_norm": 4.374649524688721, "loss_scale": 4.0, "retries": 0, "seconds": 54.4, "max_memory_gb": 16.43, "pair": 1.0232673734426498, "sign": 0.6065308228135109, "keep": 7.22475357260555, "leak": 0.01728205749532208, "replacement": 4.943229973316193}
+{"step": 64, "steps": 500, "lr": 2e-05, "grad_norm": 4.688249111175537, "loss_scale": 4.0, "retries": 0, "seconds": 55.3, "max_memory_gb": 16.44, "pair": 0.9759325683116913, "sign": 0.49956621788442135, "keep": 7.202574126422405, "leak": 0.07284784794319421, "replacement": 4.8452147245407104}
+{"step": 65, "steps": 500, "lr": 2e-05, "grad_norm": 4.304069519042969, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 16.44, "pair": 1.0085057020187378, "sign": 0.5737119130790234, "keep": 7.333806276321411, "leak": 0.04509319421777036, "replacement": 4.973538279533386}
+{"step": 66, "steps": 500, "lr": 2e-05, "grad_norm": 1.9872257709503174, "loss_scale": 4.0, "retries": 0, "seconds": 57.0, "max_memory_gb": 16.44, "pair": 1.018085092306137, "sign": 0.6355275511741638, "keep": 1.8132622155826539, "leak": 0.016024449290853227, "replacement": 2.246486097574234}
+{"step": 67, "steps": 500, "lr": 2e-05, "grad_norm": 5.406075954437256, "loss_scale": 4.0, "retries": 0, "seconds": 57.9, "max_memory_gb": 16.44, "pair": 0.9708651006221771, "sign": 0.4036771971732378, "keep": 6.854444712400436, "leak": 0.04006745414517354, "replacement": 4.609942972660065}
+{"step": 68, "steps": 500, "lr": 2e-05, "grad_norm": 1.927489161491394, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 16.44, "pair": 0.9940263628959656, "sign": 0.5020534992218018, "keep": 2.3330444213934243, "leak": 0.014978509900174686, "replacement": 2.415319949388504}
+{"step": 69, "steps": 500, "lr": 2e-05, "grad_norm": 3.1106162071228027, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 16.44, "pair": 0.9936578571796417, "sign": 0.7723587527871132, "keep": 4.454347239807248, "leak": 0.008164108619780563, "replacement": 3.6090519428253174}
+{"step": 70, "steps": 500, "lr": 2e-05, "grad_norm": 3.6053106784820557, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 16.44, "pair": 1.003837138414383, "sign": 0.6488193348050117, "keep": 4.452203824184835, "leak": 0.088447962058126, "replacement": 3.5764607787132263}
+{"step": 71, "steps": 500, "lr": 2e-05, "grad_norm": 5.3566436767578125, "loss_scale": 4.0, "retries": 0, "seconds": 61.5, "max_memory_gb": 16.44, "pair": 0.9466771334409714, "sign": 0.20198733173310757, "keep": 6.6608860567212105, "leak": 0.04682429302397395, "replacement": 4.389819860458374}
+{"step": 72, "steps": 500, "lr": 2e-05, "grad_norm": 6.111286640167236, "loss_scale": 4.0, "retries": 0, "seconds": 62.4, "max_memory_gb": 16.44, "pair": 0.920475572347641, "sign": 0.328972902148962, "keep": 7.696317672729492, "leak": 0.0532418544171378, "replacement": 4.946431219577789}
+{"step": 73, "steps": 500, "lr": 2e-05, "grad_norm": 0.7519957423210144, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 16.44, "pair": 1.0344152003526688, "sign": 0.8065996766090393, "keep": 0.4339986052364111, "leak": 0.0034464147056496586, "replacement": 1.6555759012699127}
+{"step": 74, "steps": 500, "lr": 2e-05, "grad_norm": 3.7040722370147705, "loss_scale": 4.0, "retries": 0, "seconds": 64.0, "max_memory_gb": 16.44, "pair": 0.9353613257408142, "sign": 0.5087739415466785, "keep": 4.4340859670192, "leak": 0.043996487836409415, "replacement": 3.4177902340888977}
+{"step": 75, "steps": 500, "lr": 2e-05, "grad_norm": 2.6203935146331787, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 16.44, "pair": 1.0138002038002014, "sign": 0.9053869992494583, "keep": 3.304459363222122, "leak": 0.02775756872870261, "replacement": 3.1256628334522247}
+{"step": 76, "steps": 500, "lr": 2e-05, "grad_norm": 3.5230629444122314, "loss_scale": 4.0, "retries": 0, "seconds": 65.8, "max_memory_gb": 16.44, "pair": 0.9587908089160919, "sign": 0.46292968839406967, "keep": 4.116651512682438, "leak": 0.04397328340655804, "replacement": 3.25957453250885}
+{"step": 77, "steps": 500, "lr": 2e-05, "grad_norm": 5.191216468811035, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 16.44, "pair": 0.958637535572052, "sign": 0.5146917514503002, "keep": 6.908583879470825, "leak": 0.1726026072865352, "replacement": 4.713425874710083}
+{"step": 78, "steps": 500, "lr": 2e-05, "grad_norm": 4.084500789642334, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 16.44, "pair": 0.9782777279615402, "sign": 0.7541409879922867, "keep": 3.607687057927251, "leak": 0.08055625343695283, "replacement": 3.1793309152126312}
+{"step": 79, "steps": 500, "lr": 2e-05, "grad_norm": 2.71005916595459, "loss_scale": 4.0, "retries": 0, "seconds": 68.3, "max_memory_gb": 16.44, "pair": 0.9687601923942566, "sign": 0.5847014002501965, "keep": 1.7826335304416716, "leak": 0.02541632574866526, "replacement": 2.1587817668914795}
+{"step": 80, "steps": 500, "lr": 2e-05, "grad_norm": 4.129029273986816, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 16.44, "pair": 0.9136436581611633, "sign": 0.38913628831505775, "keep": 3.981110935565084, "leak": 0.09590829256922007, "replacement": 3.1227444410324097}
+{"step": 81, "steps": 500, "lr": 2e-05, "grad_norm": 6.044498920440674, "loss_scale": 4.0, "retries": 0, "seconds": 70.0, "max_memory_gb": 16.44, "pair": 0.9172491431236267, "sign": 0.5758843645453453, "keep": 5.919967912137508, "leak": 0.10643186602464993, "replacement": 4.19178318977356}
+{"step": 82, "steps": 500, "lr": 2e-05, "grad_norm": 6.080842971801758, "loss_scale": 4.0, "retries": 0, "seconds": 70.8, "max_memory_gb": 16.44, "pair": 0.9110153019428253, "sign": 0.3956931997090578, "keep": 5.939788036048412, "leak": 0.031313237892391044, "replacement": 4.086584210395813}
+{"step": 83, "steps": 500, "lr": 2e-05, "grad_norm": 5.944197177886963, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 16.44, "pair": 0.9039703011512756, "sign": 0.5870863273739815, "keep": 5.043425345793366, "leak": 0.14801456267014146, "replacement": 3.756229817867279}
+{"step": 84, "steps": 500, "lr": 2e-05, "grad_norm": 2.3016369342803955, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 16.44, "pair": 0.9524645358324051, "sign": 0.8367407843470573, "keep": 2.2148355350364, "leak": 0.07261594750161748, "replacement": 2.4964066445827484}
+{"step": 85, "steps": 500, "lr": 2e-05, "grad_norm": 1.240829348564148, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 16.44, "pair": 0.9428984671831131, "sign": 0.5659641996026039, "keep": 0.14493259554728866, "leak": 0.040229860780527815, "replacement": 1.3084043264389038}
+{"step": 86, "steps": 500, "lr": 2e-05, "grad_norm": 5.658081531524658, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 16.44, "pair": 0.9705272167921066, "sign": 0.7262411117553711, "keep": 5.5152798891067505, "leak": 0.037760297400382115, "replacement": 4.10072785615921}
+{"step": 87, "steps": 500, "lr": 2e-05, "grad_norm": 5.822402000427246, "loss_scale": 4.0, "retries": 0, "seconds": 75.2, "max_memory_gb": 16.44, "pair": 0.9030517786741257, "sign": 0.5143106058239937, "keep": 5.56747311539948, "leak": 0.23037113809914445, "replacement": 4.00153648853302}
+{"step": 88, "steps": 500, "lr": 2e-05, "grad_norm": 4.214844703674316, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 16.44, "pair": 0.9236446022987366, "sign": 0.6479366607964039, "keep": 3.750828796532005, "leak": 0.09554988361924188, "replacement": 3.1469148099422455}
+{"step": 89, "steps": 500, "lr": 2e-05, "grad_norm": 1.1430655717849731, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 16.44, "pair": 0.9395520091056824, "sign": 0.6871403232216835, "keep": 0.2155464687384665, "leak": 0.053058616520502255, "replacement": 1.404160052537918}
+{"step": 90, "steps": 500, "lr": 2e-05, "grad_norm": 3.9179208278656006, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 16.44, "pair": 0.9028405398130417, "sign": 0.6235973984003067, "keep": 3.2068257227074355, "leak": 0.04822262885863893, "replacement": 2.8301077485084534}
+{"step": 91, "steps": 500, "lr": 2e-05, "grad_norm": 0.5686456561088562, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 16.44, "pair": 1.0419669300317764, "sign": 1.017463281750679, "keep": 0.6296183876693249, "leak": 0.008228978781517071, "replacement": 1.8675650358200073}
+{"step": 92, "steps": 500, "lr": 2e-05, "grad_norm": 2.5927200317382812, "loss_scale": 4.0, "retries": 0, "seconds": 79.7, "max_memory_gb": 16.44, "pair": 0.9248624891042709, "sign": 0.6273143589496613, "keep": 1.8727017213823274, "leak": 0.23858215380460024, "replacement": 2.234515994787216}
+{"step": 93, "steps": 500, "lr": 2e-05, "grad_norm": 6.138161659240723, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 16.44, "pair": 0.8193747699260712, "sign": 0.30620321072638035, "keep": 3.144130044034682, "leak": 0.14076236682012677, "replacement": 2.5797320902347565}
+{"step": 94, "steps": 500, "lr": 2e-05, "grad_norm": 6.8511481285095215, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 16.44, "pair": 0.8914054781198502, "sign": 0.4368528127670288, "keep": 5.4146265387535095, "leak": 0.06532959035530439, "replacement": 3.8334774672985077}
+{"step": 95, "steps": 500, "lr": 2e-05, "grad_norm": 4.5071916580200195, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 16.44, "pair": 0.9187455177307129, "sign": 0.5336770955473185, "keep": 3.6533209634944797, "leak": 0.04895551678600896, "replacement": 3.024483472108841}
+{"step": 96, "steps": 500, "lr": 2e-05, "grad_norm": 6.672280311584473, "loss_scale": 4.0, "retries": 0, "seconds": 83.1, "max_memory_gb": 16.44, "pair": 0.842144712805748, "sign": 0.5198903419077396, "keep": 5.3483670353889465, "leak": 0.4266081425284938, "replacement": 3.88292533159256}
+{"step": 97, "steps": 500, "lr": 2e-05, "grad_norm": 2.7710113525390625, "loss_scale": 4.0, "retries": 0, "seconds": 84.0, "max_memory_gb": 16.44, "pair": 0.907550647854805, "sign": 0.5419555678963661, "keep": 2.3239559084177017, "leak": 0.03758772544097155, "replacement": 2.3499032258987427}
+{"step": 98, "steps": 500, "lr": 2e-05, "grad_norm": 4.617015838623047, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 16.44, "pair": 0.8844524472951889, "sign": 0.5391490273177624, "keep": 3.2676425226964056, "leak": 0.18717070762068033, "replacement": 2.834640920162201}
+{"step": 99, "steps": 500, "lr": 2e-05, "grad_norm": 0.8401209712028503, "loss_scale": 4.0, "retries": 0, "seconds": 85.7, "max_memory_gb": 16.44, "pair": 0.9542196094989777, "sign": 0.8601306080818176, "keep": 0.394032180076465, "leak": 0.03332203649915755, "replacement": 1.5896315276622772}
+{"step": 100, "steps": 500, "lr": 2e-05, "grad_norm": 6.118719100952148, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 16.44, "pair": 0.8920648992061615, "sign": 0.7257710415869951, "keep": 4.92828081548214, "leak": 0.2329818310099654, "replacement": 3.777336299419403}
+{"step": 101, "steps": 500, "lr": 2e-05, "grad_norm": 7.015194416046143, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 16.44, "pair": 0.6888466775417328, "sign": 0.17958373483270407, "keep": 4.643523573875427, "leak": 0.1421838579699397, "replacement": 3.135946273803711}
+{"step": 102, "steps": 500, "lr": 2e-05, "grad_norm": 6.464426040649414, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 16.44, "pair": 0.8170835822820663, "sign": 0.3674066960811615, "keep": 5.283877715468407, "leak": 0.25263398536480963, "replacement": 3.705884277820587}
+{"step": 103, "steps": 500, "lr": 2e-05, "grad_norm": 4.667106628417969, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 16.44, "pair": 0.8254997432231903, "sign": 0.39426377043128014, "keep": 3.022270683897659, "leak": 0.2011658539995551, "replacement": 2.5840584337711334}
+{"step": 104, "steps": 500, "lr": 2e-05, "grad_norm": 3.1138734817504883, "loss_scale": 4.0, "retries": 0, "seconds": 90.0, "max_memory_gb": 16.44, "pair": 0.8939315527677536, "sign": 0.8847213685512543, "keep": 1.5935815708944574, "leak": 0.10321916826069355, "replacement": 2.158887803554535}
+{"step": 105, "steps": 500, "lr": 2e-05, "grad_norm": 6.299375057220459, "loss_scale": 4.0, "retries": 0, "seconds": 90.9, "max_memory_gb": 16.44, "pair": 0.8787284344434738, "sign": 0.4473239630460739, "keep": 4.529021114110947, "leak": 0.7363333740504459, "replacement": 3.550984352827072}
+{"step": 106, "steps": 500, "lr": 2e-05, "grad_norm": 7.195115089416504, "loss_scale": 4.0, "retries": 0, "seconds": 91.8, "max_memory_gb": 16.44, "pair": 0.8563813716173172, "sign": 0.6465325579047203, "keep": 4.1347169280052185, "leak": 0.6541348157043103, "replacement": 3.4105398058891296}
+{"step": 107, "steps": 500, "lr": 2e-05, "grad_norm": 2.745300769805908, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 16.44, "pair": 0.9027737826108932, "sign": 0.7970392256975174, "keep": 1.6810288252308965, "leak": 0.10293432581238449, "replacement": 2.1675414443016052}
+{"step": 108, "steps": 500, "lr": 2e-05, "grad_norm": 2.639904499053955, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 16.44, "pair": 0.8288774639368057, "sign": 0.724130742251873, "keep": 1.6768596195615828, "leak": 0.34004200001072604, "replacement": 2.1143830865621567}
+{"step": 109, "steps": 500, "lr": 2e-05, "grad_norm": 6.8197150230407715, "loss_scale": 4.0, "retries": 0, "seconds": 94.4, "max_memory_gb": 16.44, "pair": 0.7987875491380692, "sign": 0.48661655001342297, "keep": 4.442910850048065, "leak": 0.6517067048553145, "replacement": 3.4264779686927795}
+{"step": 110, "steps": 500, "lr": 2e-05, "grad_norm": 5.166257858276367, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 16.44, "pair": 0.89018115401268, "sign": 0.5647151470184326, "keep": 2.913014680147171, "leak": 0.0717624891358355, "replacement": 2.6469866931438446}
+{"step": 111, "steps": 500, "lr": 2e-05, "grad_norm": 4.9205780029296875, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 16.44, "pair": 0.7885535359382629, "sign": 0.4531057868152857, "keep": 2.62522201705724, "leak": 0.11704093217849731, "replacement": 2.356977641582489}
+{"step": 112, "steps": 500, "lr": 2e-05, "grad_norm": 5.39555025100708, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 16.44, "pair": 0.8445666432380676, "sign": 0.4918728806078434, "keep": 2.6490832404233515, "leak": 0.09805195243097842, "replacement": 2.4395576417446136}
+{"step": 113, "steps": 500, "lr": 2e-05, "grad_norm": 5.658204078674316, "loss_scale": 4.0, "retries": 0, "seconds": 97.9, "max_memory_gb": 16.44, "pair": 0.8669785410165787, "sign": 0.5777434706687927, "keep": 2.9885166566818953, "leak": 0.09975245993518911, "replacement": 2.6750467121601105}
+{"step": 114, "steps": 500, "lr": 2e-05, "grad_norm": 7.556042671203613, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 16.44, "pair": 0.8731259852647781, "sign": 0.5755693167448044, "keep": 3.957137167453766, "leak": 0.5275962365267333, "replacement": 3.2713783383369446}
+{"step": 115, "steps": 500, "lr": 2e-05, "grad_norm": 4.969389915466309, "loss_scale": 4.0, "retries": 0, "seconds": 99.6, "max_memory_gb": 16.44, "pair": 0.8541867583990097, "sign": 0.5989254228770733, "keep": 2.7242382019758224, "leak": 0.6319557975341468, "replacement": 2.673757493495941}
+{"step": 116, "steps": 500, "lr": 2e-05, "grad_norm": 1.0666468143463135, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 16.44, "pair": 0.8849731832742691, "sign": 0.6916357725858688, "keep": 0.22213004808872938, "leak": 0.07688942208960725, "replacement": 1.3610784709453583}
+{"step": 117, "steps": 500, "lr": 2e-05, "grad_norm": 5.246176242828369, "loss_scale": 4.0, "retries": 0, "seconds": 101.4, "max_memory_gb": 16.44, "pair": 0.7793736308813095, "sign": 0.45693011581897736, "keep": 2.8579207592410967, "leak": 0.40340387923424714, "replacement": 2.537650093436241}
+{"step": 118, "steps": 500, "lr": 2e-05, "grad_norm": 3.254592180252075, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 16.44, "pair": 0.7494874745607376, "sign": 0.5218506176024675, "keep": 1.3631828115321696, "leak": 0.1343415752066086, "replacement": 1.7255895882844925}
+{"step": 119, "steps": 500, "lr": 2e-05, "grad_norm": 3.878110885620117, "loss_scale": 4.0, "retries": 0, "seconds": 103.1, "max_memory_gb": 16.44, "pair": 0.8508853912353516, "sign": 0.8048405051231384, "keep": 1.0365921705961227, "leak": 0.26485279974804143, "replacement": 1.8378149271011353}
+{"step": 120, "steps": 500, "lr": 2e-05, "grad_norm": 5.284169673919678, "loss_scale": 4.0, "retries": 0, "seconds": 104.0, "max_memory_gb": 16.44, "pair": 0.7645795196294785, "sign": 0.5282460823655128, "keep": 2.40704208239913, "leak": 0.39601023064460605, "replacement": 2.33122618496418}
+{"step": 121, "steps": 500, "lr": 2e-05, "grad_norm": 2.796907663345337, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 16.44, "pair": 0.857053354382515, "sign": 0.7729605883359909, "keep": 1.6963177151046693, "leak": 0.8554274663329124, "replacement": 2.305549442768097}
+{"step": 122, "steps": 500, "lr": 2e-05, "grad_norm": 5.193599700927734, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 16.44, "pair": 0.7976123690605164, "sign": 0.6969281658530235, "keep": 1.9025968582136557, "leak": 0.3363385508127976, "replacement": 2.1814595460891724}
+{"step": 123, "steps": 500, "lr": 2e-05, "grad_norm": 2.94465970993042, "loss_scale": 4.0, "retries": 0, "seconds": 106.6, "max_memory_gb": 16.44, "pair": 0.8999342024326324, "sign": 0.7937676450237632, "keep": 1.9930501841008663, "leak": 0.07196067649056204, "replacement": 2.311333268880844}
+{"step": 124, "steps": 500, "lr": 2e-05, "grad_norm": 3.761096239089966, "loss_scale": 4.0, "retries": 0, "seconds": 107.5, "max_memory_gb": 16.44, "pair": 0.9773416221141815, "sign": 1.0622818693518639, "keep": 0.9345013685524464, "leak": 0.01860552979633212, "replacement": 1.9803846776485443}
+{"step": 125, "steps": 500, "lr": 2e-05, "grad_norm": 3.128547191619873, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 16.44, "pair": 0.8979631811380386, "sign": 0.8186189979314804, "keep": 1.1086232247762382, "leak": 0.06938207597704604, "replacement": 1.8789298236370087}
+{"step": 126, "steps": 500, "lr": 2e-05, "grad_norm": 5.602852821350098, "loss_scale": 4.0, "retries": 0, "seconds": 109.2, "max_memory_gb": 16.44, "pair": 0.7901611775159836, "sign": 0.6165797975845635, "keep": 2.4721966832876205, "leak": 0.06116520391515223, "replacement": 2.3498407900333405}
+{"step": 127, "steps": 500, "lr": 2e-05, "grad_norm": 3.30596923828125, "loss_scale": 4.0, "retries": 0, "seconds": 110.0, "max_memory_gb": 16.44, "pair": 0.6746827960014343, "sign": 0.36969945579767227, "keep": 0.6873320343438536, "leak": 0.33636425342410803, "replacement": 1.287289634346962}
+{"step": 128, "steps": 500, "lr": 2e-05, "grad_norm": 5.854991436004639, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 16.44, "pair": 0.6310852766036987, "sign": 0.3308988041244447, "keep": 1.612224094569683, "leak": 0.6972645097412169, "replacement": 1.7769628465175629}
+{"step": 129, "steps": 500, "lr": 2e-05, "grad_norm": 9.854991912841797, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 16.44, "pair": 0.6161549761891365, "sign": 0.32081906124949455, "keep": 2.9837157130241394, "leak": 1.0222717279102653, "replacement": 2.523990273475647}
+{"step": 130, "steps": 500, "lr": 2e-05, "grad_norm": 5.672709941864014, "loss_scale": 4.0, "retries": 0, "seconds": 112.6, "max_memory_gb": 16.44, "pair": 0.8403869271278381, "sign": 0.8893331680446863, "keep": 2.001325882039964, "leak": 0.27206087624654174, "replacement": 2.353731781244278}
+{"step": 131, "steps": 500, "lr": 2e-05, "grad_norm": 3.219364643096924, "loss_scale": 4.0, "retries": 0, "seconds": 113.4, "max_memory_gb": 16.44, "pair": 0.8541665971279144, "sign": 0.718796968460083, "keep": 1.0605981647968292, "leak": 0.13752451524487697, "replacement": 1.7782452702522278}
+{"step": 132, "steps": 500, "lr": 2e-05, "grad_norm": 7.820533275604248, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 16.44, "pair": 0.5496328100562096, "sign": 0.1449113953858614, "keep": 2.669587210053578, "leak": 0.2299941424280405, "replacement": 2.014380693435669}
+{"step": 133, "steps": 500, "lr": 2e-05, "grad_norm": 8.076445579528809, "loss_scale": 4.0, "retries": 0, "seconds": 115.2, "max_memory_gb": 16.44, "pair": 0.7642693817615509, "sign": 0.4856237806379795, "keep": 2.4500629752874374, "leak": 0.975906679435866, "replacement": 2.4760894179344177}
+{"step": 134, "steps": 500, "lr": 2e-05, "grad_norm": 0.8914020657539368, "loss_scale": 4.0, "retries": 0, "seconds": 116.0, "max_memory_gb": 16.44, "pair": 0.7479384392499924, "sign": 0.6853315234184265, "keep": 0.3482483020052314, "leak": 0.33864912658464164, "replacement": 1.3493906557559967}
+{"step": 135, "steps": 500, "lr": 2e-05, "grad_norm": 7.63167667388916, "loss_scale": 4.0, "retries": 0, "seconds": 116.9, "max_memory_gb": 16.44, "pair": 0.6743720769882202, "sign": 0.5113974399864674, "keep": 2.0339203774929047, "leak": 1.463760127837304, "replacement": 2.3129710257053375}
+{"step": 136, "steps": 500, "lr": 2e-05, "grad_norm": 8.25164794921875, "loss_scale": 4.0, "retries": 0, "seconds": 117.7, "max_memory_gb": 16.44, "pair": 0.7079196572303772, "sign": 0.47871727123856544, "keep": 1.214553757570684, "leak": 0.18936449952889234, "replacement": 1.6018962860107422}
+{"step": 137, "steps": 500, "lr": 2e-05, "grad_norm": 5.145836353302002, "loss_scale": 4.0, "retries": 0, "seconds": 118.6, "max_memory_gb": 16.44, "pair": 0.6411071270704269, "sign": 0.45865469612181187, "keep": 0.9758891053497791, "leak": 1.9727958276162099, "replacement": 1.8515779972076416}
+{"step": 138, "steps": 500, "lr": 2e-05, "grad_norm": 5.713685035705566, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 16.44, "pair": 0.6621807813644409, "sign": 0.31343637965619564, "keep": 1.2478142911568284, "leak": 0.1176141791511327, "replacement": 1.4722096621990204}
+{"step": 139, "steps": 500, "lr": 2e-05, "grad_norm": 7.608392238616943, "loss_scale": 4.0, "retries": 0, "seconds": 120.4, "max_memory_gb": 16.44, "pair": 0.699591189622879, "sign": 0.6609364673495293, "keep": 2.0276250690221786, "leak": 0.6484686559997499, "replacement": 2.20598903298378}
+{"step": 140, "steps": 500, "lr": 2e-05, "grad_norm": 6.038827419281006, "loss_scale": 4.0, "retries": 0, "seconds": 121.2, "max_memory_gb": 16.44, "pair": 0.6462374180555344, "sign": 0.532699316740036, "keep": 0.7299537919461727, "leak": 0.6496315188705921, "replacement": 1.4399718046188354}
+{"step": 141, "steps": 500, "lr": 2e-05, "grad_norm": 4.795385837554932, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 16.44, "pair": 0.6804669052362442, "sign": 0.572740638628602, "keep": 0.7753933556377888, "leak": 1.414232888026163, "replacement": 1.7080921679735184}
+{"step": 142, "steps": 500, "lr": 2e-05, "grad_norm": 5.2288336753845215, "loss_scale": 4.0, "retries": 0, "seconds": 123.0, "max_memory_gb": 16.44, "pair": 0.5984868556261063, "sign": 0.4101770520210266, "keep": 0.800840689567849, "leak": 0.270825067149417, "replacement": 1.2717019766569138}
+{"step": 143, "steps": 500, "lr": 2e-05, "grad_norm": 0.9600290060043335, "loss_scale": 4.0, "retries": 0, "seconds": 123.9, "max_memory_gb": 16.44, "pair": 0.8223887979984283, "sign": 1.0310754477977753, "keep": 0.241022517089732, "leak": 0.3271784421885968, "replacement": 1.5402323752641678}
+{"step": 144, "steps": 500, "lr": 2e-05, "grad_norm": 6.759958267211914, "loss_scale": 4.0, "retries": 0, "seconds": 124.8, "max_memory_gb": 16.44, "pair": 0.6978007256984711, "sign": 0.4409458115696907, "keep": 1.6984382681548595, "leak": 0.2585934577509761, "replacement": 1.8321411311626434}
+{"step": 145, "steps": 500, "lr": 2e-05, "grad_norm": 5.38758659362793, "loss_scale": 4.0, "retries": 0, "seconds": 125.6, "max_memory_gb": 16.44, "pair": 0.5134467035531998, "sign": 0.2866130890324712, "keep": 0.7869080586824566, "leak": 0.4996986836194992, "replacement": 1.1751319617033005}
+{"step": 146, "steps": 500, "lr": 2e-05, "grad_norm": 4.518143177032471, "loss_scale": 4.0, "retries": 0, "seconds": 126.5, "max_memory_gb": 16.44, "pair": 0.6357409656047821, "sign": 0.5553853791207075, "keep": 0.6357347024604678, "leak": 0.5500259982654825, "replacement": 1.3688075244426727}
+{"step": 147, "steps": 500, "lr": 2e-05, "grad_norm": 4.770992279052734, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 16.44, "pair": 0.6858667284250259, "sign": 0.5411868430674076, "keep": 0.9365731091238558, "leak": 0.34179672796744853, "replacement": 1.5101958364248276}
+{"step": 148, "steps": 500, "lr": 2e-05, "grad_norm": 3.0156755447387695, "loss_scale": 4.0, "retries": 0, "seconds": 128.2, "max_memory_gb": 16.44, "pair": 0.5853211730718613, "sign": 0.31494075059890747, "keep": 0.3022127453587018, "leak": 0.3337246812880039, "replacement": 0.9773290902376175}
+{"step": 149, "steps": 500, "lr": 2e-05, "grad_norm": 4.1975250244140625, "loss_scale": 4.0, "retries": 0, "seconds": 129.1, "max_memory_gb": 16.44, "pair": 0.6345705315470695, "sign": 0.4970891997218132, "keep": 0.4358023083768785, "leak": 0.3095943396911025, "replacement": 1.1784148514270782}
+{"step": 150, "steps": 500, "lr": 2e-05, "grad_norm": 5.518866539001465, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 16.44, "pair": 0.6345366612076759, "sign": 0.3846636824309826, "keep": 0.9931754171848297, "leak": 1.3303185959812254, "replacement": 1.6560359001159668}
+{"step": 151, "steps": 500, "lr": 2e-05, "grad_norm": 3.409515380859375, "loss_scale": 4.0, "retries": 0, "seconds": 130.9, "max_memory_gb": 16.44, "pair": 0.6384985595941544, "sign": 0.4630066640675068, "keep": 0.5781574146822095, "leak": 0.2755645103752613, "replacement": 1.2279717326164246}
+{"step": 152, "steps": 500, "lr": 2e-05, "grad_norm": 4.675472736358643, "loss_scale": 4.0, "retries": 0, "seconds": 131.7, "max_memory_gb": 16.44, "pair": 0.4921778216958046, "sign": 0.2984796557575464, "keep": 0.4248827211558819, "leak": 1.6406583115458488, "replacement": 1.2640236169099808}
+{"step": 153, "steps": 500, "lr": 2e-05, "grad_norm": 4.3086090087890625, "loss_scale": 4.0, "retries": 0, "seconds": 132.6, "max_memory_gb": 16.44, "pair": 0.5130505412817001, "sign": 0.2931598909199238, "keep": 0.43622837495058775, "leak": 1.5186037048697472, "replacement": 1.2573956400156021}
+{"step": 154, "steps": 500, "lr": 2e-05, "grad_norm": 5.9396443367004395, "loss_scale": 4.0, "retries": 0, "seconds": 133.5, "max_memory_gb": 16.44, "pair": 0.6388941556215286, "sign": 0.27948281541466713, "keep": 1.1528745219111443, "leak": 0.740264717489481, "replacement": 1.5401389300823212}
+{"step": 155, "steps": 500, "lr": 2e-05, "grad_norm": 4.880379676818848, "loss_scale": 4.0, "retries": 0, "seconds": 134.4, "max_memory_gb": 16.44, "pair": 0.6123149544000626, "sign": 0.6341833062469959, "keep": 1.7213970124721527, "leak": 1.2293000258505344, "replacement": 2.0974300652742386}
+{"step": 156, "steps": 500, "lr": 2e-05, "grad_norm": 2.1816115379333496, "loss_scale": 4.0, "retries": 0, "seconds": 135.2, "max_memory_gb": 16.44, "pair": 0.6146171018481255, "sign": 0.5364450626075268, "keep": 0.32823666010517627, "leak": 0.525681224418804, "replacement": 1.178378239274025}
+{"step": 157, "steps": 500, "lr": 2e-05, "grad_norm": 3.762197732925415, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 16.44, "pair": 0.8494693338871002, "sign": 1.2441419064998627, "keep": 0.4298589136451483, "leak": 0.3023846347350627, "replacement": 1.762065902352333}
+{"step": 158, "steps": 500, "lr": 2e-05, "grad_norm": 1.747517466545105, "loss_scale": 4.0, "retries": 0, "seconds": 136.9, "max_memory_gb": 16.44, "pair": 0.7582496553659439, "sign": 0.6987929195165634, "keep": 0.5214593280106783, "leak": 0.18237988877808675, "replacement": 1.4139707535505295}
+{"step": 159, "steps": 500, "lr": 2e-05, "grad_norm": 4.769433498382568, "loss_scale": 4.0, "retries": 0, "seconds": 137.8, "max_memory_gb": 16.44, "pair": 0.42547114193439484, "sign": 0.19994275271892548, "keep": 0.17693164898082614, "leak": 2.025808684527874, "replacement": 1.1203604638576508}
+{"step": 160, "steps": 500, "lr": 2e-05, "grad_norm": 5.924136161804199, "loss_scale": 4.0, "retries": 0, "seconds": 138.6, "max_memory_gb": 16.44, "pair": 0.4521840661764145, "sign": 0.3716352393385023, "keep": 0.329290971159935, "leak": 2.1027364729816327, "replacement": 1.3283312767744064}
+{"step": 161, "steps": 500, "lr": 2e-05, "grad_norm": 1.7820405960083008, "loss_scale": 4.0, "retries": 0, "seconds": 139.5, "max_memory_gb": 16.44, "pair": 0.6440275087952614, "sign": 0.592509638518095, "keep": 0.4078236520290375, "leak": 0.34059723852260504, "replacement": 1.2293434143066406}
+{"step": 162, "steps": 500, "lr": 2e-05, "grad_norm": 1.8469107151031494, "loss_scale": 4.0, "retries": 0, "seconds": 140.3, "max_memory_gb": 16.44, "pair": 0.6549683213233948, "sign": 0.600731585174799, "keep": 0.13831950351595879, "leak": 0.2821156978607178, "replacement": 1.0950228422880173}
+{"step": 163, "steps": 500, "lr": 2e-05, "grad_norm": 5.297004699707031, "loss_scale": 4.0, "retries": 0, "seconds": 141.1, "max_memory_gb": 16.44, "pair": 0.4864286109805107, "sign": 0.35027616657316685, "keep": 0.2952503617852926, "leak": 0.8424264648929238, "replacement": 1.0197985023260117}
+{"step": 164, "steps": 500, "lr": 2e-05, "grad_norm": 3.9266693592071533, "loss_scale": 4.0, "retries": 0, "seconds": 142.0, "max_memory_gb": 16.44, "pair": 0.7474963739514351, "sign": 0.8804974798113108, "keep": 0.627664253115654, "leak": 0.508175291299267, "replacement": 1.6286210715770721}
+{"step": 165, "steps": 500, "lr": 2e-05, "grad_norm": 4.077134609222412, "loss_scale": 4.0, "retries": 0, "seconds": 142.9, "max_memory_gb": 16.44, "pair": 0.7069617211818695, "sign": 0.651210367679596, "keep": 0.2809567851945758, "leak": 1.1568474974483252, "replacement": 1.462257206439972}
+{"step": 166, "steps": 500, "lr": 2e-05, "grad_norm": 3.2467222213745117, "loss_scale": 4.0, "retries": 0, "seconds": 143.7, "max_memory_gb": 16.44, "pair": 0.6428029984235764, "sign": 0.5629064440727234, "keep": 0.3156161792576313, "leak": 0.9099231684813276, "replacement": 1.3095451295375824}
+{"step": 167, "steps": 500, "lr": 2e-05, "grad_norm": 2.9524130821228027, "loss_scale": 4.0, "retries": 0, "seconds": 144.6, "max_memory_gb": 16.44, "pair": 0.5003189444541931, "sign": 0.4187511960044503, "keep": 0.6055940762162209, "leak": 1.0423717519443016, "replacement": 1.2730845510959625}
+{"step": 168, "steps": 500, "lr": 2e-05, "grad_norm": 5.83207368850708, "loss_scale": 4.0, "retries": 0, "seconds": 145.4, "max_memory_gb": 16.44, "pair": 0.5679522231221199, "sign": 0.5073199234902859, "keep": 0.16560635715723038, "leak": 1.454246461391449, "replacement": 1.267976999282837}
+{"step": 169, "steps": 500, "lr": 2e-05, "grad_norm": 5.029836177825928, "loss_scale": 4.0, "retries": 0, "seconds": 146.3, "max_memory_gb": 16.44, "pair": 0.6096451506018639, "sign": 0.48248269967734814, "keep": 0.6498938445001841, "leak": 1.5898254653438926, "replacement": 1.5732897371053696}
+{"step": 170, "steps": 500, "lr": 2e-05, "grad_norm": 4.037718296051025, "loss_scale": 4.0, "retries": 0, "seconds": 147.1, "max_memory_gb": 16.44, "pair": 0.32950425148010254, "sign": 0.15212522819638252, "keep": 0.06313850171864033, "leak": 1.2405685782432556, "replacement": 0.7472782582044601}
+{"step": 171, "steps": 500, "lr": 2e-05, "grad_norm": 4.658416748046875, "loss_scale": 4.0, "retries": 0, "seconds": 148.0, "max_memory_gb": 16.44, "pair": 0.5506060123443604, "sign": 0.40755344182252884, "keep": 0.10063681984320283, "leak": 1.6829236820340157, "replacement": 1.22543203830719}
+{"step": 172, "steps": 500, "lr": 2e-05, "grad_norm": 6.364818572998047, "loss_scale": 4.0, "retries": 0, "seconds": 148.9, "max_memory_gb": 16.44, "pair": 0.31209345906972885, "sign": 0.1467484701424837, "keep": 0.2881716233678162, "leak": 2.865870736539364, "replacement": 1.2460211664438248}
+{"step": 173, "steps": 500, "lr": 2e-05, "grad_norm": 4.926393985748291, "loss_scale": 4.0, "retries": 0, "seconds": 149.7, "max_memory_gb": 16.44, "pair": 0.546630285680294, "sign": 0.4132622415199876, "keep": 0.33405905589461327, "leak": 1.0232398845255375, "replacement": 1.1761009395122528}
+{"step": 174, "steps": 500, "lr": 2e-05, "grad_norm": 1.9427306652069092, "loss_scale": 4.0, "retries": 0, "seconds": 150.5, "max_memory_gb": 16.44, "pair": 0.42630667984485626, "sign": 0.3113912306725979, "keep": 0.34591666888445616, "leak": 0.820679560303688, "replacement": 0.960130512714386}
+{"step": 175, "steps": 500, "lr": 2e-05, "grad_norm": 3.5461177825927734, "loss_scale": 4.0, "retries": 0, "seconds": 151.4, "max_memory_gb": 16.44, "pair": 0.5963792130351067, "sign": 0.5940928906202316, "keep": 0.3219367954879999, "leak": 0.9179364424198866, "replacement": 1.2838781774044037}
+{"step": 176, "steps": 500, "lr": 2e-05, "grad_norm": 6.230859756469727, "loss_scale": 4.0, "retries": 0, "seconds": 152.2, "max_memory_gb": 16.44, "pair": 0.37018348276615143, "sign": 0.2189657725393772, "keep": 0.09584457240998745, "leak": 2.0673897862434387, "replacement": 1.044436126947403}
+{"step": 177, "steps": 500, "lr": 2e-05, "grad_norm": 2.8113157749176025, "loss_scale": 4.0, "retries": 0, "seconds": 153.1, "max_memory_gb": 16.44, "pair": 0.4077443368732929, "sign": 0.323786080814898, "keep": 0.3035055585205555, "leak": 0.6012878045148682, "replacement": 0.8717120960354805}
+{"step": 178, "steps": 500, "lr": 2e-05, "grad_norm": 3.44006609916687, "loss_scale": 4.0, "retries": 0, "seconds": 153.9, "max_memory_gb": 16.44, "pair": 0.6165120229125023, "sign": 0.5802096426486969, "keep": 0.3642897675745189, "leak": 1.2615729048848152, "replacement": 1.4041549563407898}
+{"step": 179, "steps": 500, "lr": 2e-05, "grad_norm": 2.4286372661590576, "loss_scale": 4.0, "retries": 0, "seconds": 154.7, "max_memory_gb": 16.44, "pair": 0.6945014446973801, "sign": 0.7555224783718586, "keep": 0.5945802591741085, "leak": 0.5991628143692651, "replacement": 1.51934352517128}
+{"step": 180, "steps": 500, "lr": 2e-05, "grad_norm": 2.871506690979004, "loss_scale": 4.0, "retries": 0, "seconds": 155.6, "max_memory_gb": 16.44, "pair": 0.5412347167730331, "sign": 0.4854494873434305, "keep": 0.5751776862889528, "leak": 0.7319105537608266, "replacement": 1.2545259147882462}
+{"step": 181, "steps": 500, "lr": 2e-05, "grad_norm": 5.4499897956848145, "loss_scale": 4.0, "retries": 0, "seconds": 156.4, "max_memory_gb": 16.44, "pair": 0.3156130425632, "sign": 0.19796289689838886, "keep": 0.13298972230404615, "leak": 1.8264106214046478, "replacement": 0.9376920163631439}
+{"step": 182, "steps": 500, "lr": 2e-05, "grad_norm": 2.9016456604003906, "loss_scale": 4.0, "retries": 0, "seconds": 157.3, "max_memory_gb": 16.44, "pair": 0.6544021889567375, "sign": 0.5844766590744257, "keep": 0.29837864410365, "leak": 0.7581400430062786, "replacement": 1.285364881157875}
+{"step": 183, "steps": 500, "lr": 2e-05, "grad_norm": 5.370334625244141, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 16.44, "pair": 0.2657626047730446, "sign": 0.14452834846451879, "keep": 0.028758768981788307, "leak": 1.3372567147016525, "replacement": 0.6867203265428543}
+{"step": 184, "steps": 500, "lr": 2e-05, "grad_norm": 5.830446720123291, "loss_scale": 4.0, "retries": 0, "seconds": 159.0, "max_memory_gb": 16.44, "pair": 0.4577190764248371, "sign": 0.3216586662456393, "keep": 0.138050434703473, "leak": 1.2771571520715952, "replacement": 1.0068629086017609}
+{"step": 185, "steps": 500, "lr": 2e-05, "grad_norm": 6.2094502449035645, "loss_scale": 4.0, "retries": 0, "seconds": 159.9, "max_memory_gb": 16.44, "pair": 0.3488446734845638, "sign": 0.2762049725279212, "keep": 0.08553262706845999, "leak": 2.5184871703386307, "replacement": 1.1593352407217026}
+{"step": 186, "steps": 500, "lr": 2e-05, "grad_norm": 3.1767654418945312, "loss_scale": 4.0, "retries": 0, "seconds": 160.7, "max_memory_gb": 16.44, "pair": 0.39216819405555725, "sign": 0.23090195935219526, "keep": 0.1276825056411326, "leak": 0.7797680124640465, "replacement": 0.7664024531841278}
+{"step": 187, "steps": 500, "lr": 2e-05, "grad_norm": 4.40224027633667, "loss_scale": 4.0, "retries": 0, "seconds": 161.6, "max_memory_gb": 16.44, "pair": 0.383652962744236, "sign": 0.10748206730931997, "keep": 0.1611008383333683, "leak": 0.9708368605934083, "replacement": 0.7606536224484444}
+{"step": 188, "steps": 500, "lr": 2e-05, "grad_norm": 0.7587531208992004, "loss_scale": 4.0, "retries": 0, "seconds": 162.4, "max_memory_gb": 16.44, "pair": 0.6009704917669296, "sign": 0.6109632030129433, "keep": 0.24827906655264087, "leak": 0.5476729974616319, "replacement": 1.1675098836421967}
+{"step": 189, "steps": 500, "lr": 2e-05, "grad_norm": 5.017818927764893, "loss_scale": 4.0, "retries": 0, "seconds": 163.3, "max_memory_gb": 16.44, "pair": 0.34367571771144867, "sign": 0.23759866878390312, "keep": 0.04854765208438039, "leak": 1.4350677132606506, "replacement": 0.8455158025026321}
+{"step": 190, "steps": 500, "lr": 2e-05, "grad_norm": 4.4055585861206055, "loss_scale": 4.0, "retries": 0, "seconds": 164.1, "max_memory_gb": 16.44, "pair": 0.4425739496946335, "sign": 0.20883291959762573, "keep": 0.0819737920537591, "leak": 0.8102386221289635, "replacement": 0.7905369400978088}
+{"step": 191, "steps": 500, "lr": 2e-05, "grad_norm": 2.2253289222717285, "loss_scale": 4.0, "retries": 0, "seconds": 164.9, "max_memory_gb": 16.44, "pair": 0.6165505461394787, "sign": 0.6000329125672579, "keep": 0.608174511231482, "leak": 0.38434940576553345, "replacement": 1.31674163043499}
+{"step": 192, "steps": 500, "lr": 2e-05, "grad_norm": 5.44843864440918, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 16.44, "pair": 0.5032356977462769, "sign": 0.4903930742293596, "keep": 0.4183294821996242, "leak": 1.3228911611173686, "replacement": 1.2883197665214539}
+{"step": 193, "steps": 500, "lr": 2e-05, "grad_norm": 3.2905921936035156, "loss_scale": 4.0, "retries": 0, "seconds": 166.6, "max_memory_gb": 16.44, "pair": 0.34274446964263916, "sign": 0.18332486413419247, "keep": 0.08181480405619368, "leak": 1.0036681592464447, "replacement": 0.7262313365936279}
+{"step": 194, "steps": 500, "lr": 2e-05, "grad_norm": 2.405794858932495, "loss_scale": 4.0, "retries": 0, "seconds": 167.4, "max_memory_gb": 16.44, "pair": 0.7453307211399078, "sign": 0.7492558993399143, "keep": 0.5728969443589449, "leak": 0.4573941114358604, "replacement": 1.5207556337118149}
+{"step": 195, "steps": 500, "lr": 2e-05, "grad_norm": 6.427821159362793, "loss_scale": 4.0, "retries": 0, "seconds": 168.3, "max_memory_gb": 16.44, "pair": 0.45686914026737213, "sign": 0.3541890559718013, "keep": 0.18646779283881187, "leak": 1.7076556194533623, "replacement": 1.1541114449501038}
+{"step": 196, "steps": 500, "lr": 2e-05, "grad_norm": 9.199070930480957, "loss_scale": 4.0, "retries": 0, "seconds": 169.2, "max_memory_gb": 16.44, "pair": 0.21944055147469044, "sign": 0.11102891759946942, "keep": 0.08177814783994108, "leak": 2.467898115515709, "replacement": 0.9328186362981796}
+{"step": 197, "steps": 500, "lr": 2e-05, "grad_norm": 4.95756721496582, "loss_scale": 4.0, "retries": 0, "seconds": 170.0, "max_memory_gb": 16.44, "pair": 0.6027813479304314, "sign": 0.5644210167229176, "keep": 0.23001372627913952, "leak": 0.7315339054912329, "replacement": 1.1828821897506714}
+{"step": 198, "steps": 500, "lr": 2e-05, "grad_norm": 2.1609597206115723, "loss_scale": 4.0, "retries": 0, "seconds": 170.9, "max_memory_gb": 16.44, "pair": 0.6934157609939575, "sign": 0.6109026744961739, "keep": 0.36834363266825676, "leak": 0.32101493049412966, "replacement": 1.263292670249939}
+{"step": 199, "steps": 500, "lr": 2e-05, "grad_norm": 3.3680944442749023, "loss_scale": 4.0, "retries": 0, "seconds": 171.7, "max_memory_gb": 16.44, "pair": 0.3201354667544365, "sign": 0.2647211840376258, "keep": 0.2379215150140226, "leak": 0.9218340516090393, "replacement": 0.8019153326749802}
+{"step": 200, "steps": 500, "lr": 2e-05, "grad_norm": 1.882300853729248, "loss_scale": 4.0, "retries": 0, "seconds": 172.6, "max_memory_gb": 16.44, "pair": 0.6690745651721954, "sign": 0.7024422865360975, "keep": 0.48911156319081783, "leak": 0.4288431556051364, "replacement": 1.3720623254776}
+{"step": 201, "steps": 500, "lr": 2e-05, "grad_norm": 3.067859649658203, "loss_scale": 4.0, "retries": 0, "seconds": 173.5, "max_memory_gb": 16.44, "pair": 0.6276529505848885, "sign": 0.6718663964420557, "keep": 0.3359531043097377, "leak": 0.43382035568356514, "replacement": 1.2400178164243698}
+{"step": 202, "steps": 500, "lr": 2e-05, "grad_norm": 3.4405670166015625, "loss_scale": 4.0, "retries": 0, "seconds": 174.3, "max_memory_gb": 16.44, "pair": 0.36578840017318726, "sign": 0.4419696740806103, "keep": 0.005915671063121408, "leak": 0.8820693641901016, "replacement": 0.8102484196424484}
+{"step": 203, "steps": 500, "lr": 2e-05, "grad_norm": 1.945796012878418, "loss_scale": 4.0, "retries": 0, "seconds": 175.1, "max_memory_gb": 16.44, "pair": 0.5269232057034969, "sign": 0.5512908287346363, "keep": 0.3324316944926977, "leak": 0.4157782872207463, "replacement": 1.0727291107177734}
+{"step": 204, "steps": 500, "lr": 2e-05, "grad_norm": 4.586785316467285, "loss_scale": 4.0, "retries": 0, "seconds": 176.0, "max_memory_gb": 16.44, "pair": 0.5042877644300461, "sign": 0.5999881383031607, "keep": 0.3098499539773911, "leak": 0.37754612043499947, "replacement": 1.053593322634697}
+{"step": 205, "steps": 500, "lr": 2e-05, "grad_norm": 2.387552499771118, "loss_scale": 4.0, "retries": 0, "seconds": 176.8, "max_memory_gb": 16.44, "pair": 0.47488611564040184, "sign": 0.6120358258485794, "keep": 0.20256164018064737, "leak": 0.6393136553524528, "replacement": 1.042013257741928}
+{"step": 206, "steps": 500, "lr": 2e-05, "grad_norm": 4.657072067260742, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 16.44, "pair": 0.4626941867172718, "sign": 0.6461641564965248, "keep": 0.38017630265676416, "leak": 1.0951045975089073, "replacement": 1.2496405094861984}
+{"step": 207, "steps": 500, "lr": 2e-05, "grad_norm": 2.2732200622558594, "loss_scale": 4.0, "retries": 0, "seconds": 178.5, "max_memory_gb": 16.44, "pair": 0.8413616642355919, "sign": 0.9866049513220787, "keep": 1.6336971407290548, "leak": 0.27520501511753537, "replacement": 2.220313936471939}
+{"step": 208, "steps": 500, "lr": 2e-05, "grad_norm": 7.204482555389404, "loss_scale": 4.0, "retries": 0, "seconds": 179.4, "max_memory_gb": 16.44, "pair": 0.1887711137533188, "sign": 0.10284154396504164, "keep": 0.1242867173277773, "leak": 1.563491776585579, "replacement": 0.6932081878185272}
+{"step": 209, "steps": 500, "lr": 2e-05, "grad_norm": 3.7301576137542725, "loss_scale": 4.0, "retries": 0, "seconds": 180.3, "max_memory_gb": 16.44, "pair": 0.6382834762334824, "sign": 0.637941999360919, "keep": 0.3734787357971072, "leak": 0.9873356372117996, "replacement": 1.3908277451992035}
+{"step": 210, "steps": 500, "lr": 2e-05, "grad_norm": 3.19437313079834, "loss_scale": 4.0, "retries": 0, "seconds": 181.1, "max_memory_gb": 16.44, "pair": 0.5035631060600281, "sign": 0.5285466630011797, "keep": 0.49745852837804705, "leak": 1.1419491665437818, "replacement": 1.3020529448986053}
+{"step": 211, "steps": 500, "lr": 2e-05, "grad_norm": 4.86866569519043, "loss_scale": 4.0, "retries": 0, "seconds": 182.0, "max_memory_gb": 16.44, "pair": 0.5228390954434872, "sign": 0.5245327381417155, "keep": 0.28139954898506403, "leak": 0.874553571746219, "replacement": 1.1444435864686966}
+{"step": 212, "steps": 500, "lr": 2e-05, "grad_norm": 4.032383918762207, "loss_scale": 4.0, "retries": 0, "seconds": 182.9, "max_memory_gb": 16.44, "pair": 0.5940731391310692, "sign": 0.812434608116746, "keep": 0.6523231007158756, "leak": 0.34674644842743874, "replacement": 1.4131386429071426}
+{"step": 213, "steps": 500, "lr": 2e-05, "grad_norm": 7.107995510101318, "loss_scale": 4.0, "retries": 0, "seconds": 183.7, "max_memory_gb": 16.44, "pair": 0.4595755524933338, "sign": 0.4908801447600126, "keep": 0.6601669387891889, "leak": 1.8306514527648687, "replacement": 1.4927619695663452}
+{"step": 214, "steps": 500, "lr": 2e-05, "grad_norm": 7.533414840698242, "loss_scale": 4.0, "retries": 0, "seconds": 184.6, "max_memory_gb": 16.44, "pair": 0.6498492546379566, "sign": 0.7067964673042297, "keep": 1.504515386710409, "leak": 0.37385685791923606, "replacement": 1.848969355225563}
+{"step": 215, "steps": 500, "lr": 2e-05, "grad_norm": 1.5233361721038818, "loss_scale": 4.0, "retries": 0, "seconds": 185.5, "max_memory_gb": 16.44, "pair": 0.7341667041182518, "sign": 0.912076361477375, "keep": 0.9158060169720557, "leak": 0.5736232117051259, "replacement": 1.7915136516094208}
+{"step": 216, "steps": 500, "lr": 2e-05, "grad_norm": 1.9969658851623535, "loss_scale": 4.0, "retries": 0, "seconds": 186.3, "max_memory_gb": 16.44, "pair": 0.45972416177392006, "sign": 0.36574285849928856, "keep": 0.5289186472073197, "leak": 0.658333320170641, "replacement": 1.0716383159160614}
+{"step": 217, "steps": 500, "lr": 2e-05, "grad_norm": 5.272961616516113, "loss_scale": 4.0, "retries": 0, "seconds": 187.2, "max_memory_gb": 16.44, "pair": 0.30006884783506393, "sign": 0.24347738921642303, "keep": 0.19976276013767347, "leak": 1.5382753163576126, "replacement": 0.9062577486038208}
+{"step": 218, "steps": 500, "lr": 2e-05, "grad_norm": 1.9828917980194092, "loss_scale": 4.0, "retries": 0, "seconds": 188.0, "max_memory_gb": 16.44, "pair": 0.81008680164814, "sign": 1.0403794199228287, "keep": 0.30137934582307935, "leak": 0.31427238111791667, "replacement": 1.5595342814922333}
+{"step": 219, "steps": 500, "lr": 2e-05, "grad_norm": 2.2113935947418213, "loss_scale": 4.0, "retries": 0, "seconds": 188.9, "max_memory_gb": 16.44, "pair": 0.5630986765027046, "sign": 0.5883898809552193, "keep": 0.15897995489649475, "leak": 0.3879946591332555, "replacement": 1.033782258629799}
+{"step": 220, "steps": 500, "lr": 2e-05, "grad_norm": 3.4334962368011475, "loss_scale": 4.0, "retries": 0, "seconds": 189.7, "max_memory_gb": 16.44, "pair": 0.38174403458833694, "sign": 0.45580537617206573, "keep": 0.3119833071832545, "leak": 0.750606412650086, "replacement": 0.9532899707555771}
+{"step": 221, "steps": 500, "lr": 2e-05, "grad_norm": 4.834813117980957, "loss_scale": 4.0, "retries": 0, "seconds": 190.6, "max_memory_gb": 16.44, "pair": 0.5039298720657825, "sign": 0.26153003983199596, "keep": 0.426529036834836, "leak": 0.2624830477871001, "replacement": 0.9135802090167999}
+{"step": 222, "steps": 500, "lr": 2e-05, "grad_norm": 4.424219131469727, "loss_scale": 4.0, "retries": 0, "seconds": 191.5, "max_memory_gb": 16.44, "pair": 0.32736752182245255, "sign": 0.2073775315657258, "keep": 0.4086694326251745, "leak": 1.742611937224865, "replacement": 1.0710439831018448}
+{"step": 223, "steps": 500, "lr": 2e-05, "grad_norm": 1.2771345376968384, "loss_scale": 4.0, "retries": 0, "seconds": 192.3, "max_memory_gb": 16.44, "pair": 0.6132726967334747, "sign": 0.670538492500782, "keep": 0.38639729795977473, "leak": 0.3633041694993153, "replacement": 1.232566624879837}
+{"step": 224, "steps": 500, "lr": 2e-05, "grad_norm": 3.0395312309265137, "loss_scale": 4.0, "retries": 0, "seconds": 193.2, "max_memory_gb": 16.44, "pair": 0.53736512362957, "sign": 0.5434944983571768, "keep": 0.2510556107154116, "leak": 0.8494182238355279, "replacement": 1.1469947695732117}
+{"step": 225, "steps": 500, "lr": 2e-05, "grad_norm": 4.562410354614258, "loss_scale": 4.0, "retries": 0, "seconds": 194.0, "max_memory_gb": 16.44, "pair": 0.5678068101406097, "sign": 0.5423976574093103, "keep": 0.24290507100522518, "leak": 1.2550032210419886, "replacement": 1.2742090374231339}
+{"step": 226, "steps": 500, "lr": 2e-05, "grad_norm": 7.034514427185059, "loss_scale": 4.0, "retries": 0, "seconds": 194.9, "max_memory_gb": 16.44, "pair": 0.42473244294524193, "sign": 0.4295292068272829, "keep": 0.3430950930342078, "leak": 1.456248041242361, "replacement": 1.1751066595315933}
+{"step": 227, "steps": 500, "lr": 2e-05, "grad_norm": 3.0814740657806396, "loss_scale": 4.0, "retries": 0, "seconds": 195.7, "max_memory_gb": 16.44, "pair": 0.2844257801771164, "sign": 0.17584084533154964, "keep": 0.07176832411641954, "leak": 0.5800469592213631, "replacement": 0.5532421097159386}
+{"step": 228, "steps": 500, "lr": 2e-05, "grad_norm": 2.9891929626464844, "loss_scale": 4.0, "retries": 0, "seconds": 196.6, "max_memory_gb": 16.44, "pair": 0.521034725010395, "sign": 0.53495953977108, "keep": 0.15440122783184052, "leak": 0.7527526319026947, "replacement": 1.0539032518863678}
+{"step": 229, "steps": 500, "lr": 2e-05, "grad_norm": 4.535582065582275, "loss_scale": 4.0, "retries": 0, "seconds": 197.4, "max_memory_gb": 16.44, "pair": 0.5803261455148458, "sign": 0.6951914962846786, "keep": 0.5178745202720165, "leak": 0.98068568110466, "replacement": 1.4320305287837982}
+{"step": 230, "steps": 500, "lr": 2e-05, "grad_norm": 1.6100213527679443, "loss_scale": 4.0, "retries": 0, "seconds": 198.3, "max_memory_gb": 16.44, "pair": 0.6709595248103142, "sign": 0.7185028046369553, "keep": 0.6790556029009167, "leak": 0.31550309737212956, "replacement": 1.4486145228147507}
+{"step": 231, "steps": 500, "lr": 2e-05, "grad_norm": 3.0788590908050537, "loss_scale": 4.0, "retries": 0, "seconds": 199.1, "max_memory_gb": 16.44, "pair": 0.6760797426104546, "sign": 0.8068156838417053, "keep": 0.422985336263082, "leak": 0.515645666513592, "replacement": 1.4198917001485825}
+{"step": 232, "steps": 500, "lr": 2e-05, "grad_norm": 8.017643928527832, "loss_scale": 4.0, "retries": 0, "seconds": 199.9, "max_memory_gb": 16.44, "pair": 0.35452403873205185, "sign": 0.49744176492094994, "keep": 0.12522710673511028, "leak": 1.2324805110692978, "replacement": 0.9739786088466644}
+{"step": 233, "steps": 500, "lr": 2e-05, "grad_norm": 5.539547443389893, "loss_scale": 4.0, "retries": 0, "seconds": 200.8, "max_memory_gb": 16.44, "pair": 0.3372954726219177, "sign": 0.19532759301364422, "keep": 0.7197034507989883, "leak": 0.9169423878192902, "replacement": 1.0240465998649597}
+{"step": 234, "steps": 500, "lr": 2e-05, "grad_norm": 4.181748867034912, "loss_scale": 4.0, "retries": 0, "seconds": 201.6, "max_memory_gb": 16.44, "pair": 0.49861152470111847, "sign": 0.49046143889427185, "keep": 0.40331404539756477, "leak": 1.0733895171433687, "replacement": 1.2138466835021973}
+{"step": 235, "steps": 500, "lr": 2e-05, "grad_norm": 3.529066801071167, "loss_scale": 4.0, "retries": 0, "seconds": 202.5, "max_memory_gb": 16.44, "pair": 0.5847100988030434, "sign": 0.7064256630837917, "keep": 0.529312402009964, "leak": 0.5898358570411801, "replacement": 1.3500381112098694}
+{"step": 236, "steps": 500, "lr": 2e-05, "grad_norm": 2.974447727203369, "loss_scale": 4.0, "retries": 0, "seconds": 203.3, "max_memory_gb": 16.44, "pair": 0.5142882838845253, "sign": 0.484850961714983, "keep": 0.17312560911523178, "leak": 0.7597531585488468, "replacement": 1.0332148522138596}
+{"step": 237, "steps": 500, "lr": 2e-05, "grad_norm": 5.249929428100586, "loss_scale": 4.0, "retries": 0, "seconds": 204.1, "max_memory_gb": 16.44, "pair": 0.5138011425733566, "sign": 0.47942889109253883, "keep": 0.24987729206623044, "leak": 0.9274097308516502, "replacement": 1.1103066205978394}
+{"step": 238, "steps": 500, "lr": 2e-05, "grad_norm": 6.495767116546631, "loss_scale": 4.0, "retries": 0, "seconds": 205.0, "max_memory_gb": 16.44, "pair": 0.24249870143830776, "sign": 0.2686746926046908, "keep": 0.09124180488288403, "leak": 1.5952774733304977, "replacement": 0.8212763369083405}
+{"step": 239, "steps": 500, "lr": 2e-05, "grad_norm": 2.9603638648986816, "loss_scale": 4.0, "retries": 0, "seconds": 205.8, "max_memory_gb": 16.44, "pair": 0.6412568464875221, "sign": 0.7467797584831715, "keep": 0.5815268140286207, "leak": 0.51351104141213, "replacement": 1.4337879121303558}
+{"step": 240, "steps": 500, "lr": 2e-05, "grad_norm": 3.643312454223633, "loss_scale": 4.0, "retries": 0, "seconds": 206.6, "max_memory_gb": 16.44, "pair": 0.4049959182739258, "sign": 0.3742731362581253, "keep": 0.25765517563559115, "leak": 0.9866804033517838, "replacement": 0.9676301926374435}
+{"step": 241, "steps": 500, "lr": 2e-05, "grad_norm": 1.0642210245132446, "loss_scale": 4.0, "retries": 0, "seconds": 207.5, "max_memory_gb": 16.44, "pair": 0.6777878105640411, "sign": 0.7759710215032101, "keep": 0.3767340835183859, "leak": 0.27469197288155556, "replacement": 1.3228134214878082}
+{"step": 242, "steps": 500, "lr": 2e-05, "grad_norm": 1.6409138441085815, "loss_scale": 4.0, "retries": 0, "seconds": 208.3, "max_memory_gb": 16.44, "pair": 0.603585384786129, "sign": 0.6451210491359234, "keep": 0.3744248989096377, "leak": 0.39979978185147047, "replacement": 1.2133083641529083}
+{"step": 243, "steps": 500, "lr": 2e-05, "grad_norm": 3.656251907348633, "loss_scale": 4.0, "retries": 0, "seconds": 209.2, "max_memory_gb": 16.44, "pair": 0.28655569814145565, "sign": 0.395024798810482, "keep": 0.25253035221248865, "leak": 1.555568978190422, "replacement": 0.9992255419492722}
+{"step": 244, "steps": 500, "lr": 2e-05, "grad_norm": 4.072937965393066, "loss_scale": 4.0, "retries": 0, "seconds": 210.1, "max_memory_gb": 16.44, "pair": 0.6623104959726334, "sign": 0.7909516450017691, "keep": 0.7459855582565069, "leak": 0.5487414291128516, "replacement": 1.567964419722557}
+{"step": 245, "steps": 500, "lr": 2e-05, "grad_norm": 4.150590896606445, "loss_scale": 4.0, "retries": 0, "seconds": 210.9, "max_memory_gb": 16.44, "pair": 0.8216238468885422, "sign": 0.9743902422487736, "keep": 0.6437520096078515, "leak": 0.4088323999749264, "replacement": 1.732903003692627}
+{"step": 246, "steps": 500, "lr": 2e-05, "grad_norm": 4.991230010986328, "loss_scale": 4.0, "retries": 0, "seconds": 211.8, "max_memory_gb": 16.44, "pair": 0.26169921830296516, "sign": 0.21651773434132338, "keep": 0.23275287405704148, "leak": 1.279558002948761, "replacement": 0.8062240183353424}
+{"step": 247, "steps": 500, "lr": 2e-05, "grad_norm": 1.675137996673584, "loss_scale": 4.0, "retries": 0, "seconds": 212.7, "max_memory_gb": 16.44, "pair": 0.49504007399082184, "sign": 0.45334357023239136, "keep": 0.24664479942293838, "leak": 0.5718630847986788, "replacement": 0.9880000352859497}
+{"step": 248, "steps": 500, "lr": 2e-05, "grad_norm": 3.9683172702789307, "loss_scale": 4.0, "retries": 0, "seconds": 213.6, "max_memory_gb": 16.44, "pair": 0.3043305017054081, "sign": 0.2065002042800188, "keep": 0.16596372053027153, "leak": 0.7573116028506774, "replacement": 0.679890364408493}
+{"step": 249, "steps": 500, "lr": 2e-05, "grad_norm": 5.238026142120361, "loss_scale": 4.0, "retries": 0, "seconds": 214.5, "max_memory_gb": 16.44, "pair": 0.6478542536497116, "sign": 0.9638802707195282, "keep": 0.12062177434563637, "leak": 0.4164919632185047, "replacement": 1.2942282855510712}
+{"step": 250, "steps": 500, "lr": 2e-05, "grad_norm": 5.629882335662842, "loss_scale": 4.0, "retries": 0, "seconds": 215.4, "max_memory_gb": 16.44, "pair": 0.3460237830877304, "sign": 0.26441544108092785, "keep": 0.3608210129314102, "leak": 1.146484985947609, "replacement": 0.9452632963657379}
+{"step": 251, "steps": 500, "lr": 2e-05, "grad_norm": 3.0774669647216797, "loss_scale": 4.0, "retries": 0, "seconds": 216.2, "max_memory_gb": 16.44, "pair": 0.44658543542027473, "sign": 0.47408073395490646, "keep": 0.25407327109132893, "leak": 0.9833217486739159, "replacement": 1.0564928650856018}
+{"step": 252, "steps": 500, "lr": 2e-05, "grad_norm": 6.212688446044922, "loss_scale": 4.0, "retries": 0, "seconds": 217.1, "max_memory_gb": 16.44, "pair": 0.24576019495725632, "sign": 0.25242495723068714, "keep": 0.03091755835339427, "leak": 1.2496666386723518, "replacement": 0.6998481303453445}
+{"step": 253, "steps": 500, "lr": 2e-05, "grad_norm": 5.476564884185791, "loss_scale": 4.0, "retries": 0, "seconds": 218.0, "max_memory_gb": 16.44, "pair": 0.3573076259344816, "sign": 0.3315194044262171, "keep": 0.16448695119470358, "leak": 1.3046134866854118, "replacement": 0.9314641803503036}
+{"step": 254, "steps": 500, "lr": 2e-05, "grad_norm": 3.3380980491638184, "loss_scale": 4.0, "retries": 0, "seconds": 218.8, "max_memory_gb": 16.44, "pair": 0.41654037684202194, "sign": 0.5599467810243368, "keep": 0.5560240708291531, "leak": 0.8237252738326788, "replacement": 1.1804570853710175}
+{"step": 255, "steps": 500, "lr": 2e-05, "grad_norm": 1.0607120990753174, "loss_scale": 4.0, "retries": 0, "seconds": 219.6, "max_memory_gb": 16.44, "pair": 0.6291516721248627, "sign": 0.7877049893140793, "keep": 0.2889813557267189, "leak": 0.5799738243222237, "replacement": 1.3124883025884628}
+{"step": 256, "steps": 500, "lr": 2e-05, "grad_norm": 2.1937062740325928, "loss_scale": 4.0, "retries": 0, "seconds": 220.5, "max_memory_gb": 16.44, "pair": 0.567228652536869, "sign": 0.6281436458230019, "keep": 0.27963297767564654, "leak": 0.2695949003100395, "replacement": 1.0885157138109207}
+{"step": 257, "steps": 500, "lr": 2e-05, "grad_norm": 3.7275590896606445, "loss_scale": 4.0, "retries": 0, "seconds": 221.3, "max_memory_gb": 16.44, "pair": 0.47345202043652534, "sign": 0.4970520371571183, "keep": 0.10063313949649455, "leak": 0.9463616120629013, "replacement": 1.008884996175766}
+{"step": 258, "steps": 500, "lr": 2e-05, "grad_norm": 3.9255096912384033, "loss_scale": 4.0, "retries": 0, "seconds": 222.2, "max_memory_gb": 16.44, "pair": 0.21838614717125893, "sign": 0.17707631178200245, "keep": 0.1273157401010394, "leak": 1.6487147063016891, "replacement": 0.7827608436346054}
+{"step": 259, "steps": 500, "lr": 2e-05, "grad_norm": 1.3275374174118042, "loss_scale": 4.0, "retries": 0, "seconds": 223.0, "max_memory_gb": 16.44, "pair": 0.24729258567094803, "sign": 0.22562450543045998, "keep": 0.029500083124730736, "leak": 1.2137743532657623, "replacement": 0.6782984584569931}
+{"step": 260, "steps": 500, "lr": 2e-05, "grad_norm": 3.0864920616149902, "loss_scale": 4.0, "retries": 0, "seconds": 223.9, "max_memory_gb": 16.44, "pair": 0.4489145949482918, "sign": 0.4444351904094219, "keep": 0.3716133274137974, "leak": 0.826842050999403, "replacement": 1.063649371266365}
+{"step": 261, "steps": 500, "lr": 2e-05, "grad_norm": 3.5673515796661377, "loss_scale": 4.0, "retries": 0, "seconds": 224.7, "max_memory_gb": 16.44, "pair": 0.4438071921467781, "sign": 0.3539803456515074, "keep": 0.20752851851284504, "leak": 0.9488678905181587, "replacement": 0.9617785960435867}
+{"step": 262, "steps": 500, "lr": 2e-05, "grad_norm": 4.212146759033203, "loss_scale": 4.0, "retries": 0, "seconds": 225.6, "max_memory_gb": 16.44, "pair": 0.4467686489224434, "sign": 0.5535674658603966, "keep": 0.5768463276326656, "leak": 1.5613188706338406, "replacement": 1.4023052603006363}
+{"step": 263, "steps": 500, "lr": 2e-05, "grad_norm": 4.255216121673584, "loss_scale": 4.0, "retries": 0, "seconds": 226.4, "max_memory_gb": 16.44, "pair": 0.4561854600906372, "sign": 0.4201150070875883, "keep": 0.54011495038867, "leak": 1.137719314545393, "replacement": 1.220730260014534}
+{"step": 264, "steps": 500, "lr": 2e-05, "grad_norm": 3.1831235885620117, "loss_scale": 4.0, "retries": 0, "seconds": 227.3, "max_memory_gb": 16.44, "pair": 0.2821117714047432, "sign": 0.20339988265186548, "keep": 0.8416769927425776, "leak": 1.0466485098004341, "replacement": 1.0663123279809952}
+{"step": 265, "steps": 500, "lr": 2e-05, "grad_norm": 4.17952299118042, "loss_scale": 4.0, "retries": 0, "seconds": 228.1, "max_memory_gb": 16.44, "pair": 0.41920366510748863, "sign": 0.4424675051122904, "keep": 0.18023038154933602, "leak": 1.4477438414469361, "replacement": 1.0924885869026184}
+{"step": 266, "steps": 500, "lr": 2e-05, "grad_norm": 4.741971492767334, "loss_scale": 4.0, "retries": 0, "seconds": 229.0, "max_memory_gb": 16.44, "pair": 0.442585714161396, "sign": 0.3949211537837982, "keep": 0.8229628060944378, "leak": 0.1897475173464045, "replacement": 1.0989645719528198}
+{"step": 267, "steps": 500, "lr": 2e-05, "grad_norm": 5.6830830574035645, "loss_scale": 4.0, "retries": 0, "seconds": 229.9, "max_memory_gb": 16.44, "pair": 0.4397796541452408, "sign": 0.4700186988338828, "keep": 0.2402641822118312, "leak": 1.2291685375967063, "replacement": 1.1022132188081741}
+{"step": 268, "steps": 500, "lr": 2e-05, "grad_norm": 0.8372278213500977, "loss_scale": 4.0, "retries": 0, "seconds": 230.7, "max_memory_gb": 16.44, "pair": 0.6752856895327568, "sign": 0.7779404297471046, "keep": 0.4191163812065497, "leak": 0.3655518264276907, "replacement": 1.3652020990848541}
+{"step": 269, "steps": 500, "lr": 2e-05, "grad_norm": 4.1795148849487305, "loss_scale": 4.0, "retries": 0, "seconds": 231.6, "max_memory_gb": 16.44, "pair": 0.48286040872335434, "sign": 0.5613998025655746, "keep": 0.8988230745308101, "leak": 0.9889146780769806, "replacement": 1.4602005183696747}
+{"step": 270, "steps": 500, "lr": 2e-05, "grad_norm": 2.7027082443237305, "loss_scale": 4.0, "retries": 0, "seconds": 232.4, "max_memory_gb": 16.44, "pair": 0.4492792636156082, "sign": 0.39659673906862736, "keep": 0.25018392922356725, "leak": 0.7746117722708732, "replacement": 0.966322511434555}
+{"step": 271, "steps": 500, "lr": 2e-05, "grad_norm": 5.8769426345825195, "loss_scale": 4.0, "retries": 0, "seconds": 233.3, "max_memory_gb": 16.44, "pair": 0.4272003844380379, "sign": 0.66612309217453, "keep": 0.5897565647028387, "leak": 0.4544653035700321, "replacement": 1.1687565445899963}
+{"step": 272, "steps": 500, "lr": 2e-05, "grad_norm": 9.200750350952148, "loss_scale": 4.0, "retries": 0, "seconds": 234.1, "max_memory_gb": 16.44, "pair": 0.22184259817004204, "sign": 0.10209189355373383, "keep": 0.03197428863495588, "leak": 2.439397618174553, "replacement": 0.8987251073122025}
+{"step": 273, "steps": 500, "lr": 2e-05, "grad_norm": 2.4082772731781006, "loss_scale": 4.0, "retries": 0, "seconds": 235.0, "max_memory_gb": 16.44, "pair": 0.29960381612181664, "sign": 0.17287529818713665, "keep": 0.021570663717284333, "leak": 0.5696497783064842, "replacement": 0.5392392203211784}
+{"step": 274, "steps": 500, "lr": 2e-05, "grad_norm": 5.293117046356201, "loss_scale": 4.0, "retries": 0, "seconds": 235.8, "max_memory_gb": 16.44, "pair": 0.6486045122146606, "sign": 0.9143705368041992, "keep": 0.2976188282482326, "leak": 0.6061495514586568, "replacement": 1.4061364978551865}
+{"step": 275, "steps": 500, "lr": 2e-05, "grad_norm": 4.995746612548828, "loss_scale": 4.0, "retries": 0, "seconds": 236.7, "max_memory_gb": 16.44, "pair": 0.42704472318291664, "sign": 0.4895106302574277, "keep": 1.2587724486365914, "leak": 0.8672813605517149, "replacement": 1.5180065780878067}
+{"step": 276, "steps": 500, "lr": 2e-05, "grad_norm": 6.210148811340332, "loss_scale": 4.0, "retries": 0, "seconds": 237.6, "max_memory_gb": 16.44, "pair": 0.17433782294392586, "sign": 0.1253665266558528, "keep": 0.016145664390933234, "leak": 1.5925743281841278, "replacement": 0.6432375013828278}
+{"step": 277, "steps": 500, "lr": 2e-05, "grad_norm": 2.8068387508392334, "loss_scale": 4.0, "retries": 0, "seconds": 238.4, "max_memory_gb": 16.44, "pair": 0.381512776017189, "sign": 0.31092015840113163, "keep": 0.012181169338873588, "leak": 0.7395439520478249, "replacement": 0.7279494181275368}
+{"step": 278, "steps": 500, "lr": 2e-05, "grad_norm": 2.6626124382019043, "loss_scale": 4.0, "retries": 0, "seconds": 239.3, "max_memory_gb": 16.44, "pair": 0.43249450996518135, "sign": 0.48762958496809006, "keep": 0.2775432542402996, "leak": 0.7530163602350513, "replacement": 1.0033349990844727}
+{"step": 279, "steps": 500, "lr": 2e-05, "grad_norm": 3.1042513847351074, "loss_scale": 4.0, "retries": 0, "seconds": 240.1, "max_memory_gb": 16.44, "pair": 0.4146520122885704, "sign": 0.34258410334587097, "keep": 0.15470802478375845, "leak": 1.1037113973870873, "replacement": 0.9392259418964386}
+{"step": 280, "steps": 500, "lr": 2e-05, "grad_norm": 7.031767845153809, "loss_scale": 4.0, "retries": 0, "seconds": 240.9, "max_memory_gb": 16.44, "pair": 0.2803127933293581, "sign": 0.3003728725016117, "keep": 0.05165074465912767, "leak": 1.4557785466313362, "replacement": 0.8202692419290543}
+{"step": 281, "steps": 500, "lr": 2e-05, "grad_norm": 3.1197257041931152, "loss_scale": 4.0, "retries": 0, "seconds": 241.8, "max_memory_gb": 16.44, "pair": 0.3960119038820267, "sign": 0.3353199418634176, "keep": 0.6341457414673641, "leak": 0.6544576259329915, "replacement": 1.0443591326475143}
+{"step": 282, "steps": 500, "lr": 2e-05, "grad_norm": 7.507023811340332, "loss_scale": 4.0, "retries": 0, "seconds": 242.6, "max_memory_gb": 16.44, "pair": 0.3989300765097141, "sign": 0.44798213988542557, "keep": 0.3478203612467041, "leak": 1.395834468305111, "replacement": 1.1457899436354637}
+{"step": 283, "steps": 500, "lr": 2e-05, "grad_norm": 6.013274669647217, "loss_scale": 4.0, "retries": 0, "seconds": 243.5, "max_memory_gb": 16.44, "pair": 0.18474168330430984, "sign": 0.12301377020776272, "keep": 0.1719590649008751, "leak": 1.9423574656248093, "replacement": 0.8178174644708633}
+{"step": 284, "steps": 500, "lr": 2e-05, "grad_norm": 5.117859840393066, "loss_scale": 4.0, "retries": 0, "seconds": 244.4, "max_memory_gb": 16.44, "pair": 0.3882302790880203, "sign": 0.3483766121789813, "keep": 0.2682018559426069, "leak": 0.9194143861532211, "replacement": 0.9263730943202972}
+{"step": 285, "steps": 500, "lr": 2e-05, "grad_norm": 5.730623722076416, "loss_scale": 4.0, "retries": 0, "seconds": 245.2, "max_memory_gb": 16.44, "pair": 0.5361395925283432, "sign": 0.7949801608920097, "keep": 0.17735232738777995, "leak": 0.8942121900618076, "replacement": 1.2458589002490044}
+{"step": 286, "steps": 500, "lr": 2e-05, "grad_norm": 2.5149693489074707, "loss_scale": 4.0, "retries": 0, "seconds": 246.1, "max_memory_gb": 16.44, "pair": 0.43422459810972214, "sign": 0.8843826130032539, "keep": 0.04789141728542745, "leak": 1.0722612589597702, "replacement": 1.1684269309043884}
+{"step": 287, "steps": 500, "lr": 2e-05, "grad_norm": 6.213514804840088, "loss_scale": 4.0, "retries": 0, "seconds": 247.0, "max_memory_gb": 16.44, "pair": 0.4323645383119583, "sign": 0.5034974650479853, "keep": 0.5749012832529843, "leak": 1.0965556465089321, "replacement": 1.245702862739563}
+{"step": 288, "steps": 500, "lr": 2e-05, "grad_norm": 5.956619739532471, "loss_scale": 4.0, "retries": 0, "seconds": 247.9, "max_memory_gb": 16.44, "pair": 0.3371661603450775, "sign": 0.5674911621026695, "keep": 0.02574373537208885, "leak": 1.2205309718847275, "replacement": 0.9389163255691528}
+{"step": 289, "steps": 500, "lr": 2e-05, "grad_norm": 3.6797611713409424, "loss_scale": 4.0, "retries": 0, "seconds": 248.7, "max_memory_gb": 16.44, "pair": 0.3690881244838238, "sign": 0.30867696925997734, "keep": 0.31389315612614155, "leak": 0.6042225658893585, "replacement": 0.8314287811517715}
+{"step": 290, "steps": 500, "lr": 2e-05, "grad_norm": 4.516815662384033, "loss_scale": 4.0, "retries": 0, "seconds": 249.6, "max_memory_gb": 16.44, "pair": 0.5080957151949406, "sign": 0.5969848521053791, "keep": 0.5767054698953871, "leak": 0.7642595288343728, "replacement": 1.2860057353973389}
+{"step": 291, "steps": 500, "lr": 2e-05, "grad_norm": 4.246055603027344, "loss_scale": 4.0, "retries": 0, "seconds": 250.4, "max_memory_gb": 16.44, "pair": 0.19604165479540825, "sign": 0.08354145660996437, "keep": 0.029405590787064284, "leak": 1.263093762099743, "replacement": 0.5682886093854904}
+{"step": 292, "steps": 500, "lr": 2e-05, "grad_norm": 5.055490970611572, "loss_scale": 4.0, "retries": 0, "seconds": 251.3, "max_memory_gb": 16.44, "pair": 0.3799942843616009, "sign": 0.28445993922650814, "keep": 0.13979499228298664, "leak": 0.6571018844842911, "replacement": 0.7563972175121307}
+{"step": 293, "steps": 500, "lr": 2e-05, "grad_norm": 3.7831623554229736, "loss_scale": 4.0, "retries": 0, "seconds": 252.1, "max_memory_gb": 16.44, "pair": 0.45493385195732117, "sign": 0.32520119473338127, "keep": 0.182526697171852, "leak": 0.7486390098929405, "replacement": 0.8959575593471527}
+{"step": 294, "steps": 500, "lr": 2e-05, "grad_norm": 4.0405168533325195, "loss_scale": 4.0, "retries": 0, "seconds": 253.0, "max_memory_gb": 16.44, "pair": 0.299953680485487, "sign": 0.19837494753301144, "keep": 0.4187353290617466, "leak": 0.5160240810364485, "replacement": 0.7375148385763168}
+{"step": 295, "steps": 500, "lr": 2e-05, "grad_norm": 4.142764568328857, "loss_scale": 4.0, "retries": 0, "seconds": 253.8, "max_memory_gb": 16.44, "pair": 0.528603620827198, "sign": 0.8824191465973854, "keep": 0.9019427355378866, "leak": 0.9954214096069336, "replacement": 1.6696398854255676}
+{"step": 296, "steps": 500, "lr": 2e-05, "grad_norm": 2.7061657905578613, "loss_scale": 4.0, "retries": 0, "seconds": 254.6, "max_memory_gb": 16.44, "pair": 0.5941376276314259, "sign": 0.7043588533997536, "keep": 0.3428893154487014, "leak": 0.5524188386189053, "replacement": 1.2558664232492447}
+{"step": 297, "steps": 500, "lr": 2e-05, "grad_norm": 1.3336549997329712, "loss_scale": 4.0, "retries": 0, "seconds": 255.5, "max_memory_gb": 16.44, "pair": 0.5196148380637169, "sign": 0.7277663126587868, "keep": 0.070823430782184, "leak": 0.5415126420557499, "replacement": 1.054287850856781}
+{"step": 298, "steps": 500, "lr": 2e-05, "grad_norm": 6.716101169586182, "loss_scale": 4.0, "retries": 0, "seconds": 256.3, "max_memory_gb": 16.44, "pair": 0.6260397732257843, "sign": 0.8326456993818283, "keep": 1.052859346353216, "leak": 0.19225841481238604, "replacement": 1.6168569326400757}
+{"step": 299, "steps": 500, "lr": 2e-05, "grad_norm": 5.682188034057617, "loss_scale": 4.0, "retries": 0, "seconds": 257.2, "max_memory_gb": 16.44, "pair": 0.3163127191364765, "sign": 0.1323979366570711, "keep": 0.04306641570292413, "leak": 1.0149705782532692, "replacement": 0.6577875316143036}
+{"step": 300, "steps": 500, "lr": 2e-05, "grad_norm": 3.075500249862671, "loss_scale": 4.0, "retries": 0, "seconds": 258.1, "max_memory_gb": 16.44, "pair": 0.7001256160438061, "sign": 0.8337885178625584, "keep": 0.38054525665938854, "leak": 0.6398489306739066, "replacement": 1.4672547429800034}
+{"step": 301, "steps": 500, "lr": 2e-05, "grad_norm": 3.9765212535858154, "loss_scale": 4.0, "retries": 0, "seconds": 258.9, "max_memory_gb": 16.44, "pair": 0.42771680280566216, "sign": 0.4580339342355728, "keep": 0.21316359378397465, "leak": 0.8717751931399107, "replacement": 0.9812593758106232}
+{"step": 302, "steps": 500, "lr": 2e-05, "grad_norm": 6.986327171325684, "loss_scale": 4.0, "retries": 0, "seconds": 259.7, "max_memory_gb": 16.44, "pair": 0.2090741004794836, "sign": 0.0940596479922533, "keep": 0.042074709897860885, "leak": 1.7969230264425278, "replacement": 0.7263720482587814}
+{"step": 303, "steps": 500, "lr": 2e-05, "grad_norm": 4.073938846588135, "loss_scale": 4.0, "retries": 0, "seconds": 260.6, "max_memory_gb": 16.44, "pair": 0.301094226539135, "sign": 0.21324225375428796, "keep": 0.030715200773556717, "leak": 1.4107962250709534, "replacement": 0.7757719978690147}
+{"step": 304, "steps": 500, "lr": 2e-05, "grad_norm": 6.606196880340576, "loss_scale": 4.0, "retries": 0, "seconds": 261.5, "max_memory_gb": 16.44, "pair": 0.46733658760786057, "sign": 0.6151573732495308, "keep": 0.1120939701795578, "leak": 1.2210844904184341, "replacement": 1.1362333595752716}
+{"step": 305, "steps": 500, "lr": 2e-05, "grad_norm": 5.2039923667907715, "loss_scale": 4.0, "retries": 0, "seconds": 262.3, "max_memory_gb": 16.44, "pair": 0.2728828489780426, "sign": 0.2129428619518876, "keep": 0.15599097423546482, "leak": 1.5175404220353812, "replacement": 0.836734876036644}
+{"step": 306, "steps": 500, "lr": 2e-05, "grad_norm": 5.366095066070557, "loss_scale": 4.0, "retries": 0, "seconds": 263.2, "max_memory_gb": 16.44, "pair": 0.21786253154277802, "sign": 0.06411970779299736, "keep": 0.08603920228779316, "leak": 1.2476270273327827, "replacement": 0.6048487275838852}
+{"step": 307, "steps": 500, "lr": 2e-05, "grad_norm": 6.0055251121521, "loss_scale": 4.0, "retries": 0, "seconds": 264.0, "max_memory_gb": 16.44, "pair": 0.354471443220973, "sign": 0.3383810706436634, "keep": 0.0931481309235096, "leak": 1.6411232719983673, "replacement": 0.9805168509483337}
+{"step": 308, "steps": 500, "lr": 2e-05, "grad_norm": 5.143176555633545, "loss_scale": 4.0, "retries": 0, "seconds": 264.9, "max_memory_gb": 16.44, "pair": 0.4755620155483484, "sign": 0.6331864576786757, "keep": 0.13834380575281102, "leak": 1.1756043005734682, "replacement": 1.155228167772293}
+{"step": 309, "steps": 500, "lr": 2e-05, "grad_norm": 4.082196235656738, "loss_scale": 4.0, "retries": 0, "seconds": 265.7, "max_memory_gb": 16.44, "pair": 0.5817653276026249, "sign": 0.6464096501003951, "keep": 0.32779455417767167, "leak": 0.6414212612435222, "replacement": 1.229222759604454}
+{"step": 310, "steps": 500, "lr": 2e-05, "grad_norm": 2.9245269298553467, "loss_scale": 4.0, "retries": 0, "seconds": 266.6, "max_memory_gb": 16.44, "pair": 0.25932255387306213, "sign": 0.13752661319449544, "keep": 0.02181752929027425, "leak": 0.5999709218740463, "replacement": 0.48898735642433167}
+{"step": 311, "steps": 500, "lr": 2e-05, "grad_norm": 1.5851062536239624, "loss_scale": 4.0, "retries": 0, "seconds": 267.4, "max_memory_gb": 16.44, "pair": 0.4068928714841604, "sign": 0.4664296954870224, "keep": 0.19212680065538734, "leak": 0.9827630114741623, "replacement": 0.9818618670105934}
+{"step": 312, "steps": 500, "lr": 2e-05, "grad_norm": 4.261401176452637, "loss_scale": 4.0, "retries": 0, "seconds": 268.2, "max_memory_gb": 16.44, "pair": 0.32906705513596535, "sign": 0.3342722291126847, "keep": 0.0499671066791052, "leak": 0.7057854514569044, "replacement": 0.6976330950856209}
+{"step": 313, "steps": 500, "lr": 2e-05, "grad_norm": 8.12452507019043, "loss_scale": 4.0, "retries": 0, "seconds": 269.1, "max_memory_gb": 16.44, "pair": 0.45116445422172546, "sign": 0.5452116867527366, "keep": 0.35205054795369506, "leak": 1.4100744184106588, "replacement": 1.2523141950368881}
+{"step": 314, "steps": 500, "lr": 2e-05, "grad_norm": 5.157650947570801, "loss_scale": 4.0, "retries": 0, "seconds": 270.0, "max_memory_gb": 16.44, "pair": 0.3791367094963789, "sign": 0.4891507728025317, "keep": 0.21784556150669232, "leak": 1.082156353048049, "replacement": 1.0031740069389343}
+{"step": 315, "steps": 500, "lr": 2e-05, "grad_norm": 5.439639091491699, "loss_scale": 4.0, "retries": 0, "seconds": 270.9, "max_memory_gb": 16.44, "pair": 0.33728119917213917, "sign": 0.2532818764448166, "keep": 0.07457201508805156, "leak": 1.2391846384853125, "replacement": 0.8110043108463287}
+{"step": 316, "steps": 500, "lr": 2e-05, "grad_norm": 5.016757488250732, "loss_scale": 4.0, "retries": 0, "seconds": 271.7, "max_memory_gb": 16.44, "pair": 0.60389369353652, "sign": 0.7272027311846614, "keep": 0.7822401188313961, "leak": 0.3797939922660589, "replacement": 1.4535636603832245}
+{"step": 317, "steps": 500, "lr": 2e-05, "grad_norm": 3.148660898208618, "loss_scale": 4.0, "retries": 0, "seconds": 272.6, "max_memory_gb": 16.44, "pair": 0.5847746506333351, "sign": 0.4907381385564804, "keep": 0.47568007558584213, "leak": 0.6747339740395546, "replacement": 1.2366672456264496}
+{"step": 318, "steps": 500, "lr": 2e-05, "grad_norm": 3.0884149074554443, "loss_scale": 4.0, "retries": 0, "seconds": 273.4, "max_memory_gb": 16.44, "pair": 0.44063203036785126, "sign": 0.34105900675058365, "keep": 0.07882163010071963, "leak": 0.439085740596056, "replacement": 0.7603437677025795}
+{"step": 319, "steps": 500, "lr": 2e-05, "grad_norm": 3.511502504348755, "loss_scale": 4.0, "retries": 0, "seconds": 274.2, "max_memory_gb": 16.44, "pair": 0.49041537195444107, "sign": 0.4040325414389372, "keep": 0.23321669455617666, "leak": 0.5816797688603401, "replacement": 0.9544599056243896}
+{"step": 320, "steps": 500, "lr": 2e-05, "grad_norm": 2.628509044647217, "loss_scale": 4.0, "retries": 0, "seconds": 275.1, "max_memory_gb": 16.44, "pair": 0.39455683901906013, "sign": 0.4007055386900902, "keep": 0.32352249988616677, "leak": 0.7255314141511917, "replacement": 0.938053697347641}
+{"step": 321, "steps": 500, "lr": 2e-05, "grad_norm": 3.493344306945801, "loss_scale": 4.0, "retries": 0, "seconds": 275.9, "max_memory_gb": 16.44, "pair": 0.3309783861041069, "sign": 0.3254032814875245, "keep": 0.21105023357085884, "leak": 0.8431283188983798, "replacement": 0.8099872171878815}
+{"step": 322, "steps": 500, "lr": 2e-05, "grad_norm": 5.490199565887451, "loss_scale": 4.0, "retries": 0, "seconds": 276.8, "max_memory_gb": 16.44, "pair": 0.5026938877999783, "sign": 0.5449795201420784, "keep": 0.043603355530649424, "leak": 0.5710094289388508, "replacement": 0.9397376775741577}
+{"step": 323, "steps": 500, "lr": 2e-05, "grad_norm": 7.125697612762451, "loss_scale": 4.0, "retries": 0, "seconds": 277.6, "max_memory_gb": 16.44, "pair": 0.3073924817144871, "sign": 0.27944930735975504, "keep": 0.13083165418356657, "leak": 2.0629041492938995, "replacement": 1.0282589942216873}
+{"step": 324, "steps": 500, "lr": 2e-05, "grad_norm": 3.3546981811523438, "loss_scale": 4.0, "retries": 0, "seconds": 278.5, "max_memory_gb": 16.44, "pair": 0.2736281231045723, "sign": 0.13302204012870789, "keep": 0.060647445294307545, "leak": 1.0708849653601646, "replacement": 0.6381840780377388}
+{"step": 325, "steps": 500, "lr": 2e-05, "grad_norm": 2.31613826751709, "loss_scale": 4.0, "retries": 0, "seconds": 279.3, "max_memory_gb": 16.44, "pair": 0.5363790020346642, "sign": 0.5738040127325803, "keep": 0.47059041159809567, "leak": 0.24091430701082572, "replacement": 1.1188048347830772}
+{"step": 326, "steps": 500, "lr": 2e-05, "grad_norm": 3.5102362632751465, "loss_scale": 4.0, "retries": 0, "seconds": 280.2, "max_memory_gb": 16.44, "pair": 0.472063273191452, "sign": 0.6808320973068476, "keep": 0.19229218023247086, "leak": 0.6421262146250228, "replacement": 1.069156974554062}
+{"step": 327, "steps": 500, "lr": 2e-05, "grad_norm": 4.201755046844482, "loss_scale": 4.0, "retries": 0, "seconds": 281.1, "max_memory_gb": 16.44, "pair": 0.2801585979759693, "sign": 0.11811301996931434, "keep": 0.37634893879294395, "leak": 0.5618542358279228, "replacement": 0.6678531169891357}
+{"step": 328, "steps": 500, "lr": 2e-05, "grad_norm": 3.1296918392181396, "loss_scale": 4.0, "retries": 0, "seconds": 281.9, "max_memory_gb": 16.44, "pair": 0.6116558238863945, "sign": 0.6514239385724068, "keep": 0.2052497597760521, "leak": 0.18218113703187555, "replacement": 1.0855379402637482}
+{"step": 329, "steps": 500, "lr": 2e-05, "grad_norm": 3.6275970935821533, "loss_scale": 4.0, "retries": 0, "seconds": 282.8, "max_memory_gb": 16.44, "pair": 0.45567335560917854, "sign": 0.41770828142762184, "keep": 0.1548562702955678, "leak": 0.7248031637864187, "replacement": 0.9231564179062843}
+{"step": 330, "steps": 500, "lr": 2e-05, "grad_norm": 4.303316116333008, "loss_scale": 4.0, "retries": 0, "seconds": 283.6, "max_memory_gb": 16.44, "pair": 0.5037628412246704, "sign": 0.5234738569706678, "keep": 0.27471677038556663, "leak": 0.6333245225250721, "replacement": 1.0611892938613892}
+{"step": 331, "steps": 500, "lr": 2e-05, "grad_norm": 2.9543073177337646, "loss_scale": 4.0, "retries": 0, "seconds": 284.5, "max_memory_gb": 16.44, "pair": 0.6312031373381615, "sign": 0.6090917661786079, "keep": 0.4977423113887198, "leak": 0.6439804985420778, "replacement": 1.345615342259407}
+{"step": 332, "steps": 500, "lr": 2e-05, "grad_norm": 4.075744152069092, "loss_scale": 4.0, "retries": 0, "seconds": 285.3, "max_memory_gb": 16.44, "pair": 0.422577828168869, "sign": 0.40698258206248283, "keep": 0.34636515354941366, "leak": 0.7072119815275073, "replacement": 0.9760546684265137}
+{"step": 333, "steps": 500, "lr": 2e-05, "grad_norm": 2.300813674926758, "loss_scale": 4.0, "retries": 0, "seconds": 286.2, "max_memory_gb": 16.44, "pair": 0.7177461087703705, "sign": 0.8050845935940742, "keep": 0.4560876961913891, "leak": 0.2932690759698744, "replacement": 1.421649530529976}
+{"step": 334, "steps": 500, "lr": 2e-05, "grad_norm": 4.227468490600586, "loss_scale": 4.0, "retries": 0, "seconds": 287.0, "max_memory_gb": 16.44, "pair": 0.5540605112910271, "sign": 0.5242401286959648, "keep": 0.7673383811488748, "leak": 0.7859110906720161, "replacement": 1.3963275253772736}
+{"step": 335, "steps": 500, "lr": 2e-05, "grad_norm": 4.728090286254883, "loss_scale": 4.0, "retries": 0, "seconds": 287.9, "max_memory_gb": 16.44, "pair": 0.48660973832011223, "sign": 0.5664173439145088, "keep": 0.8352522779532592, "leak": 0.5529411807656288, "replacement": 1.325679898262024}
+{"step": 336, "steps": 500, "lr": 2e-05, "grad_norm": 3.6619577407836914, "loss_scale": 4.0, "retries": 0, "seconds": 288.8, "max_memory_gb": 16.44, "pair": 0.2735487725585699, "sign": 0.22817336581647396, "keep": 0.03883336552098626, "leak": 1.6661624684929848, "replacement": 0.8235927671194077}
+{"step": 337, "steps": 500, "lr": 2e-05, "grad_norm": 4.292754650115967, "loss_scale": 4.0, "retries": 0, "seconds": 289.6, "max_memory_gb": 16.44, "pair": 0.5426398776471615, "sign": 1.1513542104512453, "keep": 0.5950979068875313, "leak": 0.6193408248946071, "replacement": 1.5707011073827744}
+{"step": 338, "steps": 500, "lr": 2e-05, "grad_norm": 2.4629194736480713, "loss_scale": 4.0, "retries": 0, "seconds": 290.5, "max_memory_gb": 16.44, "pair": 0.46378055214881897, "sign": 0.5450549684464931, "keep": 0.27227400336414576, "leak": 0.961771705377032, "replacement": 1.1128879338502884}
+{"step": 339, "steps": 500, "lr": 2e-05, "grad_norm": 3.1984496116638184, "loss_scale": 4.0, "retries": 0, "seconds": 291.4, "max_memory_gb": 16.44, "pair": 0.5724059604108334, "sign": 0.7785014994442463, "keep": 0.23269115563016385, "leak": 0.8390167528996244, "replacement": 1.2877565175294876}
+{"step": 340, "steps": 500, "lr": 2e-05, "grad_norm": 2.488448143005371, "loss_scale": 4.0, "retries": 0, "seconds": 292.2, "max_memory_gb": 16.44, "pair": 0.4094429388642311, "sign": 0.3703976795077324, "keep": 0.49532262585125864, "leak": 0.5481718068476766, "replacement": 0.9793460518121719}
+{"step": 341, "steps": 500, "lr": 2e-05, "grad_norm": 3.728019952774048, "loss_scale": 4.0, "retries": 0, "seconds": 293.1, "max_memory_gb": 16.44, "pair": 0.44162584096193314, "sign": 0.43148930836468935, "keep": 0.5890941014513373, "leak": 0.5962888140056748, "replacement": 1.1009897291660309}
+{"step": 342, "steps": 500, "lr": 2e-05, "grad_norm": 1.9982523918151855, "loss_scale": 4.0, "retries": 0, "seconds": 294.0, "max_memory_gb": 16.44, "pair": 0.6013182178139687, "sign": 0.6295845173299313, "keep": 0.41151961497962475, "leak": 0.24542592151556164, "replacement": 1.183226764202118}
+{"step": 343, "steps": 500, "lr": 2e-05, "grad_norm": 3.8511483669281006, "loss_scale": 4.0, "retries": 0, "seconds": 294.8, "max_memory_gb": 16.44, "pair": 0.4290155656635761, "sign": 0.5542074553668499, "keep": 0.029198538642958738, "leak": 0.6241882275789976, "replacement": 0.8767656087875366}
+{"step": 344, "steps": 500, "lr": 2e-05, "grad_norm": 5.3324971199035645, "loss_scale": 4.0, "retries": 0, "seconds": 295.6, "max_memory_gb": 16.44, "pair": 0.2001001052558422, "sign": 0.15301155857741833, "keep": 0.021504736505448818, "leak": 1.3601947128772736, "replacement": 0.6274069398641586}
+{"step": 345, "steps": 500, "lr": 2e-05, "grad_norm": 4.879385948181152, "loss_scale": 4.0, "retries": 0, "seconds": 296.5, "max_memory_gb": 16.44, "pair": 0.39562220126390457, "sign": 0.39870656467974186, "keep": 0.19384692003950477, "leak": 0.942464198495145, "replacement": 0.9275150001049042}
+{"step": 346, "steps": 500, "lr": 2e-05, "grad_norm": 2.9949886798858643, "loss_scale": 4.0, "retries": 0, "seconds": 297.3, "max_memory_gb": 16.44, "pair": 0.4775160476565361, "sign": 0.6933378912508488, "keep": 0.7397255495743593, "leak": 0.7170820310711861, "replacement": 1.3733182549476624}
+{"step": 347, "steps": 500, "lr": 2e-05, "grad_norm": 5.487990379333496, "loss_scale": 4.0, "retries": 0, "seconds": 298.2, "max_memory_gb": 16.44, "pair": 0.37793920934200287, "sign": 0.39904433814808726, "keep": 0.32697533164173365, "leak": 0.9637276113498956, "replacement": 0.9818809479475021}
+{"step": 348, "steps": 500, "lr": 2e-05, "grad_norm": 5.493454933166504, "loss_scale": 4.0, "retries": 0, "seconds": 299.0, "max_memory_gb": 16.44, "pair": 0.1984014492481947, "sign": 0.20630869176238775, "keep": 0.023639294086024165, "leak": 1.4755661636590958, "replacement": 0.6822669953107834}
+{"step": 349, "steps": 500, "lr": 2e-05, "grad_norm": 5.969968795776367, "loss_scale": 4.0, "retries": 0, "seconds": 299.9, "max_memory_gb": 16.44, "pair": 0.405019398778677, "sign": 0.41794642992317677, "keep": 0.1468179855728522, "leak": 1.5647747457260266, "replacement": 1.0785952806472778}
+{"step": 350, "steps": 500, "lr": 2e-05, "grad_norm": 1.8191378116607666, "loss_scale": 4.0, "retries": 0, "seconds": 300.8, "max_memory_gb": 16.44, "pair": 0.4900229722261429, "sign": 0.7130970805883408, "keep": 0.20989113673567772, "leak": 0.10918643073819112, "replacement": 0.9788136780261993}
+{"step": 351, "steps": 500, "lr": 2e-05, "grad_norm": 3.4125397205352783, "loss_scale": 4.0, "retries": 0, "seconds": 301.6, "max_memory_gb": 16.44, "pair": 0.29363198205828667, "sign": 0.18523413687944412, "keep": 0.09544880734756589, "leak": 0.2568891132250428, "replacement": 0.49819572269916534}
+{"step": 352, "steps": 500, "lr": 2e-05, "grad_norm": 2.335865020751953, "loss_scale": 4.0, "retries": 0, "seconds": 302.4, "max_memory_gb": 16.44, "pair": 0.6118197813630104, "sign": 0.7488361895084381, "keep": 0.8487119907513261, "leak": 0.5836260921787471, "replacement": 1.5565003380179405}
+{"step": 353, "steps": 500, "lr": 2e-05, "grad_norm": 5.618778705596924, "loss_scale": 4.0, "retries": 0, "seconds": 303.3, "max_memory_gb": 16.44, "pair": 0.42334912717342377, "sign": 0.5970572717487812, "keep": 0.6807105075567961, "leak": 1.1171835958957672, "replacement": 1.3415288925170898}
+{"step": 354, "steps": 500, "lr": 2e-05, "grad_norm": 3.8339173793792725, "loss_scale": 4.0, "retries": 0, "seconds": 304.1, "max_memory_gb": 16.44, "pair": 0.46256381645798683, "sign": 0.48259636387228966, "keep": 0.3096477667313593, "leak": 0.49347888933334616, "replacement": 0.9820556119084358}
+{"step": 355, "steps": 500, "lr": 2e-05, "grad_norm": 4.725985527038574, "loss_scale": 4.0, "retries": 0, "seconds": 305.0, "max_memory_gb": 16.44, "pair": 0.48256276175379753, "sign": 0.4917770763859153, "keep": 0.204279842087999, "leak": 1.0551265931135276, "replacement": 1.0943728983402252}
+{"step": 356, "steps": 500, "lr": 2e-05, "grad_norm": 2.9169087409973145, "loss_scale": 4.0, "retries": 0, "seconds": 305.9, "max_memory_gb": 16.44, "pair": 0.1852820087224245, "sign": 0.20433442946523428, "keep": 0.07048849482089281, "leak": 1.1441571414470673, "replacement": 0.608732745051384}
+{"step": 357, "steps": 500, "lr": 2e-05, "grad_norm": 14.792224884033203, "loss_scale": 4.0, "retries": 0, "seconds": 306.7, "max_memory_gb": 16.44, "pair": 0.20806316658854485, "sign": 0.17745840968564153, "keep": 0.22349347919225693, "leak": 3.7033756375312805, "replacement": 1.3343830406665802}
+{"step": 358, "steps": 500, "lr": 2e-05, "grad_norm": 3.846219301223755, "loss_scale": 4.0, "retries": 0, "seconds": 307.6, "max_memory_gb": 16.44, "pair": 0.568776298314333, "sign": 0.6710845865309238, "keep": 0.3893406766001135, "leak": 0.5715334974229336, "replacement": 1.2418723106384277}
+{"step": 359, "steps": 500, "lr": 2e-05, "grad_norm": 1.742060899734497, "loss_scale": 4.0, "retries": 0, "seconds": 308.5, "max_memory_gb": 16.44, "pair": 0.46012788265943527, "sign": 0.6249059066176414, "keep": 0.05539504997432232, "leak": 0.43179755281016696, "replacement": 0.908227726817131}
+{"step": 360, "steps": 500, "lr": 2e-05, "grad_norm": 3.2802374362945557, "loss_scale": 4.0, "retries": 0, "seconds": 309.3, "max_memory_gb": 16.44, "pair": 0.2555167432874441, "sign": 0.2385238092392683, "keep": 0.0870361060515279, "leak": 0.8445998765528202, "replacement": 0.6294466778635979}
+{"step": 361, "steps": 500, "lr": 2e-05, "grad_norm": 3.663027763366699, "loss_scale": 4.0, "retries": 0, "seconds": 310.2, "max_memory_gb": 16.44, "pair": 0.33189692348241806, "sign": 0.5367508977651596, "keep": 0.0641545970865991, "leak": 1.0627979263663292, "replacement": 0.8980491608381271}
+{"step": 362, "steps": 500, "lr": 2e-05, "grad_norm": 8.02877140045166, "loss_scale": 4.0, "retries": 0, "seconds": 311.0, "max_memory_gb": 16.44, "pair": 0.47491971030831337, "sign": 0.47483911644667387, "keep": 1.6581914456619415, "leak": 0.3847678587771952, "replacement": 1.637626938521862}
+{"step": 363, "steps": 500, "lr": 2e-05, "grad_norm": 8.31628131866455, "loss_scale": 4.0, "retries": 0, "seconds": 311.8, "max_memory_gb": 16.44, "pair": 0.6604619696736336, "sign": 1.389636404812336, "keep": 0.08593960665166378, "leak": 1.2817104905843735, "replacement": 1.7186776399612427}
+{"step": 364, "steps": 500, "lr": 2e-05, "grad_norm": 6.39764404296875, "loss_scale": 4.0, "retries": 0, "seconds": 312.7, "max_memory_gb": 16.44, "pair": 0.2092198207974434, "sign": 0.08321162452921271, "keep": 0.06975390436127782, "leak": 2.1219828128814697, "replacement": 0.8161982893943787}
+{"step": 365, "steps": 500, "lr": 2e-05, "grad_norm": 3.2791664600372314, "loss_scale": 4.0, "retries": 0, "seconds": 313.6, "max_memory_gb": 16.44, "pair": 0.3731190674006939, "sign": 0.41861524153500795, "keep": 0.30204146314645186, "leak": 0.4019543118774891, "replacement": 0.8339359760284424}
+{"step": 366, "steps": 500, "lr": 2e-05, "grad_norm": 8.931947708129883, "loss_scale": 4.0, "retries": 0, "seconds": 314.4, "max_memory_gb": 16.44, "pair": 0.3775697164237499, "sign": 0.34685636311769485, "keep": 0.06187992636114359, "leak": 2.0728150755167007, "replacement": 1.1001416593790054}
+{"step": 367, "steps": 500, "lr": 2e-05, "grad_norm": 2.401315212249756, "loss_scale": 4.0, "retries": 0, "seconds": 315.3, "max_memory_gb": 16.44, "pair": 0.34621887654066086, "sign": 0.389534430578351, "keep": 0.2185392762767151, "leak": 0.809617662394885, "replacement": 0.8526601493358612}
+{"step": 368, "steps": 500, "lr": 2e-05, "grad_norm": 4.418602466583252, "loss_scale": 4.0, "retries": 0, "seconds": 316.1, "max_memory_gb": 16.44, "pair": 0.23156272247433662, "sign": 0.12158621475100517, "keep": 0.06649953679152532, "leak": 1.0353917129687034, "replacement": 0.5844535380601883}
+{"step": 369, "steps": 500, "lr": 2e-05, "grad_norm": 8.109701156616211, "loss_scale": 4.0, "retries": 0, "seconds": 316.9, "max_memory_gb": 16.44, "pair": 0.3217795295640826, "sign": 0.3900116574950516, "keep": 0.08373962929181289, "leak": 2.471372861415148, "replacement": 1.1764984130859375}
+{"step": 370, "steps": 500, "lr": 2e-05, "grad_norm": 3.9774603843688965, "loss_scale": 4.0, "retries": 0, "seconds": 317.8, "max_memory_gb": 16.44, "pair": 0.2926938235759735, "sign": 0.33099116012454033, "keep": 1.1335104629397392, "leak": 0.9185148328542709, "replacement": 1.2545733749866486}
+{"step": 371, "steps": 500, "lr": 2e-05, "grad_norm": 4.930727005004883, "loss_scale": 4.0, "retries": 0, "seconds": 318.7, "max_memory_gb": 16.44, "pair": 0.257175549864769, "sign": 0.1035647289827466, "keep": 0.15931070782244205, "leak": 1.3429471291601658, "replacement": 0.7243500500917435}
+{"step": 372, "steps": 500, "lr": 2e-05, "grad_norm": 7.990305423736572, "loss_scale": 4.0, "retries": 0, "seconds": 319.5, "max_memory_gb": 16.44, "pair": 0.24733998626470566, "sign": 0.271622683852911, "keep": 0.1359518278914038, "leak": 1.384576829150319, "replacement": 0.7972714751958847}
+{"step": 373, "steps": 500, "lr": 2e-05, "grad_norm": 1.3924078941345215, "loss_scale": 4.0, "retries": 0, "seconds": 320.3, "max_memory_gb": 16.44, "pair": 0.6076878644526005, "sign": 0.7968176454305649, "keep": 0.09877812354534399, "leak": 0.4889385560527444, "replacement": 1.1777203679084778}
+{"step": 374, "steps": 500, "lr": 2e-05, "grad_norm": 4.311777591705322, "loss_scale": 4.0, "retries": 0, "seconds": 321.2, "max_memory_gb": 16.44, "pair": 0.3320755157619715, "sign": 0.3209719080477953, "keep": 0.48149181163171306, "leak": 1.3142106477171183, "replacement": 1.061860054731369}
+{"step": 375, "steps": 500, "lr": 2e-05, "grad_norm": 5.4369215965271, "loss_scale": 4.0, "retries": 0, "seconds": 322.1, "max_memory_gb": 16.44, "pair": 0.4774831235408783, "sign": 0.5548222362995148, "keep": 0.33444343163864687, "leak": 1.1084950142540038, "replacement": 1.1992397159337997}
+{"step": 376, "steps": 500, "lr": 2e-05, "grad_norm": 3.1969902515411377, "loss_scale": 4.0, "retries": 0, "seconds": 322.9, "max_memory_gb": 16.44, "pair": 0.3651581443846226, "sign": 0.29610441625118256, "keep": 0.31908808622392826, "leak": 0.3895535357296467, "replacement": 0.7701427936553955}
+{"step": 377, "steps": 500, "lr": 2e-05, "grad_norm": 2.929729700088501, "loss_scale": 4.0, "retries": 0, "seconds": 323.8, "max_memory_gb": 16.44, "pair": 0.3666304051876068, "sign": 0.3845168808475137, "keep": 0.12492850597482175, "leak": 0.7086146827787161, "replacement": 0.7985067516565323}
+{"step": 378, "steps": 500, "lr": 2e-05, "grad_norm": 5.291869163513184, "loss_scale": 4.0, "retries": 0, "seconds": 324.6, "max_memory_gb": 16.44, "pair": 0.2705332562327385, "sign": 0.21543312445282936, "keep": 0.15137514751404524, "leak": 0.7962364703416824, "replacement": 0.6529965102672577}
+{"step": 379, "steps": 500, "lr": 2e-05, "grad_norm": 3.381925344467163, "loss_scale": 4.0, "retries": 0, "seconds": 325.5, "max_memory_gb": 16.44, "pair": 0.39016933739185333, "sign": 0.37045764178037643, "keep": 0.14251969382166862, "leak": 0.6238327547907829, "replacement": 0.802616223692894}
+{"step": 380, "steps": 500, "lr": 2e-05, "grad_norm": 3.348581314086914, "loss_scale": 4.0, "retries": 0, "seconds": 326.3, "max_memory_gb": 16.44, "pair": 0.37016433104872704, "sign": 0.6029141079634428, "keep": 0.3417027108371258, "leak": 0.8187707103788853, "replacement": 1.047165386378765}
+{"step": 381, "steps": 500, "lr": 2e-05, "grad_norm": 2.08664608001709, "loss_scale": 4.0, "retries": 0, "seconds": 327.2, "max_memory_gb": 16.44, "pair": 0.7430867105722427, "sign": 0.9000866562128067, "keep": 0.4263777658343315, "leak": 0.20971607984392904, "replacement": 1.4587479829788208}
+{"step": 382, "steps": 500, "lr": 2e-05, "grad_norm": 7.171914100646973, "loss_scale": 4.0, "retries": 0, "seconds": 328.0, "max_memory_gb": 16.44, "pair": 0.4926557093858719, "sign": 0.7547450205311179, "keep": 0.7600804776884615, "leak": 1.4790240786969662, "replacement": 1.619824543595314}
+{"step": 383, "steps": 500, "lr": 2e-05, "grad_norm": 4.025608539581299, "loss_scale": 4.0, "retries": 0, "seconds": 328.8, "max_memory_gb": 16.44, "pair": 0.5194468051195145, "sign": 0.4587411545217037, "keep": 0.21470555500127375, "leak": 0.25309598263993394, "replacement": 0.9194441586732864}
+{"step": 384, "steps": 500, "lr": 2e-05, "grad_norm": 7.001031398773193, "loss_scale": 4.0, "retries": 0, "seconds": 329.7, "max_memory_gb": 16.44, "pair": 0.22798607498407364, "sign": 0.1652450105175376, "keep": 0.03167851665057242, "leak": 1.3718887381255627, "replacement": 0.6694200187921524}
+{"step": 385, "steps": 500, "lr": 2e-05, "grad_norm": 1.2360732555389404, "loss_scale": 4.0, "retries": 0, "seconds": 330.6, "max_memory_gb": 16.44, "pair": 0.43302490562200546, "sign": 0.5028251335024834, "keep": 0.3431672124133911, "leak": 0.4284348808287177, "replacement": 0.9631298370659351}
+{"step": 386, "steps": 500, "lr": 2e-05, "grad_norm": 4.321227073669434, "loss_scale": 4.0, "retries": 0, "seconds": 331.4, "max_memory_gb": 16.44, "pair": 0.25887163914740086, "sign": 0.3158154634293169, "keep": 0.17669348791241646, "leak": 1.0923223122954369, "replacement": 0.7782066762447357}
+{"step": 387, "steps": 500, "lr": 2e-05, "grad_norm": 5.804262161254883, "loss_scale": 4.0, "retries": 0, "seconds": 332.3, "max_memory_gb": 16.44, "pair": 0.4580708649009466, "sign": 0.5879090670496225, "keep": 0.23958412883803248, "leak": 0.8730789348483086, "replacement": 1.0900872349739075}
+{"step": 388, "steps": 500, "lr": 2e-05, "grad_norm": 3.9145305156707764, "loss_scale": 4.0, "retries": 0, "seconds": 333.2, "max_memory_gb": 16.44, "pair": 0.2170250304043293, "sign": 0.18225715309381485, "keep": 0.061486406571930274, "leak": 0.9036405310034752, "replacement": 0.5648069232702255}
+{"step": 389, "steps": 500, "lr": 2e-05, "grad_norm": 6.858337879180908, "loss_scale": 4.0, "retries": 0, "seconds": 334.1, "max_memory_gb": 16.44, "pair": 0.3127692639827728, "sign": 0.22326100710779428, "keep": 0.06792328786104918, "leak": 1.8130008056759834, "replacement": 0.9116116166114807}
+{"step": 390, "steps": 500, "lr": 2e-05, "grad_norm": 3.0773589611053467, "loss_scale": 4.0, "retries": 0, "seconds": 334.9, "max_memory_gb": 16.44, "pair": 0.29154195077717304, "sign": 0.31741662696003914, "keep": 0.07978659542277455, "leak": 0.9402590394020081, "replacement": 0.7252083197236061}
+{"step": 391, "steps": 500, "lr": 2e-05, "grad_norm": 2.9817113876342773, "loss_scale": 4.0, "retries": 0, "seconds": 335.8, "max_memory_gb": 16.44, "pair": 0.5889932513237, "sign": 0.7398961931467056, "keep": 0.8358435593545437, "leak": 0.14998370828107, "replacement": 1.414358988404274}
+{"step": 392, "steps": 500, "lr": 2e-05, "grad_norm": 2.5669264793395996, "loss_scale": 4.0, "retries": 0, "seconds": 336.7, "max_memory_gb": 16.44, "pair": 0.14283808134496212, "sign": 0.08596932096406817, "keep": 0.10906415097997524, "leak": 0.7598277777433395, "replacement": 0.43031175434589386}
+{"step": 393, "steps": 500, "lr": 2e-05, "grad_norm": 5.664012432098389, "loss_scale": 4.0, "retries": 0, "seconds": 337.5, "max_memory_gb": 16.44, "pair": 0.4231233671307564, "sign": 0.5027804803103209, "keep": 0.12947908020578325, "leak": 0.4929301039665006, "replacement": 0.8624856621026993}
+{"step": 394, "steps": 500, "lr": 2e-05, "grad_norm": 3.957714080810547, "loss_scale": 4.0, "retries": 0, "seconds": 338.5, "max_memory_gb": 16.44, "pair": 0.4441745448857546, "sign": 0.698105875402689, "keep": 0.13284611818380654, "leak": 0.7321511097252369, "replacement": 1.0426883473992348}
+{"step": 395, "steps": 500, "lr": 2e-05, "grad_norm": 4.213891983032227, "loss_scale": 4.0, "retries": 0, "seconds": 339.4, "max_memory_gb": 16.44, "pair": 0.25168656930327415, "sign": 0.25540851056575775, "keep": 0.03551154988235794, "leak": 1.1527447327971458, "replacement": 0.685332790017128}
+{"step": 396, "steps": 500, "lr": 2e-05, "grad_norm": 2.839977741241455, "loss_scale": 4.0, "retries": 0, "seconds": 340.2, "max_memory_gb": 16.44, "pair": 0.23356042057275772, "sign": 0.09022377245128155, "keep": 0.06703737564384937, "leak": 0.4047149792313576, "replacement": 0.41336973011493683}
+{"step": 397, "steps": 500, "lr": 2e-05, "grad_norm": 1.8690000772476196, "loss_scale": 4.0, "retries": 0, "seconds": 341.1, "max_memory_gb": 16.44, "pair": 0.45845832116901875, "sign": 0.5757544133812189, "keep": 0.35521978372707963, "leak": 0.5229594931006432, "replacement": 1.0546853244304657}
+{"step": 398, "steps": 500, "lr": 2e-05, "grad_norm": 3.6938343048095703, "loss_scale": 4.0, "retries": 0, "seconds": 342.0, "max_memory_gb": 16.44, "pair": 0.395957563072443, "sign": 0.37318708561360836, "keep": 0.17888813227182254, "leak": 0.9267837908118963, "replacement": 0.9036911278963089}
+{"step": 399, "steps": 500, "lr": 2e-05, "grad_norm": 7.9435319900512695, "loss_scale": 4.0, "retries": 0, "seconds": 343.0, "max_memory_gb": 16.44, "pair": 0.4506076257675886, "sign": 0.6042429953813553, "keep": 0.4785745022818446, "leak": 1.2617787718772888, "replacement": 1.3074611276388168}
+{"step": 400, "steps": 500, "lr": 2e-05, "grad_norm": 3.9905989170074463, "loss_scale": 4.0, "retries": 0, "seconds": 343.9, "max_memory_gb": 16.44, "pair": 0.3986760713160038, "sign": 0.40232049115002155, "keep": 0.36886376421898603, "leak": 0.5718486569821835, "replacement": 0.9272303730249405}
+{"step": 401, "steps": 500, "lr": 2e-05, "grad_norm": 7.411082744598389, "loss_scale": 4.0, "retries": 0, "seconds": 344.9, "max_memory_gb": 16.44, "pair": 0.23229235783219337, "sign": 0.5661897175014019, "keep": 0.6377978066302603, "leak": 1.0888971537351608, "replacement": 1.1065104119479656}
+{"step": 402, "steps": 500, "lr": 2e-05, "grad_norm": 2.060931444168091, "loss_scale": 4.0, "retries": 0, "seconds": 345.8, "max_memory_gb": 16.44, "pair": 0.6629023551940918, "sign": 0.8284812718629837, "keep": 0.49793848767876625, "leak": 0.19871922489255667, "replacement": 1.3757920265197754}
+{"step": 403, "steps": 500, "lr": 2e-05, "grad_norm": 5.0218892097473145, "loss_scale": 4.0, "retries": 0, "seconds": 346.6, "max_memory_gb": 16.44, "pair": 0.1885838359594345, "sign": 0.14129131101071835, "keep": 0.017977302573854104, "leak": 1.0136302160099149, "replacement": 0.5216256864368916}
+{"step": 404, "steps": 500, "lr": 2e-05, "grad_norm": 2.3913629055023193, "loss_scale": 4.0, "retries": 0, "seconds": 347.5, "max_memory_gb": 16.44, "pair": 0.35658705420792103, "sign": 0.4065638482570648, "keep": 0.20521807111799717, "leak": 0.469350622035563, "replacement": 0.779815673828125}
+{"step": 405, "steps": 500, "lr": 2e-05, "grad_norm": 3.5107457637786865, "loss_scale": 4.0, "retries": 0, "seconds": 348.4, "max_memory_gb": 16.44, "pair": 0.17927268706262112, "sign": 0.19352862983942032, "keep": 0.11316034485207638, "leak": 1.0489321984350681, "replacement": 0.5948502197861671}
+{"step": 406, "steps": 500, "lr": 2e-05, "grad_norm": 4.779277324676514, "loss_scale": 4.0, "retries": 0, "seconds": 349.2, "max_memory_gb": 16.44, "pair": 0.5840884577482939, "sign": 0.8109020795673132, "keep": 0.3742227666079998, "leak": 1.1943612569011748, "replacement": 1.4752411991357803}
+{"step": 407, "steps": 500, "lr": 2e-05, "grad_norm": 7.293337821960449, "loss_scale": 4.0, "retries": 0, "seconds": 350.1, "max_memory_gb": 16.44, "pair": 0.7895727064460516, "sign": 1.0306150428950787, "keep": 2.4966658977791667, "leak": 0.8408290967345238, "replacement": 2.763420432806015}
+{"step": 408, "steps": 500, "lr": 2e-05, "grad_norm": 4.4906487464904785, "loss_scale": 4.0, "retries": 0, "seconds": 350.9, "max_memory_gb": 16.44, "pair": 0.5132023235782981, "sign": 0.5860455227084458, "keep": 0.7194495497387834, "leak": 1.0155980251729488, "replacement": 1.4198493659496307}
+{"step": 409, "steps": 500, "lr": 2e-05, "grad_norm": 3.020552158355713, "loss_scale": 4.0, "retries": 0, "seconds": 351.8, "max_memory_gb": 16.44, "pair": 0.4287637248635292, "sign": 0.4786321744322777, "keep": 0.39279259584873216, "leak": 0.5826417922507972, "replacement": 1.0101365596055984}
+{"step": 410, "steps": 500, "lr": 2e-05, "grad_norm": 4.013343811035156, "loss_scale": 4.0, "retries": 0, "seconds": 352.6, "max_memory_gb": 16.44, "pair": 0.16896368423476815, "sign": 0.15912377554923296, "keep": 0.30287383048562333, "leak": 0.9028276344761252, "replacement": 0.62566938996315}
+{"step": 411, "steps": 500, "lr": 2e-05, "grad_norm": 6.705026149749756, "loss_scale": 4.0, "retries": 0, "seconds": 353.5, "max_memory_gb": 16.44, "pair": 0.5257521569728851, "sign": 0.9521614126861095, "keep": 1.0778702312527457, "leak": 0.7670483067631721, "replacement": 1.7325300313532352}
+{"step": 412, "steps": 500, "lr": 2e-05, "grad_norm": 4.30256462097168, "loss_scale": 4.0, "retries": 0, "seconds": 354.4, "max_memory_gb": 16.44, "pair": 0.4680285435169935, "sign": 0.6811064090579748, "keep": 0.2620919862674782, "leak": 0.5929968538403045, "replacement": 1.0878770053386688}
+{"step": 413, "steps": 500, "lr": 2e-05, "grad_norm": 4.433744430541992, "loss_scale": 4.0, "retries": 0, "seconds": 355.2, "max_memory_gb": 16.44, "pair": 0.2275179661810398, "sign": 0.18096737004816532, "keep": 0.10152497980743647, "leak": 1.3952007917687297, "replacement": 0.7175643444061279}
+{"step": 414, "steps": 500, "lr": 2e-05, "grad_norm": 3.73179030418396, "loss_scale": 4.0, "retries": 0, "seconds": 356.1, "max_memory_gb": 16.44, "pair": 0.40499135106801987, "sign": 0.3540685949847102, "keep": 0.4245382258668542, "leak": 0.7517240578308702, "replacement": 0.982225812971592}
+{"step": 415, "steps": 500, "lr": 2e-05, "grad_norm": 4.6151580810546875, "loss_scale": 4.0, "retries": 0, "seconds": 356.9, "max_memory_gb": 16.44, "pair": 0.46793902292847633, "sign": 0.4057741276919842, "keep": 1.2423112457618117, "leak": 0.6022677719593048, "replacement": 1.442548707127571}
+{"step": 416, "steps": 500, "lr": 2e-05, "grad_norm": 2.4009552001953125, "loss_scale": 4.0, "retries": 0, "seconds": 357.7, "max_memory_gb": 16.44, "pair": 0.24047434702515602, "sign": 0.195750979706645, "keep": 0.08328973071911605, "leak": 0.24176249839365482, "replacement": 0.44043532758951187}
+{"step": 417, "steps": 500, "lr": 2e-05, "grad_norm": 3.385558605194092, "loss_scale": 4.0, "retries": 0, "seconds": 358.6, "max_memory_gb": 16.44, "pair": 0.7841668650507927, "sign": 1.1337567120790482, "keep": 1.1639998015016317, "leak": 0.5779640860855579, "replacement": 2.0775360763072968}
+{"step": 418, "steps": 500, "lr": 2e-05, "grad_norm": 1.8969939947128296, "loss_scale": 4.0, "retries": 0, "seconds": 359.4, "max_memory_gb": 16.44, "pair": 0.21355325728654861, "sign": 0.19745394214987755, "keep": 0.03414782910840586, "leak": 0.770717867417261, "replacement": 0.5220336094498634}
+{"step": 419, "steps": 500, "lr": 2e-05, "grad_norm": 5.043666839599609, "loss_scale": 4.0, "retries": 0, "seconds": 360.2, "max_memory_gb": 16.44, "pair": 0.39292684383690357, "sign": 0.5159321017563343, "keep": 0.06969250021938933, "leak": 2.3183853067457676, "replacement": 1.2653354704380035}
+{"step": 420, "steps": 500, "lr": 2e-05, "grad_norm": 4.178281307220459, "loss_scale": 4.0, "retries": 0, "seconds": 361.0, "max_memory_gb": 16.44, "pair": 0.35393422842025757, "sign": 0.3890570802614093, "keep": 0.6563747460022569, "leak": 0.7734415708109736, "replacement": 1.0700105503201485}
+{"step": 421, "steps": 500, "lr": 2e-05, "grad_norm": 5.33239221572876, "loss_scale": 4.0, "retries": 0, "seconds": 361.9, "max_memory_gb": 16.44, "pair": 0.28707308880984783, "sign": 0.2729872688651085, "keep": 0.18974834925029427, "leak": 1.8520279377698898, "replacement": 0.9814478904008865}
+{"step": 422, "steps": 500, "lr": 2e-05, "grad_norm": 7.119990825653076, "loss_scale": 4.0, "retries": 0, "seconds": 362.7, "max_memory_gb": 16.44, "pair": 0.4587039425969124, "sign": 0.45516046043485403, "keep": 0.3650458031333983, "leak": 1.292382239131257, "replacement": 1.1919026672840118}
+{"step": 423, "steps": 500, "lr": 2e-05, "grad_norm": 2.5680885314941406, "loss_scale": 4.0, "retries": 0, "seconds": 363.6, "max_memory_gb": 16.44, "pair": 0.7035999372601509, "sign": 0.917926512658596, "keep": 0.6932824850082397, "leak": 0.3580821801442653, "replacement": 1.5987249463796616}
+{"step": 424, "steps": 500, "lr": 2e-05, "grad_norm": 2.753139019012451, "loss_scale": 4.0, "retries": 0, "seconds": 364.4, "max_memory_gb": 16.44, "pair": 0.21686283499002457, "sign": 0.11397766880691051, "keep": 0.1894520027562976, "leak": 0.6392392441630363, "replacement": 0.5283874943852425}
+{"step": 425, "steps": 500, "lr": 2e-05, "grad_norm": 7.9235124588012695, "loss_scale": 4.0, "retries": 0, "seconds": 365.3, "max_memory_gb": 16.44, "pair": 0.2766735441982746, "sign": 0.2581221517175436, "keep": 0.23179428046569228, "leak": 1.4821950271725655, "replacement": 0.8921805024147034}
+{"step": 426, "steps": 500, "lr": 2e-05, "grad_norm": 3.036271572113037, "loss_scale": 4.0, "retries": 0, "seconds": 366.2, "max_memory_gb": 16.44, "pair": 0.6325934324413538, "sign": 0.7001304109580815, "keep": 0.6586119383573532, "leak": 0.623208769572102, "replacement": 1.4677667617797852}
+{"step": 427, "steps": 500, "lr": 2e-05, "grad_norm": 5.831328392028809, "loss_scale": 4.0, "retries": 0, "seconds": 367.0, "max_memory_gb": 16.44, "pair": 0.3158824145793915, "sign": 0.36617568135261536, "keep": 0.08993077068589628, "leak": 1.8894015550613403, "replacement": 1.0162859857082367}
+{"step": 428, "steps": 500, "lr": 2e-05, "grad_norm": 4.408199310302734, "loss_scale": 4.0, "retries": 0, "seconds": 367.9, "max_memory_gb": 16.44, "pair": 0.2285076156258583, "sign": 0.148190313950181, "keep": 0.015866080473642796, "leak": 0.9800635054707527, "replacement": 0.5555516928434372}
+{"step": 429, "steps": 500, "lr": 2e-05, "grad_norm": 8.340577125549316, "loss_scale": 4.0, "retries": 0, "seconds": 368.8, "max_memory_gb": 16.44, "pair": 0.13130428362637758, "sign": 0.07300834660418332, "keep": 0.025237880210625008, "leak": 1.9270725548267365, "replacement": 0.6621955335140228}
+{"step": 430, "steps": 500, "lr": 2e-05, "grad_norm": 3.590510129928589, "loss_scale": 4.0, "retries": 0, "seconds": 369.6, "max_memory_gb": 16.44, "pair": 0.2970183286815882, "sign": 0.2693568370305002, "keep": 0.3342274378519505, "leak": 0.7736475542187691, "replacement": 0.7922223508358002}
+{"step": 431, "steps": 500, "lr": 2e-05, "grad_norm": 5.317432880401611, "loss_scale": 4.0, "retries": 0, "seconds": 370.5, "max_memory_gb": 16.44, "pair": 0.26495759561657906, "sign": 0.09011764917522669, "keep": 0.3415118488774169, "leak": 0.4610162805765867, "replacement": 0.5960264205932617}
+{"step": 432, "steps": 500, "lr": 2e-05, "grad_norm": 4.116678714752197, "loss_scale": 4.0, "retries": 0, "seconds": 371.3, "max_memory_gb": 16.44, "pair": 0.30366989597678185, "sign": 0.24676277860999107, "keep": 0.13671567803248763, "leak": 1.0308631174266338, "replacement": 0.7531248927116394}
+{"step": 433, "steps": 500, "lr": 2e-05, "grad_norm": 3.596151113510132, "loss_scale": 4.0, "retries": 0, "seconds": 372.1, "max_memory_gb": 16.44, "pair": 0.3767848014831543, "sign": 0.3948235549032688, "keep": 0.2196784378029406, "leak": 0.9516734182834625, "replacement": 0.9219541475176811}
+{"step": 434, "steps": 500, "lr": 2e-05, "grad_norm": 2.6121742725372314, "loss_scale": 4.0, "retries": 0, "seconds": 373.0, "max_memory_gb": 16.44, "pair": 0.2077296618372202, "sign": 0.16978151723742485, "keep": 0.005021655080781784, "leak": 0.3034356404095888, "replacement": 0.3709901534020901}
+{"step": 435, "steps": 500, "lr": 2e-05, "grad_norm": 2.7286722660064697, "loss_scale": 4.0, "retries": 0, "seconds": 373.8, "max_memory_gb": 16.44, "pair": 0.4286481402814388, "sign": 0.3674333393573761, "keep": 0.22384925917140208, "leak": 0.3479338064789772, "replacement": 0.8112728893756866}
+{"step": 436, "steps": 500, "lr": 2e-05, "grad_norm": 3.17533016204834, "loss_scale": 4.0, "retries": 0, "seconds": 374.7, "max_memory_gb": 16.44, "pair": 0.35606880486011505, "sign": 0.30190870724618435, "keep": 0.39980168640613556, "leak": 0.45997305400669575, "replacement": 0.8219172805547714}
+{"step": 437, "steps": 500, "lr": 2e-05, "grad_norm": 6.370937347412109, "loss_scale": 4.0, "retries": 0, "seconds": 375.5, "max_memory_gb": 16.44, "pair": 0.536037240177393, "sign": 0.49361753836274147, "keep": 0.42904446460306644, "leak": 0.5151842962950468, "replacement": 1.126164272427559}
+{"step": 438, "steps": 500, "lr": 2e-05, "grad_norm": 2.212644577026367, "loss_scale": 4.0, "retries": 0, "seconds": 376.4, "max_memory_gb": 16.44, "pair": 0.19515586644411087, "sign": 0.11345887929201126, "keep": 0.20976676532882266, "leak": 0.37701001204550266, "replacement": 0.4510211795568466}
+{"step": 439, "steps": 500, "lr": 2e-05, "grad_norm": 7.526273250579834, "loss_scale": 4.0, "retries": 0, "seconds": 377.2, "max_memory_gb": 16.44, "pair": 0.4281202480196953, "sign": 0.47133878339082, "keep": 0.9140780474990606, "leak": 0.23881798516958952, "replacement": 1.1805331781506538}
+{"step": 440, "steps": 500, "lr": 2e-05, "grad_norm": 1.454984188079834, "loss_scale": 4.0, "retries": 0, "seconds": 378.1, "max_memory_gb": 16.44, "pair": 0.31770186126232147, "sign": 0.42055144906044006, "keep": 0.03283159527927637, "leak": 0.8336470121867023, "replacement": 0.7528051137924194}
+{"step": 441, "steps": 500, "lr": 2e-05, "grad_norm": 2.8484561443328857, "loss_scale": 4.0, "retries": 0, "seconds": 378.9, "max_memory_gb": 16.44, "pair": 0.5595102868974209, "sign": 0.6146553512662649, "keep": 0.6289777031925041, "leak": 0.5169057212769985, "replacement": 1.310553252696991}
+{"step": 442, "steps": 500, "lr": 2e-05, "grad_norm": 3.0506396293640137, "loss_scale": 4.0, "retries": 0, "seconds": 379.7, "max_memory_gb": 16.44, "pair": 0.43346258997917175, "sign": 0.4202625956386328, "keep": 0.12113447394222021, "leak": 1.2796864323318005, "replacement": 1.0240827649831772}
+{"step": 443, "steps": 500, "lr": 2e-05, "grad_norm": 1.4440443515777588, "loss_scale": 4.0, "retries": 0, "seconds": 380.6, "max_memory_gb": 16.44, "pair": 0.25854424200952053, "sign": 0.33876803144812584, "keep": 0.09773621917702258, "leak": 0.5680108815431595, "replacement": 0.6187991052865982}
+{"step": 444, "steps": 500, "lr": 2e-05, "grad_norm": 3.6195833683013916, "loss_scale": 4.0, "retries": 0, "seconds": 381.4, "max_memory_gb": 16.44, "pair": 0.32770828902721405, "sign": 0.4659405667334795, "keep": 0.287266852799803, "leak": 0.5509766153991222, "replacement": 0.842056155204773}
+{"step": 445, "steps": 500, "lr": 2e-05, "grad_norm": 4.1260576248168945, "loss_scale": 4.0, "retries": 0, "seconds": 382.3, "max_memory_gb": 16.44, "pair": 0.5970749510452151, "sign": 0.7919061090797186, "keep": 0.7800998028978938, "leak": 1.1906677602091804, "replacement": 1.680744856595993}
+{"step": 446, "steps": 500, "lr": 2e-05, "grad_norm": 4.398200988769531, "loss_scale": 4.0, "retries": 0, "seconds": 383.1, "max_memory_gb": 16.44, "pair": 0.2547470610588789, "sign": 0.21806665509939194, "keep": 0.2855879073613323, "leak": 0.8202798888087273, "replacement": 0.7116443067789078}
+{"step": 447, "steps": 500, "lr": 2e-05, "grad_norm": 2.5162031650543213, "loss_scale": 4.0, "retries": 0, "seconds": 384.0, "max_memory_gb": 16.44, "pair": 0.5465068407356739, "sign": 0.7092879042029381, "keep": 0.7554487472807523, "leak": 0.661726838226059, "replacement": 1.4443068578839302}
+{"step": 448, "steps": 500, "lr": 2e-05, "grad_norm": 3.1080474853515625, "loss_scale": 4.0, "retries": 0, "seconds": 384.8, "max_memory_gb": 16.44, "pair": 0.3480922393500805, "sign": 0.3926831744611263, "keep": 0.7251815533963963, "leak": 0.6319136414676905, "replacement": 1.0650029927492142}
+{"step": 449, "steps": 500, "lr": 2e-05, "grad_norm": 2.232853651046753, "loss_scale": 4.0, "retries": 0, "seconds": 385.6, "max_memory_gb": 16.44, "pair": 0.1013784995302558, "sign": 0.17191719077527523, "keep": 0.05196308216545731, "leak": 0.9208079855889082, "replacement": 0.4435206390917301}
+{"step": 450, "steps": 500, "lr": 2e-05, "grad_norm": 5.313226222991943, "loss_scale": 4.0, "retries": 0, "seconds": 386.5, "max_memory_gb": 16.44, "pair": 0.26443141233175993, "sign": 0.14567757956683636, "keep": 0.11112366922316141, "leak": 1.0217188149690628, "replacement": 0.6482617557048798}
+{"step": 451, "steps": 500, "lr": 2e-05, "grad_norm": 5.273784637451172, "loss_scale": 4.0, "retries": 0, "seconds": 387.3, "max_memory_gb": 16.44, "pair": 0.1559360083192587, "sign": 0.11494983732700348, "keep": 0.02774876578769181, "leak": 0.906047984957695, "replacement": 0.453797310590744}
+{"step": 452, "steps": 500, "lr": 2e-05, "grad_norm": 3.514935255050659, "loss_scale": 4.0, "retries": 0, "seconds": 388.1, "max_memory_gb": 16.44, "pair": 0.4428619295358658, "sign": 0.5062482841312885, "keep": 0.6231861711712554, "leak": 0.8647226095199585, "replacement": 1.2237598448991776}
+{"step": 453, "steps": 500, "lr": 2e-05, "grad_norm": 4.461999416351318, "loss_scale": 4.0, "retries": 0, "seconds": 389.0, "max_memory_gb": 16.44, "pair": 0.2692574243992567, "sign": 0.2733004689216614, "keep": 0.03701971028931439, "leak": 0.9043278281242237, "replacement": 0.6504994854331017}
+{"step": 454, "steps": 500, "lr": 2e-05, "grad_norm": 4.637383460998535, "loss_scale": 4.0, "retries": 0, "seconds": 389.8, "max_memory_gb": 16.44, "pair": 0.3430273309350014, "sign": 0.26992647582665086, "keep": 0.5641533955931664, "leak": 0.48964467365294695, "replacement": 0.8824784532189369}
+{"step": 455, "steps": 500, "lr": 2e-05, "grad_norm": 5.759079933166504, "loss_scale": 4.0, "retries": 0, "seconds": 390.7, "max_memory_gb": 16.44, "pair": 0.27291618660092354, "sign": 0.43259172514081, "keep": 0.7618608735501766, "leak": 0.7580282175913453, "replacement": 1.0596495121717453}
+{"step": 456, "steps": 500, "lr": 2e-05, "grad_norm": 7.912717819213867, "loss_scale": 4.0, "retries": 0, "seconds": 391.5, "max_memory_gb": 16.44, "pair": 0.308264534920454, "sign": 0.7302990974858403, "keep": 0.03518152004107833, "leak": 1.441351069137454, "replacement": 1.0513425841927528}
+{"step": 457, "steps": 500, "lr": 2e-05, "grad_norm": 1.9291549921035767, "loss_scale": 4.0, "retries": 0, "seconds": 392.4, "max_memory_gb": 16.44, "pair": 0.2402223739773035, "sign": 0.3189792279154062, "keep": 0.17285112659737933, "leak": 0.5258911997079849, "replacement": 0.6176103353500366}
+{"step": 458, "steps": 500, "lr": 2e-05, "grad_norm": 3.6425647735595703, "loss_scale": 4.0, "retries": 0, "seconds": 393.2, "max_memory_gb": 16.44, "pair": 0.50572070479393, "sign": 0.583447303622961, "keep": 0.16805209685117006, "leak": 0.5629830937832594, "replacement": 1.0222161710262299}
+{"step": 459, "steps": 500, "lr": 2e-05, "grad_norm": 2.378077983856201, "loss_scale": 4.0, "retries": 0, "seconds": 394.1, "max_memory_gb": 16.44, "pair": 0.33951128646731377, "sign": 0.397917072288692, "keep": 0.14911967515945435, "leak": 0.22745841211872175, "replacement": 0.6698942556977272}
+{"step": 460, "steps": 500, "lr": 2e-05, "grad_norm": 5.3370490074157715, "loss_scale": 4.0, "retries": 0, "seconds": 395.0, "max_memory_gb": 16.44, "pair": 0.32555141020566225, "sign": 0.3324099499732256, "keep": 0.035250813205493614, "leak": 1.2488862723112106, "replacement": 0.8216033428907394}
+{"step": 461, "steps": 500, "lr": 2e-05, "grad_norm": 1.9814633131027222, "loss_scale": 4.0, "retries": 0, "seconds": 395.8, "max_memory_gb": 16.44, "pair": 0.3094922862946987, "sign": 0.5823934711515903, "keep": 0.444995179772377, "leak": 0.823281274875626, "replacement": 1.029006913304329}
+{"step": 462, "steps": 500, "lr": 2e-05, "grad_norm": 3.719169855117798, "loss_scale": 4.0, "retries": 0, "seconds": 396.7, "max_memory_gb": 16.44, "pair": 0.24921117164194584, "sign": 0.27033508010208607, "keep": 0.041835454052488785, "leak": 1.1341430209577084, "replacement": 0.6888322159647942}
+{"step": 463, "steps": 500, "lr": 2e-05, "grad_norm": 2.635803461074829, "loss_scale": 4.0, "retries": 0, "seconds": 397.5, "max_memory_gb": 16.44, "pair": 0.1462012054398656, "sign": 0.16679206676781178, "keep": 0.011531113064847887, "leak": 0.5751542206853628, "replacement": 0.379151351749897}
+{"step": 464, "steps": 500, "lr": 2e-05, "grad_norm": 4.34995174407959, "loss_scale": 4.0, "retries": 0, "seconds": 398.4, "max_memory_gb": 16.44, "pair": 0.1676467228680849, "sign": 0.27157579362392426, "keep": 0.18103502071244293, "leak": 0.6450695842504501, "replacement": 0.5552195236086845}
+{"step": 465, "steps": 500, "lr": 2e-05, "grad_norm": 3.5992395877838135, "loss_scale": 4.0, "retries": 0, "seconds": 399.2, "max_memory_gb": 16.44, "pair": 0.2898836135864258, "sign": 0.41466130316257477, "keep": 0.194861417170614, "leak": 0.9559944719076157, "replacement": 0.8336435705423355}
+{"step": 466, "steps": 500, "lr": 2e-05, "grad_norm": 9.12413215637207, "loss_scale": 4.0, "retries": 0, "seconds": 400.1, "max_memory_gb": 16.44, "pair": 0.30405457131564617, "sign": 0.4216197612695396, "keep": 0.16129540096153505, "leak": 1.6262579555623233, "replacement": 1.0020766481757164}
+{"step": 467, "steps": 500, "lr": 2e-05, "grad_norm": 5.510632038116455, "loss_scale": 4.0, "retries": 0, "seconds": 400.9, "max_memory_gb": 16.44, "pair": 0.14884941093623638, "sign": 0.12392166815698147, "keep": 0.08948599873110652, "leak": 1.3260800205171108, "replacement": 0.5870732516050339}
+{"step": 468, "steps": 500, "lr": 2e-05, "grad_norm": 5.978009223937988, "loss_scale": 4.0, "retries": 0, "seconds": 401.8, "max_memory_gb": 16.44, "pair": 0.15086373314261436, "sign": 0.15270334295928478, "keep": 0.019776380067924038, "leak": 1.00201216340065, "replacement": 0.48760663717985153}
+{"step": 469, "steps": 500, "lr": 2e-05, "grad_norm": 4.239388942718506, "loss_scale": 4.0, "retries": 0, "seconds": 402.6, "max_memory_gb": 16.44, "pair": 0.21277421712875366, "sign": 0.13607256673276424, "keep": 0.3355360833229497, "leak": 0.6485113333910704, "replacement": 0.6107063740491867}
+{"step": 470, "steps": 500, "lr": 2e-05, "grad_norm": 4.192856311798096, "loss_scale": 4.0, "retries": 0, "seconds": 403.4, "max_memory_gb": 16.44, "pair": 0.14664147421717644, "sign": 0.0686575286090374, "keep": 0.007040338547085412, "leak": 0.4548677168786526, "replacement": 0.2982073351740837}
+{"step": 471, "steps": 500, "lr": 2e-05, "grad_norm": 3.688117265701294, "loss_scale": 4.0, "retries": 0, "seconds": 404.3, "max_memory_gb": 16.44, "pair": 0.4281265325844288, "sign": 0.6972269089892507, "keep": 0.41344629088416696, "leak": 0.37791604100493714, "replacement": 1.077942132949829}
+{"step": 472, "steps": 500, "lr": 2e-05, "grad_norm": 5.017667293548584, "loss_scale": 4.0, "retries": 0, "seconds": 405.1, "max_memory_gb": 16.44, "pair": 0.29049904318526387, "sign": 0.2346911933273077, "keep": 0.11161781955161132, "leak": 0.6026115342974663, "replacement": 0.6143064349889755}
+{"step": 473, "steps": 500, "lr": 2e-05, "grad_norm": 4.330976486206055, "loss_scale": 4.0, "retries": 0, "seconds": 405.9, "max_memory_gb": 16.44, "pair": 0.2474612519145012, "sign": 0.19373969361186028, "keep": 0.09528249854338355, "leak": 0.3509415378794074, "replacement": 0.4797077439725399}
+{"step": 474, "steps": 500, "lr": 2e-05, "grad_norm": 2.4384663105010986, "loss_scale": 4.0, "retries": 0, "seconds": 406.8, "max_memory_gb": 16.44, "pair": 0.2783779362216592, "sign": 0.3861595280468464, "keep": 0.12923531742126215, "leak": 0.1521123064449057, "replacement": 0.5741034112870693}
+{"step": 475, "steps": 500, "lr": 2e-05, "grad_norm": 2.0885672569274902, "loss_scale": 4.0, "retries": 0, "seconds": 407.7, "max_memory_gb": 16.44, "pair": 0.2794621246866882, "sign": 0.3961640875786543, "keep": 0.14009221484593581, "leak": 0.5016879723407328, "replacement": 0.6730122789740562}
+{"step": 476, "steps": 500, "lr": 2e-05, "grad_norm": 4.033813953399658, "loss_scale": 4.0, "retries": 0, "seconds": 408.5, "max_memory_gb": 16.44, "pair": 0.19690553192049265, "sign": 0.2602649014443159, "keep": 0.2113253073184751, "leak": 1.3379792124032974, "replacement": 0.767195425927639}
+{"step": 477, "steps": 500, "lr": 2e-05, "grad_norm": 7.844718933105469, "loss_scale": 4.0, "retries": 0, "seconds": 409.3, "max_memory_gb": 16.44, "pair": 0.3028308469802141, "sign": 0.44853852968662977, "keep": 0.020679327586549334, "leak": 1.3486680500209332, "replacement": 0.8746068105101585}
+{"step": 478, "steps": 500, "lr": 2e-05, "grad_norm": 4.6538567543029785, "loss_scale": 4.0, "retries": 0, "seconds": 410.2, "max_memory_gb": 16.44, "pair": 0.1758253090083599, "sign": 0.20638774475082755, "keep": 0.00421822392672766, "leak": 1.3157616085372865, "replacement": 0.610068678855896}
+{"step": 479, "steps": 500, "lr": 2e-05, "grad_norm": 6.416682720184326, "loss_scale": 4.0, "retries": 0, "seconds": 411.1, "max_memory_gb": 16.44, "pair": 0.5821870034560561, "sign": 0.5703279115259647, "keep": 0.7038102697115391, "leak": 0.4884977429173887, "replacement": 1.341380551457405}
+{"step": 480, "steps": 500, "lr": 2e-05, "grad_norm": 2.9193904399871826, "loss_scale": 4.0, "retries": 0, "seconds": 411.9, "max_memory_gb": 16.44, "pair": 0.5438872464001179, "sign": 0.5718398671597242, "keep": 0.6329043470323086, "leak": 0.22317139129154384, "replacement": 1.202052265405655}
+{"step": 481, "steps": 500, "lr": 2e-05, "grad_norm": 6.079529285430908, "loss_scale": 4.0, "retries": 0, "seconds": 412.8, "max_memory_gb": 16.44, "pair": 0.23596717324107885, "sign": 0.19412934011779726, "keep": 0.036840006556303706, "leak": 1.232866058970103, "replacement": 0.659668356180191}
+{"step": 482, "steps": 500, "lr": 2e-05, "grad_norm": 1.8104209899902344, "loss_scale": 4.0, "retries": 0, "seconds": 413.6, "max_memory_gb": 16.44, "pair": 0.3585272068157792, "sign": 0.46210511587560177, "keep": 0.21653421490918845, "leak": 0.9248418443603441, "replacement": 0.9290573224425316}
+{"step": 483, "steps": 500, "lr": 2e-05, "grad_norm": 3.7509982585906982, "loss_scale": 4.0, "retries": 0, "seconds": 414.4, "max_memory_gb": 16.44, "pair": 0.4839678704738617, "sign": 0.7393349856138229, "keep": 0.46036656433716416, "leak": 0.17451717355288565, "replacement": 1.127447932958603}
+{"step": 484, "steps": 500, "lr": 2e-05, "grad_norm": 3.981421709060669, "loss_scale": 4.0, "retries": 0, "seconds": 415.3, "max_memory_gb": 16.44, "pair": 0.5642735213041306, "sign": 0.8532850090414286, "keep": 0.803492627106607, "leak": 0.36199864000082016, "replacement": 1.4831620156764984}
+{"step": 485, "steps": 500, "lr": 2e-05, "grad_norm": 3.951927661895752, "loss_scale": 4.0, "retries": 0, "seconds": 416.1, "max_memory_gb": 16.44, "pair": 0.4772937949746847, "sign": 0.8616871647536755, "keep": 0.35799469612538815, "leak": 1.2382236495614052, "replacement": 1.3966906443238258}
+{"step": 486, "steps": 500, "lr": 2e-05, "grad_norm": 4.966502666473389, "loss_scale": 4.0, "retries": 0, "seconds": 416.9, "max_memory_gb": 16.44, "pair": 0.4713774770498276, "sign": 0.6098711788654327, "keep": 0.5657203278969973, "leak": 0.5099848806858063, "replacement": 1.186669409275055}
+{"step": 487, "steps": 500, "lr": 2e-05, "grad_norm": 4.067269325256348, "loss_scale": 4.0, "retries": 0, "seconds": 417.8, "max_memory_gb": 16.44, "pair": 0.31584956869482994, "sign": 0.21994465589523315, "keep": 0.049760772148147225, "leak": 1.2906504608690739, "replacement": 0.7733649015426636}
+{"step": 488, "steps": 500, "lr": 2e-05, "grad_norm": 2.7090113162994385, "loss_scale": 4.0, "retries": 0, "seconds": 418.6, "max_memory_gb": 16.44, "pair": 0.3117199093103409, "sign": 0.5983247496187687, "keep": 0.6135835871100426, "leak": 0.9361413139849901, "replacement": 1.1517093852162361}
+{"step": 489, "steps": 500, "lr": 2e-05, "grad_norm": 4.212612628936768, "loss_scale": 4.0, "retries": 0, "seconds": 419.5, "max_memory_gb": 16.44, "pair": 0.2843818739056587, "sign": 0.4898154307156801, "keep": 0.8427286883888883, "leak": 1.1900021396577358, "replacement": 1.248154442757368}
+{"step": 490, "steps": 500, "lr": 2e-05, "grad_norm": 8.637441635131836, "loss_scale": 4.0, "retries": 0, "seconds": 420.3, "max_memory_gb": 16.44, "pair": 0.21404758282005787, "sign": 0.4010673644952476, "keep": 0.05667542282026261, "leak": 1.6786685238112113, "replacement": 0.862586110830307}
+{"step": 491, "steps": 500, "lr": 2e-05, "grad_norm": 6.897251129150391, "loss_scale": 4.0, "retries": 0, "seconds": 421.1, "max_memory_gb": 16.44, "pair": 0.12376144574955106, "sign": 0.2641478516161442, "keep": 0.13766503903025296, "leak": 3.263816799968481, "replacement": 1.1406220979988575}
+{"step": 492, "steps": 500, "lr": 2e-05, "grad_norm": 4.012920379638672, "loss_scale": 4.0, "retries": 0, "seconds": 422.0, "max_memory_gb": 16.44, "pair": 0.1866952758282423, "sign": 0.13635973446071148, "keep": 0.0075836601899936795, "leak": 1.0379877537488937, "replacement": 0.5181639268994331}
+{"step": 493, "steps": 500, "lr": 2e-05, "grad_norm": 3.5655691623687744, "loss_scale": 4.0, "retries": 0, "seconds": 422.8, "max_memory_gb": 16.44, "pair": 0.6726234182715416, "sign": 0.8477877825498581, "keep": 0.4638275261459057, "leak": 0.23529422339652228, "replacement": 1.3872546255588531}
+{"step": 494, "steps": 500, "lr": 2e-05, "grad_norm": 6.058865547180176, "loss_scale": 4.0, "retries": 0, "seconds": 423.7, "max_memory_gb": 16.44, "pair": 0.11623796727508307, "sign": 0.23041696613654494, "keep": 0.037152229109779, "leak": 1.537511058151722, "replacement": 0.634400337934494}
+{"step": 495, "steps": 500, "lr": 2e-05, "grad_norm": 5.009451866149902, "loss_scale": 4.0, "retries": 0, "seconds": 424.5, "max_memory_gb": 16.44, "pair": 0.20370196923613548, "sign": 0.29365446511656046, "keep": 0.09312315774150193, "leak": 0.6691274866461754, "replacement": 0.564372643828392}
+{"step": 496, "steps": 500, "lr": 2e-05, "grad_norm": 5.238865852355957, "loss_scale": 4.0, "retries": 0, "seconds": 425.4, "max_memory_gb": 16.44, "pair": 0.14743565768003464, "sign": 0.05439091566950083, "keep": 0.0834796231938526, "leak": 1.5585724264383316, "replacement": 0.6060140430927277}
+{"step": 497, "steps": 500, "lr": 2e-05, "grad_norm": 6.222712516784668, "loss_scale": 4.0, "retries": 0, "seconds": 426.2, "max_memory_gb": 16.44, "pair": 0.3647892950102687, "sign": 0.632240878418088, "keep": 1.1773772499291226, "leak": 1.2711927661002846, "replacement": 1.5873964820057154}
+{"step": 498, "steps": 500, "lr": 2e-05, "grad_norm": 5.684950828552246, "loss_scale": 4.0, "retries": 0, "seconds": 427.0, "max_memory_gb": 16.44, "pair": 0.4218146875500679, "sign": 0.40541674196720123, "keep": 0.204066849546507, "leak": 1.088819329826947, "replacement": 0.9987613409757614}
+{"step": 499, "steps": 500, "lr": 2e-05, "grad_norm": 2.804603099822998, "loss_scale": 4.0, "retries": 0, "seconds": 427.9, "max_memory_gb": 16.44, "pair": 0.3892014492303133, "sign": 0.3517630463466048, "keep": 0.15840869807288982, "leak": 0.3559276550076902, "replacement": 0.7332692444324493}
+{"step": 500, "steps": 500, "lr": 2e-05, "grad_norm": 5.017227649688721, "loss_scale": 4.0, "retries": 0, "seconds": 428.7, "max_memory_gb": 16.44, "pair": 0.35522397607564926, "sign": 0.49589458107948303, "keep": 0.5817540396237746, "leak": 0.8200249494984746, "replacement": 1.0990545339882374}