Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- runs/single_internvl/context/L01_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L01_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L01_s13/config.json +24 -0
- runs/single_internvl/context/L01_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L02_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L02_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L02_s13/config.json +24 -0
- runs/single_internvl/context/L02_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L07_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L07_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L07_s13/config.json +24 -0
- runs/single_internvl/context/L07_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L09_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L09_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L09_s13/config.json +24 -0
- runs/single_internvl/context/L09_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L11_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L11_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L11_s13/config.json +24 -0
- runs/single_internvl/context/L11_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L16_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L16_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L16_s13/config.json +24 -0
- runs/single_internvl/context/L16_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L20_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L20_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L20_s13/config.json +24 -0
- runs/single_internvl/context/L20_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L25_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L25_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L25_s13/config.json +24 -0
- runs/single_internvl/context/L25_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L29_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L29_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L29_s13/config.json +24 -0
- runs/single_internvl/context/L29_s13/steps.jsonl +250 -0
- runs/single_internvl/context/L34_s13/adapter/README.md +207 -0
- runs/single_internvl/context/L34_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/context/L34_s13/config.json +24 -0
- runs/single_internvl/context/L34_s13/steps.jsonl +250 -0
- runs/single_internvl/replacement/L01_s13/adapter/README.md +207 -0
- runs/single_internvl/replacement/L01_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/replacement/L01_s13/config.json +24 -0
- runs/single_internvl/replacement/L01_s13/steps.jsonl +250 -0
- runs/single_internvl/replacement/L02_s13/adapter/README.md +207 -0
- runs/single_internvl/replacement/L02_s13/adapter/adapter_config.json +50 -0
- runs/single_internvl/replacement/L02_s13/config.json +24 -0
- runs/single_internvl/replacement/L02_s13/steps.jsonl +250 -0
- runs/single_internvl/replacement/L07_s13/adapter/README.md +207 -0
- runs/single_internvl/replacement/L07_s13/adapter/adapter_config.json +50 -0
runs/single_internvl/context/L01_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L01_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.1.self_attn.k_proj",
|
| 35 |
+
"model.language_model.layers.1.mlp.down_proj",
|
| 36 |
+
"model.language_model.layers.1.mlp.up_proj",
|
| 37 |
+
"model.language_model.layers.1.self_attn.q_proj",
|
| 38 |
+
"model.language_model.layers.1.mlp.gate_proj",
|
| 39 |
+
"model.language_model.layers.1.self_attn.v_proj",
|
| 40 |
+
"model.language_model.layers.1.self_attn.o_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L01_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 1,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L01_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.41529902815818787, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 18.94, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.630935549736023, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 18.95, "ce": 0.6404551289961091, "kl": 0.4672839170342513, "context": 0.8740971157312742}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 1.6668903827667236, "loss_scale": 4.0, "retries": 0, "seconds": 2.5, "max_memory_gb": 18.95, "ce": 1.096772275865078, "kl": 0.06840587593615055, "context": 1.1309751942753792}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 0.8812512755393982, "loss_scale": 4.0, "retries": 0, "seconds": 3.0, "max_memory_gb": 18.95, "ce": 0.45131903726723976, "kl": 0.42920650115434, "context": 0.6659222840607981}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.017900390550494194, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 18.95, "ce": 0.009307772110332735, "kl": 0.008224977031204617, "context": 0.013420260773273185}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.4201781153678894, "loss_scale": 4.0, "retries": 0, "seconds": 4.0, "max_memory_gb": 18.95, "ce": 0.2907808331801789, "kl": 0.07358878948252823, "context": 0.3275752314075362}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 1.3890644311904907, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 18.95, "ce": 0.8263224340080342, "kl": 0.12455667414030813, "context": 0.8886007743385562}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 1.3712068796157837, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 18.96, "ce": 0.7885623220354319, "kl": 0.28962204651907086, "context": 0.9333733469247818}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 0.9882053732872009, "loss_scale": 4.0, "retries": 0, "seconds": 5.5, "max_memory_gb": 18.96, "ce": 0.5176718682050705, "kl": 0.13548408180940896, "context": 0.5854139216244221}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.17280255258083344, "loss_scale": 4.0, "retries": 0, "seconds": 6.0, "max_memory_gb": 18.96, "ce": 0.05230281721014762, "kl": 0.04618364811176434, "context": 0.07539463957073167}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 1.0913310050964355, "loss_scale": 4.0, "retries": 0, "seconds": 6.5, "max_memory_gb": 18.96, "ce": 0.6889468172521447, "kl": 0.004747790688952591, "context": 0.6913207115358091}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 0.8883460760116577, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 18.96, "ce": 0.4559845983703781, "kl": 0.0027409624003666977, "context": 0.45735508319012297}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 0.5871704816818237, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 18.96, "ce": 0.41896960307713016, "kl": 0.019890877421858022, "context": 0.4289150576769316}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 1.4011739492416382, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 18.96, "ce": 0.7459917217493057, "kl": 0.9444645934272557, "context": 1.2182240188121796}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 1.2687474489212036, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 18.96, "ce": 0.4896892928145462, "kl": 0.46457273134967636, "context": 0.7219756473061807}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.40850716829299927, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 18.96, "ce": 0.20205964230262907, "kl": 0.00092566255057136, "context": 0.20252247389726108}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.5144745707511902, "loss_scale": 4.0, "retries": 0, "seconds": 9.5, "max_memory_gb": 18.96, "ce": 0.3119380276184529, "kl": 0.8303875376586802, "context": 0.7271318020066246}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.3989078998565674, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 18.96, "ce": 0.505332664353773, "kl": 0.20895072488929145, "context": 0.6098080199299147}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.441981315612793, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 18.96, "ce": 0.7192528173327446, "kl": 0.16943835164420307, "context": 0.8039719834923744}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.03257448598742485, "loss_scale": 4.0, "retries": 0, "seconds": 11.1, "max_memory_gb": 18.96, "ce": 0.0024086263565550325, "kl": 0.020944038364177686, "context": 0.012880645335826557}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.2985447645187378, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 18.96, "ce": 0.33468019276460836, "kl": 0.04821694786005537, "context": 0.3587886760069523}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 1.2551078796386719, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 18.96, "ce": 0.35539993245038204, "kl": 0.03179829720102134, "context": 0.37129907045164146}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 1.625778317451477, "loss_scale": 4.0, "retries": 0, "seconds": 12.6, "max_memory_gb": 18.96, "ce": 0.5366190668974014, "kl": 0.9957966441995723, "context": 1.0345173653067832}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 1.2230366468429565, "loss_scale": 4.0, "retries": 0, "seconds": 13.1, "max_memory_gb": 18.96, "ce": 0.2796066291630268, "kl": 0.30150755820795894, "context": 0.43036041036248207}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.581318974494934, "loss_scale": 4.0, "retries": 0, "seconds": 13.6, "max_memory_gb": 18.96, "ce": 0.6545408715210215, "kl": 0.20354427956044674, "context": 0.7563130296766758}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 1.482092022895813, "loss_scale": 4.0, "retries": 0, "seconds": 14.1, "max_memory_gb": 18.96, "ce": 0.2845962792634964, "kl": 0.7069850598927587, "context": 0.6380888223648071}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 1.6222131252288818, "loss_scale": 4.0, "retries": 0, "seconds": 14.6, "max_memory_gb": 18.96, "ce": 0.37228576313646045, "kl": 0.4234901958025148, "context": 0.5840308593469672}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.95625901222229, "loss_scale": 4.0, "retries": 0, "seconds": 15.1, "max_memory_gb": 18.96, "ce": 0.6020690612494946, "kl": 0.9375541990157217, "context": 1.0708461254835129}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 1.3546438217163086, "loss_scale": 4.0, "retries": 0, "seconds": 15.6, "max_memory_gb": 18.96, "ce": 0.2546321381814778, "kl": 0.3638458698987961, "context": 0.43655507266521454}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 1.3015401363372803, "loss_scale": 4.0, "retries": 0, "seconds": 16.2, "max_memory_gb": 18.96, "ce": 0.3217579457723332, "kl": 0.011077957261477422, "context": 0.3272969345034653}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 2.178178548812866, "loss_scale": 4.0, "retries": 0, "seconds": 16.7, "max_memory_gb": 18.96, "ce": 0.6041071526706219, "kl": 0.15398449823260307, "context": 0.6810994036495686}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 2.1632025241851807, "loss_scale": 4.0, "retries": 0, "seconds": 17.2, "max_memory_gb": 18.96, "ce": 0.3851197846233845, "kl": 0.017631475086091086, "context": 0.39393552392721176}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.823045551776886, "loss_scale": 4.0, "retries": 0, "seconds": 17.7, "max_memory_gb": 18.96, "ce": 0.21530713699758053, "kl": 0.06403093424160033, "context": 0.2473226049914956}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.9138999581336975, "loss_scale": 4.0, "retries": 0, "seconds": 18.2, "max_memory_gb": 18.96, "ce": 0.15053801928115718, "kl": 0.768636640906152, "context": 0.5348563246589038}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.9930412769317627, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 18.96, "ce": 0.17215012093947735, "kl": 0.003027120965271024, "context": 0.17366368159127887}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.7445875406265259, "loss_scale": 4.0, "retries": 0, "seconds": 19.2, "max_memory_gb": 18.96, "ce": 0.21996188519551652, "kl": 0.056002757817623205, "context": 0.24796326718933415}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.8231956958770752, "loss_scale": 4.0, "retries": 0, "seconds": 19.7, "max_memory_gb": 18.96, "ce": 0.31887743176775984, "kl": 0.7800079201115295, "context": 0.708881382481195}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.8759904503822327, "loss_scale": 4.0, "retries": 0, "seconds": 20.2, "max_memory_gb": 18.96, "ce": 0.13414412178099155, "kl": 0.5065826356876642, "context": 0.3874354362487793}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.35606151819229126, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 18.96, "ce": 0.17632671473256778, "kl": 0.012248083497979678, "context": 0.18245075378217734}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.6511844992637634, "loss_scale": 4.0, "retries": 0, "seconds": 21.2, "max_memory_gb": 18.96, "ce": 0.08877622777072247, "kl": 0.012029766716295853, "context": 0.09479111022665165}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 0.8662262558937073, "loss_scale": 4.0, "retries": 0, "seconds": 21.7, "max_memory_gb": 18.96, "ce": 0.40725398319773376, "kl": 0.12376161874271929, "context": 0.46913477778434753}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.36961913108825684, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 18.96, "ce": 0.03265774571627844, "kl": 0.0012094670764781768, "context": 0.033262478609685786}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.653836190700531, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 18.96, "ce": 0.0630049918036093, "kl": 0.0900785189223825, "context": 0.10804425216338132}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 1.9988741874694824, "loss_scale": 4.0, "retries": 0, "seconds": 23.2, "max_memory_gb": 18.96, "ce": 0.6567837692273315, "kl": 1.406298268258979, "context": 1.3599329182470683}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 0.7796401977539062, "loss_scale": 4.0, "retries": 0, "seconds": 23.7, "max_memory_gb": 18.96, "ce": 0.9091887520044111, "kl": 0.24430093786213547, "context": 1.0313391968375072}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 4.715503692626953, "loss_scale": 4.0, "retries": 0, "seconds": 24.2, "max_memory_gb": 18.96, "ce": 0.1277937963604927, "kl": 0.2929844274185598, "context": 0.2742860112339258}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.527878999710083, "loss_scale": 4.0, "retries": 0, "seconds": 24.7, "max_memory_gb": 18.96, "ce": 0.04322355194017291, "kl": 0.05331542179919779, "context": 0.06988126377109438}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.8197498917579651, "loss_scale": 4.0, "retries": 0, "seconds": 25.2, "max_memory_gb": 18.96, "ce": 0.12411393629736267, "kl": 0.009469441041801474, "context": 0.1288486560806632}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.6398574709892273, "loss_scale": 4.0, "retries": 0, "seconds": 25.7, "max_memory_gb": 18.96, "ce": 0.06697774384883814, "kl": 0.37099982965503386, "context": 0.2524776642640063}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.8370528817176819, "loss_scale": 4.0, "retries": 0, "seconds": 26.2, "max_memory_gb": 18.96, "ce": 0.1866778254516248, "kl": 0.008382773223274853, "context": 0.1908692189899739}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.5910791158676147, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 18.96, "ce": 0.6366300064655661, "kl": 0.021767326145891275, "context": 0.6475136970075255}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.9754644632339478, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 18.96, "ce": 0.18621333510964178, "kl": 0.2591856258222833, "context": 0.31580613972619176}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.4058437645435333, "loss_scale": 4.0, "retries": 0, "seconds": 27.7, "max_memory_gb": 18.96, "ce": 0.1761016531381756, "kl": 0.4517017435282469, "context": 0.40195252373814583}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.9106970429420471, "loss_scale": 4.0, "retries": 0, "seconds": 28.2, "max_memory_gb": 18.96, "ce": 0.05733488214900717, "kl": 0.3693843043874949, "context": 0.24202703847549856}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.054107196629047394, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 18.96, "ce": 0.016388643824029714, "kl": 0.015054533443276341, "context": 0.02391591010382399}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.33712777495384216, "loss_scale": 4.0, "retries": 0, "seconds": 29.2, "max_memory_gb": 18.96, "ce": 0.04439991433173418, "kl": 0.30272785876877606, "context": 0.19576384127140045}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.16109968721866608, "loss_scale": 4.0, "retries": 0, "seconds": 29.7, "max_memory_gb": 18.96, "ce": 0.016743099025916308, "kl": 0.019652200004202314, "context": 0.026569199966615997}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.30084240436553955, "loss_scale": 4.0, "retries": 0, "seconds": 30.2, "max_memory_gb": 18.96, "ce": 0.05128277537005488, "kl": 0.00283143303526856, "context": 0.05269849352771416}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.1316792070865631, "loss_scale": 4.0, "retries": 0, "seconds": 30.7, "max_memory_gb": 18.96, "ce": 0.02450391925231088, "kl": 0.04635546883218922, "context": 0.04768165361019783}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.1019187793135643, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 18.96, "ce": 0.0720350428018719, "kl": 0.061443222453817725, "context": 0.10275665618246421}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.6655155420303345, "loss_scale": 4.0, "retries": 0, "seconds": 31.7, "max_memory_gb": 18.96, "ce": 0.043357341550290585, "kl": 0.9950835090130568, "context": 0.5408990979194641}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.22032321989536285, "loss_scale": 4.0, "retries": 0, "seconds": 32.2, "max_memory_gb": 18.96, "ce": 0.04638195445295423, "kl": 0.009682625415734947, "context": 0.051223268295871094}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.14040741324424744, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 18.96, "ce": 0.03994716424494982, "kl": 0.1237264961891924, "context": 0.10181041387841105}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.1458318680524826, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 18.96, "ce": 0.026001896942034364, "kl": 0.014457380457315594, "context": 0.0332305864430964}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.5951931476593018, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 18.96, "ce": 0.2044749795459211, "kl": 0.1739400397054851, "context": 0.2914449982345104}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.4100831151008606, "loss_scale": 4.0, "retries": 0, "seconds": 34.3, "max_memory_gb": 18.96, "ce": 0.1610533402120211, "kl": 0.07454902027794219, "context": 0.1983278503364545}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.4895205497741699, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 18.96, "ce": 0.04390105959464563, "kl": 0.04947941959835589, "context": 0.0686407700413838}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.20726370811462402, "loss_scale": 4.0, "retries": 0, "seconds": 35.3, "max_memory_gb": 18.96, "ce": 0.01710079710392165, "kl": 0.054022367559809936, "context": 0.04411198027264618}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.5213683843612671, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 18.96, "ce": 0.2848783084191382, "kl": 0.04515450727194548, "context": 0.3074555713683367}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.20511581003665924, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 18.96, "ce": 0.014576786779798567, "kl": 0.001354803407593863, "context": 0.015254188503604382}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.4987012445926666, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 18.96, "ce": 0.04872734340278839, "kl": 0.016553360444959253, "context": 0.05700402316142572}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.29828161001205444, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 18.96, "ce": 0.03794885822571814, "kl": 0.43808947689831257, "context": 0.25699360109865665}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.4907500743865967, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 18.96, "ce": 0.4175184343002911, "kl": 0.15271727762228693, "context": 0.4938770731241675}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.6462470889091492, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 18.96, "ce": 0.10638844376308043, "kl": 0.12937032300396822, "context": 0.17107360503359814}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 0.8233071565628052, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 18.96, "ce": 0.4431042771029752, "kl": 1.0514792064204812, "context": 0.9688438987359405}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 0.43098267912864685, "loss_scale": 4.0, "retries": 0, "seconds": 39.3, "max_memory_gb": 18.96, "ce": 0.021766084362752736, "kl": 1.8231912070768885, "context": 0.9333617084193975}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.056420572102069855, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 18.96, "ce": 0.009407816745806485, "kl": 0.013229883918029373, "context": 0.016022759169572964}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.13427695631980896, "loss_scale": 4.0, "retries": 0, "seconds": 40.3, "max_memory_gb": 18.96, "ce": 0.0035367312957532704, "kl": 0.0742490008706227, "context": 0.04066123062511906}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.5610014200210571, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 18.96, "ce": 0.05136837621466839, "kl": 0.06963474837357353, "context": 0.08618575041600707}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.18953104317188263, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 18.96, "ce": 0.011239749489504902, "kl": 0.08980649053273737, "context": 0.056142995454138145}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.5339422821998596, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 18.96, "ce": 0.11582899279892445, "kl": 0.09279448608867824, "context": 0.16222624038346112}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.30350780487060547, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 18.96, "ce": 0.03838304668897763, "kl": 0.010404451764770783, "context": 0.04358527215663344}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.3309577703475952, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 18.96, "ce": 0.17120755137875676, "kl": 0.8258629122865386, "context": 0.5841390072600916}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.22325891256332397, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 18.96, "ce": 0.021560225426128454, "kl": 0.05866225290355942, "context": 0.050891352274447854}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 2.0988142490386963, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 18.96, "ce": 0.09791610882223267, "kl": 0.08249838096307371, "context": 0.13916529930486377}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.6013784408569336, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 18.96, "ce": 0.26567687664646655, "kl": 0.5634722396498546, "context": 0.5474129817448556}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.78244948387146, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 18.96, "ce": 0.01142694233567454, "kl": 0.7967089731246233, "context": 0.4097814252600074}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.08671758323907852, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 18.96, "ce": 0.007394502040824591, "kl": 0.033042527527868515, "context": 0.02391576668514972}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.10365752130746841, "loss_scale": 4.0, "retries": 0, "seconds": 45.8, "max_memory_gb": 18.96, "ce": 0.010549725739792848, "kl": 0.00832796507893363, "context": 0.014713708106683043}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.20420098304748535, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 18.96, "ce": 0.022955902233661618, "kl": 0.012992301140911877, "context": 0.029452052694978192}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 0.4648401141166687, "loss_scale": 4.0, "retries": 0, "seconds": 46.8, "max_memory_gb": 18.96, "ce": 0.17000147222279338, "kl": 0.6714539211006922, "context": 0.5057284409231215}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.5578516721725464, "loss_scale": 4.0, "retries": 0, "seconds": 47.4, "max_memory_gb": 18.96, "ce": 0.4453744730872131, "kl": 0.5221160057762972, "context": 0.7064324468356062}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.2088490724563599, "loss_scale": 4.0, "retries": 0, "seconds": 47.9, "max_memory_gb": 18.96, "ce": 0.0703731972607784, "kl": 0.5181256934301928, "context": 0.3294360414147377}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.6625189781188965, "loss_scale": 4.0, "retries": 0, "seconds": 48.4, "max_memory_gb": 18.96, "ce": 0.32796623674221337, "kl": 0.15993681442341767, "context": 0.40793465985916555}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.8752552270889282, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 18.96, "ce": 0.6420130477054045, "kl": 0.029452618095092475, "context": 0.6567393448203802}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 0.7073447704315186, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 18.96, "ce": 0.534109107684344, "kl": 0.00568621646380052, "context": 0.5369522353867069}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 2.61953067779541, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 18.96, "ce": 0.02887815331632737, "kl": 0.4844901212491095, "context": 0.2711232090368867}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.08754952251911163, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 18.96, "ce": 0.0049501672610858805, "kl": 0.0988186548856902, "context": 0.05435949303864618}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.3461315333843231, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 18.96, "ce": 0.03453727294208875, "kl": 0.013974352174045634, "context": 0.041524449679855024}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.6663182973861694, "loss_scale": 4.0, "retries": 0, "seconds": 51.4, "max_memory_gb": 18.96, "ce": 0.028162088012322783, "kl": 0.8004364707667264, "context": 0.42838032252620906}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.28902667760849, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 18.96, "ce": 0.006148882530396804, "kl": 0.0995752906310372, "context": 0.055936527147423476}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.11631490290164948, "loss_scale": 4.0, "retries": 0, "seconds": 52.4, "max_memory_gb": 18.96, "ce": 0.04690119909355417, "kl": 0.039179587038233876, "context": 0.06649099406786263}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.1967485398054123, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 18.96, "ce": 0.019680983417856623, "kl": 0.02022741342352674, "context": 0.02979468982448452}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.046305056661367416, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 18.96, "ce": 0.003638441958173644, "kl": 0.0056495059111227874, "context": 0.006463194695243146}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.21142996847629547, "loss_scale": 4.0, "retries": 0, "seconds": 54.0, "max_memory_gb": 18.96, "ce": 0.06746011739596725, "kl": 0.38595246884506196, "context": 0.2604363556019962}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.16882707178592682, "loss_scale": 4.0, "retries": 0, "seconds": 54.5, "max_memory_gb": 18.96, "ce": 0.009285319996706676, "kl": 0.009808907114347676, "context": 0.014189773464750033}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 1.6739987134933472, "loss_scale": 4.0, "retries": 0, "seconds": 55.0, "max_memory_gb": 18.96, "ce": 0.8426965724029287, "kl": 0.08854798271886466, "context": 0.8869705698743928}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.17834429442882538, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 18.96, "ce": 0.0137188092667202, "kl": 0.002809380345581758, "context": 0.015123499149922282}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.3872220516204834, "loss_scale": 4.0, "retries": 0, "seconds": 56.0, "max_memory_gb": 18.96, "ce": 0.046501971752149984, "kl": 0.38637590780854225, "context": 0.23968991823494434}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 2.0720765590667725, "loss_scale": 4.0, "retries": 0, "seconds": 56.5, "max_memory_gb": 18.96, "ce": 0.32240709103643894, "kl": 0.33162558800540864, "context": 0.4882198730483651}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.01806967705488205, "loss_scale": 4.0, "retries": 0, "seconds": 57.0, "max_memory_gb": 18.96, "ce": 0.0014339681265482795, "kl": 0.0018688087229747907, "context": 0.0023683724357397296}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.025271175429224968, "loss_scale": 4.0, "retries": 0, "seconds": 57.6, "max_memory_gb": 18.96, "ce": 0.0034598659549374133, "kl": 0.0005149464554961014, "context": 0.003717339161084965}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.6220430135726929, "loss_scale": 4.0, "retries": 0, "seconds": 58.1, "max_memory_gb": 18.96, "ce": 0.008134944480843842, "kl": 0.5000790535705164, "context": 0.2581744643393904}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.2349562644958496, "loss_scale": 4.0, "retries": 0, "seconds": 58.6, "max_memory_gb": 18.96, "ce": 0.016232651192694902, "kl": 1.240225177258253, "context": 0.6363452402874827}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.19357512891292572, "loss_scale": 4.0, "retries": 0, "seconds": 59.1, "max_memory_gb": 18.96, "ce": 0.01220725635994313, "kl": 0.054612674117379356, "context": 0.03951359330312698}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 1.244156002998352, "loss_scale": 4.0, "retries": 0, "seconds": 59.6, "max_memory_gb": 18.96, "ce": 0.22801323342901014, "kl": 0.6681981122463299, "context": 0.562112319355947}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 2.222731828689575, "loss_scale": 4.0, "retries": 0, "seconds": 60.1, "max_memory_gb": 18.97, "ce": 0.17652968525726465, "kl": 0.29329155880986946, "context": 0.32317546463309554}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.3917120397090912, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 18.97, "ce": 0.18515902590843325, "kl": 0.01907473486392064, "context": 0.19469639776070835}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.23290924727916718, "loss_scale": 4.0, "retries": 0, "seconds": 61.1, "max_memory_gb": 18.97, "ce": 0.014112862441834295, "kl": 0.017130525485612225, "context": 0.022678125471429667}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.1884337663650513, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 18.97, "ce": 0.8004684394982178, "kl": 0.20221318372932728, "context": 0.9015750506951008}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 1.1067469120025635, "loss_scale": 4.0, "retries": 0, "seconds": 62.1, "max_memory_gb": 18.97, "ce": 0.055410970862794784, "kl": 0.049679361924063414, "context": 0.08025065185393032}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.11171060055494308, "loss_scale": 4.0, "retries": 0, "seconds": 62.6, "max_memory_gb": 18.97, "ce": 0.03568339610501425, "kl": 0.01274519205071556, "context": 0.04205599157648976}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.2115929275751114, "loss_scale": 4.0, "retries": 0, "seconds": 63.1, "max_memory_gb": 18.97, "ce": 0.029075018013827503, "kl": 0.025210241845343262, "context": 0.04168013937305659}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.367719054222107, "loss_scale": 4.0, "retries": 0, "seconds": 63.6, "max_memory_gb": 18.97, "ce": 0.7136793443060014, "kl": 0.7366017602616921, "context": 1.0819802535697818}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 1.0015259981155396, "loss_scale": 4.0, "retries": 0, "seconds": 64.1, "max_memory_gb": 18.97, "ce": 0.11642471965751611, "kl": 0.13906041852897033, "context": 0.18595492333406582}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.2558355331420898, "loss_scale": 4.0, "retries": 0, "seconds": 64.6, "max_memory_gb": 18.97, "ce": 0.4884591866284609, "kl": 0.15004092128947377, "context": 0.5634796358644962}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.049282222986221313, "loss_scale": 4.0, "retries": 0, "seconds": 65.1, "max_memory_gb": 18.97, "ce": 0.012633966615794634, "kl": 0.0015113023400772363, "context": 0.013389617790380726}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.8383291959762573, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 18.97, "ce": 0.47038151627930347, "kl": 1.6573186521563912, "context": 1.2990408520272467}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.18006093800067902, "loss_scale": 4.0, "retries": 0, "seconds": 66.2, "max_memory_gb": 18.97, "ce": 0.08355676825158298, "kl": 0.0368889095261693, "context": 0.10200122557580471}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.14378507435321808, "loss_scale": 4.0, "retries": 0, "seconds": 66.7, "max_memory_gb": 18.97, "ce": 0.04114120258600451, "kl": 0.31942464257190295, "context": 0.20085352833848447}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 2.077575445175171, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 18.97, "ce": 0.13041764221998164, "kl": 0.30488520092330873, "context": 0.28286024287808686}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.2610035538673401, "loss_scale": 4.0, "retries": 0, "seconds": 67.7, "max_memory_gb": 18.97, "ce": 0.18431325217170524, "kl": 0.03508622491699498, "context": 0.20185636113819783}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.48632341623306274, "loss_scale": 4.0, "retries": 0, "seconds": 68.2, "max_memory_gb": 18.97, "ce": 0.34639231563778594, "kl": 0.20117653533816338, "context": 0.44698057882487774}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.74669349193573, "loss_scale": 4.0, "retries": 0, "seconds": 68.7, "max_memory_gb": 18.97, "ce": 0.24432065858673013, "kl": 0.005657983643043707, "context": 0.24714965692328406}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.12908843159675598, "loss_scale": 4.0, "retries": 0, "seconds": 69.2, "max_memory_gb": 18.97, "ce": 0.03785982300178148, "kl": 0.09848007370601408, "context": 0.08709985774476081}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 4.125553131103516, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 18.97, "ce": 0.20316723825817462, "kl": 0.27714152506086975, "context": 0.34173798582924064}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.06894055753946304, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 18.97, "ce": 0.008769909807597287, "kl": 0.007171174991526641, "context": 0.012355497688986361}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 3.325916051864624, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 18.97, "ce": 0.11350376787595451, "kl": 0.2521299282088876, "context": 0.23956872429698706}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.7467379570007324, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 18.97, "ce": 0.0034422428943798877, "kl": 0.4290576507046353, "context": 0.2179710645796149}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.042760975658893585, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 18.97, "ce": 0.005663720869051758, "kl": 0.0054394764883909374, "context": 0.008383459062315524}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.014881729148328304, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 18.97, "ce": 0.003131317724182736, "kl": 0.005968341569769109, "context": 0.006115488562500104}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.9214199185371399, "loss_scale": 4.0, "retries": 0, "seconds": 72.7, "max_memory_gb": 18.97, "ce": 0.05730243215839437, "kl": 0.04072082648053765, "context": 0.07766284584067762}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.2487913370132446, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 18.97, "ce": 0.7206597110762232, "kl": 0.05686473673085857, "context": 0.7490920969030412}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.6882110238075256, "loss_scale": 4.0, "retries": 0, "seconds": 73.7, "max_memory_gb": 18.97, "ce": 0.09586014179512858, "kl": 0.23392554430756718, "context": 0.21282292273826897}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.05322660505771637, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 18.97, "ce": 0.0034693200141191483, "kl": 0.011830216032251428, "context": 0.00938442822007346}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.06664061546325684, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 18.97, "ce": 0.0067826874437741935, "kl": 0.011876699116328382, "context": 0.012721037084702402}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 0.6190969944000244, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 18.97, "ce": 0.4830112859453948, "kl": 0.6172582402141416, "context": 0.7916404074494494}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.05152052640914917, "loss_scale": 4.0, "retries": 0, "seconds": 75.8, "max_memory_gb": 18.97, "ce": 0.004413216029433897, "kl": 0.008239653379860101, "context": 0.00853304282463796}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.2330605387687683, "loss_scale": 4.0, "retries": 0, "seconds": 76.3, "max_memory_gb": 18.97, "ce": 0.030890114372596145, "kl": 0.01181194577657152, "context": 0.03679608725360595}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.17089878022670746, "loss_scale": 4.0, "retries": 0, "seconds": 76.8, "max_memory_gb": 18.97, "ce": 0.050796293071471155, "kl": 0.03019681735895574, "context": 0.0658947005867958}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.09490539878606796, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 18.97, "ce": 0.010387602242190042, "kl": 0.016476459031252944, "context": 0.018625831757162814}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.2693023383617401, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 18.97, "ce": 0.01531800904103875, "kl": 0.7253150598451867, "context": 0.3779755333753201}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.2959217131137848, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 18.97, "ce": 0.010297489092408796, "kl": 0.4974973120974937, "context": 0.2590461394365775}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 2.8318052291870117, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 18.97, "ce": 0.12412031344138086, "kl": 0.1376490427646786, "context": 0.19294483866542578}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.2852438688278198, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 18.97, "ce": 0.20988452534356838, "kl": 0.011278326216142887, "context": 0.21552368473021488}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 4.3682074546813965, "loss_scale": 4.0, "retries": 0, "seconds": 79.8, "max_memory_gb": 18.97, "ce": 0.011769015895424673, "kl": 0.4405403404152821, "context": 0.23203918860053818}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 1.9768022298812866, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 18.97, "ce": 0.1435489461873658, "kl": 0.10951289893273497, "context": 0.19830539188114926}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.335369348526001, "loss_scale": 4.0, "retries": 0, "seconds": 80.8, "max_memory_gb": 18.97, "ce": 0.16255593898040388, "kl": 0.0873932357353624, "context": 0.20625255120830843}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.3600740134716034, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 18.98, "ce": 0.0032806728304421995, "kl": 0.04949203215073794, "context": 0.028026688465615734}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.8391955494880676, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 18.98, "ce": 0.01995205710409209, "kl": 0.5038218500558287, "context": 0.2718629811424762}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 1.2608169317245483, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 18.98, "ce": 0.09343571961653652, "kl": 0.11857361241595754, "context": 0.15272252769500483}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.1041406393051147, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 18.98, "ce": 0.15397512686104164, "kl": 0.29293865807994734, "context": 0.30044445589737734}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 1.6480916738510132, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 18.98, "ce": 0.2856952587608248, "kl": 0.953682441264391, "context": 0.7625364817213267}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 5.773058891296387, "loss_scale": 4.0, "retries": 0, "seconds": 83.9, "max_memory_gb": 18.98, "ce": 0.1491083505388815, "kl": 0.27798689348855987, "context": 0.2881017976324074}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.05439922958612442, "loss_scale": 4.0, "retries": 0, "seconds": 84.4, "max_memory_gb": 18.98, "ce": 0.0022571998515559244, "kl": 0.01891255639930023, "context": 0.011713477804732975}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.7474507093429565, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 18.98, "ce": 0.2565801796299638, "kl": 0.13496075454168022, "context": 0.3240605624887394}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.1972493678331375, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 18.98, "ce": 0.0354742988711223, "kl": 0.005083370189822745, "context": 0.038015984115190804}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.11330270022153854, "loss_scale": 4.0, "retries": 0, "seconds": 85.9, "max_memory_gb": 18.98, "ce": 0.03145776747260243, "kl": 0.004042641943669878, "context": 0.033479088931926526}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 0.9190447926521301, "loss_scale": 4.0, "retries": 0, "seconds": 86.4, "max_memory_gb": 18.98, "ce": 0.1113847597152926, "kl": 1.6336406469345093, "context": 0.9282051101326942}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.11079658567905426, "loss_scale": 4.0, "retries": 0, "seconds": 86.9, "max_memory_gb": 18.98, "ce": 0.004615597268639249, "kl": 0.0032201120775425807, "context": 0.006225653298315592}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.510127067565918, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 18.98, "ce": 0.03507966661209139, "kl": 0.31855554642970674, "context": 0.19435744192742277}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.08791660517454147, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 18.98, "ce": 0.01645446929614991, "kl": 0.01520232018083334, "context": 0.024055629735812545}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.15108653903007507, "loss_scale": 4.0, "retries": 0, "seconds": 88.5, "max_memory_gb": 18.98, "ce": 0.05284354658215307, "kl": 0.03461645451898221, "context": 0.07015177491120994}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.23246392607688904, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 18.98, "ce": 0.05626359941197734, "kl": 0.030253342460639487, "context": 0.07139027250195795}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 2.5122969150543213, "loss_scale": 4.0, "retries": 0, "seconds": 89.6, "max_memory_gb": 18.98, "ce": 0.6937809803057462, "kl": 0.10485951835289598, "context": 0.7462107352912426}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.08984461426734924, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 18.98, "ce": 0.009840656304731965, "kl": 0.002888279983380926, "context": 0.011284796230029315}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.37388190627098083, "loss_scale": 4.0, "retries": 0, "seconds": 90.6, "max_memory_gb": 18.98, "ce": 0.008172553854819853, "kl": 0.9434828018347616, "context": 0.479913941497216}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.9274818897247314, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 18.98, "ce": 0.03602658032468753, "kl": 0.2058732916120789, "context": 0.13896322798973415}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.6005780696868896, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 18.98, "ce": 0.3447389024331642, "kl": 0.039891827409519465, "context": 0.3646848067837709}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.08659005910158157, "loss_scale": 4.0, "retries": 0, "seconds": 92.1, "max_memory_gb": 18.98, "ce": 0.030035919276997447, "kl": 0.021924684406258166, "context": 0.04099826142191887}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.15513546764850616, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 18.98, "ce": 0.009498732746578753, "kl": 0.11192346836469369, "context": 0.0654604674782604}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 3.5124008655548096, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 18.98, "ce": 0.046218769435654394, "kl": 0.4246017233999737, "context": 0.25851962869637646}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.023386266082525253, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 18.98, "ce": 0.0022230041836337477, "kl": 0.0006404414093594823, "context": 0.0025432248003198765}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.07818631827831268, "loss_scale": 4.0, "retries": 0, "seconds": 94.2, "max_memory_gb": 18.98, "ce": 0.004802599323738832, "kl": 0.0029455789299390744, "context": 0.006275388892390765}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.29358386993408203, "loss_scale": 4.0, "retries": 0, "seconds": 94.7, "max_memory_gb": 18.98, "ce": 0.028919396398123354, "kl": 0.128206392750144, "context": 0.09302259725518525}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 4.325638771057129, "loss_scale": 4.0, "retries": 0, "seconds": 95.2, "max_memory_gb": 18.98, "ce": 0.2607893957210763, "kl": 0.5764876204520988, "context": 0.5490332220997516}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.8024309277534485, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 18.98, "ce": 0.1421125393389957, "kl": 0.566490052267909, "context": 0.4253575773909688}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.2086436748504639, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 18.98, "ce": 0.3504050384990478, "kl": 0.015541225068773201, "context": 0.35817564893841336}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.52878999710083, "loss_scale": 4.0, "retries": 0, "seconds": 96.7, "max_memory_gb": 18.98, "ce": 0.3562849891022779, "kl": 0.4399126132193487, "context": 0.5762412956682965}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.06981436163187027, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 18.98, "ce": 0.004208537167869508, "kl": 0.036032077245181426, "context": 0.02222457667812705}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 7.076426982879639, "loss_scale": 4.0, "retries": 0, "seconds": 97.7, "max_memory_gb": 18.98, "ce": 0.19781881657218037, "kl": 0.30148935728857396, "context": 0.34856347851200553}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.276957631111145, "loss_scale": 4.0, "retries": 0, "seconds": 98.2, "max_memory_gb": 18.98, "ce": 0.5102504066308029, "kl": 0.11476722898078151, "context": 0.5676340211066417}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.10395245999097824, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 18.98, "ce": 0.010168041440920206, "kl": 0.018802635536758316, "context": 0.019569358394619485}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.2493802309036255, "loss_scale": 4.0, "retries": 0, "seconds": 99.2, "max_memory_gb": 18.98, "ce": 0.021816864471475128, "kl": 0.024713891791179776, "context": 0.03417381035978906}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.02485785260796547, "loss_scale": 4.0, "retries": 0, "seconds": 99.7, "max_memory_gb": 18.98, "ce": 0.00778044734033756, "kl": 0.006314644044323359, "context": 0.01093776899506338}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.47263216972351074, "loss_scale": 4.0, "retries": 0, "seconds": 100.2, "max_memory_gb": 18.98, "ce": 0.0022769295824218716, "kl": 0.6611237913830124, "context": 0.3328388311820163}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 0.9383494257926941, "loss_scale": 4.0, "retries": 0, "seconds": 100.8, "max_memory_gb": 18.98, "ce": 0.0022877223964314908, "kl": 0.9108470051141921, "context": 0.4577112279366702}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.48888057470321655, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 18.98, "ce": 0.05491276651082444, "kl": 0.01774496415600879, "context": 0.0637852475992986}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.5311357975006104, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 18.98, "ce": 0.1997921514630434, "kl": 0.22646693600700019, "context": 0.3130256208605715}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 1.9523428678512573, "loss_scale": 4.0, "retries": 0, "seconds": 102.3, "max_memory_gb": 18.98, "ce": 0.432535801679478, "kl": 0.05552561691365554, "context": 0.4602986061363481}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 1.075562596321106, "loss_scale": 4.0, "retries": 0, "seconds": 102.8, "max_memory_gb": 18.98, "ce": 0.0707029315381078, "kl": 0.16471292808273574, "context": 0.15305939561221749}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.029294639825820923, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 18.98, "ce": 0.0019277989536021778, "kl": 0.002745613201625474, "context": 0.003300605658296263}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.4088200628757477, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 18.98, "ce": 0.012283415868296288, "kl": 0.026889100729022175, "context": 0.025727966480189934}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.11881604045629501, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 18.98, "ce": 0.032972912420518696, "kl": 0.014392861485248432, "context": 0.04016934451647103}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.2215890884399414, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 18.98, "ce": 0.0015036104305181652, "kl": 0.6460757015665877, "context": 0.3245414472767152}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.056184571236371994, "loss_scale": 4.0, "retries": 0, "seconds": 105.3, "max_memory_gb": 18.98, "ce": 0.00740640157391681, "kl": 0.003182952027373176, "context": 0.00899787732578261}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.12145908921957016, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 18.98, "ce": 0.051357803680730285, "kl": 0.07206586823303951, "context": 0.08739074153709225}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.07247952371835709, "loss_scale": 4.0, "retries": 0, "seconds": 106.3, "max_memory_gb": 18.98, "ce": 0.005362880976463202, "kl": 0.0021831555059179664, "context": 0.006454458802181762}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.6271920204162598, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 18.98, "ce": 0.1379451782850083, "kl": 0.19017858827282907, "context": 0.2330344673100626}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 1.682898759841919, "loss_scale": 4.0, "retries": 0, "seconds": 107.3, "max_memory_gb": 18.98, "ce": 0.5158633854266554, "kl": 0.18077340282798104, "context": 0.6062500830867066}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.35970282554626465, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 18.98, "ce": 0.06818124884739518, "kl": 0.2185630047461018, "context": 0.17746275896206498}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 0.37219077348709106, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 18.98, "ce": 0.015250517521053553, "kl": 1.1745796625036746, "context": 0.60254037741106}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.45927923917770386, "loss_scale": 4.0, "retries": 0, "seconds": 108.9, "max_memory_gb": 18.98, "ce": 0.23562622417375678, "kl": 0.06291248667457694, "context": 0.2670824637862097}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.5675139427185059, "loss_scale": 4.0, "retries": 0, "seconds": 109.4, "max_memory_gb": 18.98, "ce": 0.4940532047985471, "kl": 0.20468345152039547, "context": 0.5963949164724909}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.1753995418548584, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 18.98, "ce": 0.6518336723920584, "kl": 0.02018178930728709, "context": 0.6619245837946437}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.9219538569450378, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 18.98, "ce": 0.022916375561180757, "kl": 0.17131841169612017, "context": 0.10857557861527312}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.014599842950701714, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 18.98, "ce": 0.0007626533042639494, "kl": 0.003919585065887077, "context": 0.0027224457371630706}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.8641185164451599, "loss_scale": 4.0, "retries": 0, "seconds": 111.4, "max_memory_gb": 18.98, "ce": 0.127925387583673, "kl": 0.3007251173257828, "context": 0.27828794717788696}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 1.3180837631225586, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 18.98, "ce": 0.005148557836946566, "kl": 0.2747845639823936, "context": 0.14254083238483872}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.11102703958749771, "loss_scale": 4.0, "retries": 0, "seconds": 112.4, "max_memory_gb": 18.98, "ce": 0.053502118522374076, "kl": 0.007500465055272798, "context": 0.05725234941201052}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.29060178995132446, "loss_scale": 4.0, "retries": 0, "seconds": 112.9, "max_memory_gb": 18.98, "ce": 0.003925756609532982, "kl": 0.6514936300227419, "context": 0.32967255695257336}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.15971165895462036, "loss_scale": 4.0, "retries": 0, "seconds": 113.5, "max_memory_gb": 18.98, "ce": 0.03683381643713801, "kl": 0.07616139100715458, "context": 0.07491451168243657}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 0.1708502620458603, "loss_scale": 4.0, "retries": 0, "seconds": 114.0, "max_memory_gb": 18.98, "ce": 0.006482746022811625, "kl": 0.00042421620310051367, "context": 0.006694854204397416}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.3116990923881531, "loss_scale": 4.0, "retries": 0, "seconds": 114.5, "max_memory_gb": 18.98, "ce": 0.008104746186290868, "kl": 0.2523290235449167, "context": 0.13426925854582805}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.1807195395231247, "loss_scale": 4.0, "retries": 0, "seconds": 115.0, "max_memory_gb": 18.98, "ce": 0.0029644981004821602, "kl": 0.29927087222495175, "context": 0.15259992852224968}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 2.1209359169006348, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 18.98, "ce": 0.15223212714772671, "kl": 0.0973063075216487, "context": 0.20088528087944724}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.245442733168602, "loss_scale": 4.0, "retries": 0, "seconds": 116.0, "max_memory_gb": 18.98, "ce": 0.02513497412110155, "kl": 0.003610693325754255, "context": 0.026940320180074195}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.4937261641025543, "loss_scale": 4.0, "retries": 0, "seconds": 116.5, "max_memory_gb": 18.98, "ce": 0.0016577949172642548, "kl": 0.5495242518372834, "context": 0.2764199276425643}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.3317440450191498, "loss_scale": 4.0, "retries": 0, "seconds": 117.0, "max_memory_gb": 18.98, "ce": 0.06311184691548988, "kl": 0.016503276320236182, "context": 0.07136348300264217}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 2.5097782611846924, "loss_scale": 4.0, "retries": 0, "seconds": 117.5, "max_memory_gb": 18.98, "ce": 0.536767880046682, "kl": 0.6664086254313588, "context": 0.86997215935844}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.15092229843139648, "loss_scale": 4.0, "retries": 0, "seconds": 118.0, "max_memory_gb": 18.98, "ce": 0.02031589940088452, "kl": 0.14670809423842002, "context": 0.09366994768424775}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.11961091309785843, "loss_scale": 4.0, "retries": 0, "seconds": 118.5, "max_memory_gb": 18.98, "ce": 0.021503361640498042, "kl": 0.02094500692328438, "context": 0.03197586559690535}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.45496058464050293, "loss_scale": 4.0, "retries": 0, "seconds": 119.0, "max_memory_gb": 18.98, "ce": 0.008016634732484818, "kl": 0.13776437286287546, "context": 0.07689882308477536}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.6353839635848999, "loss_scale": 4.0, "retries": 0, "seconds": 119.5, "max_memory_gb": 18.98, "ce": 0.008714014787983615, "kl": 0.16091385892650578, "context": 0.08917094324715436}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.327990174293518, "loss_scale": 4.0, "retries": 0, "seconds": 120.0, "max_memory_gb": 18.98, "ce": 0.44797027437380166, "kl": 0.059361370111219, "context": 0.4776509631774388}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.032696157693862915, "loss_scale": 4.0, "retries": 0, "seconds": 120.5, "max_memory_gb": 18.98, "ce": 0.005291718014632352, "kl": 0.012696591322310269, "context": 0.011640013733995147}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.6939123868942261, "loss_scale": 4.0, "retries": 0, "seconds": 121.0, "max_memory_gb": 18.98, "ce": 0.07645469328781473, "kl": 0.05248076940102919, "context": 0.10269507807606715}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.3137120306491852, "loss_scale": 4.0, "retries": 0, "seconds": 121.5, "max_memory_gb": 18.98, "ce": 0.0264911656267941, "kl": 0.018280836005942547, "context": 0.03563158237375319}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 1.514756202697754, "loss_scale": 4.0, "retries": 0, "seconds": 122.0, "max_memory_gb": 18.98, "ce": 0.10683096142020077, "kl": 0.04932699473283719, "context": 0.131494456960354}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.20388109982013702, "loss_scale": 4.0, "retries": 0, "seconds": 122.5, "max_memory_gb": 18.98, "ce": 0.11825199065060588, "kl": 0.04540035521131358, "context": 0.14095216906571295}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 0.6219296455383301, "loss_scale": 4.0, "retries": 0, "seconds": 123.0, "max_memory_gb": 18.98, "ce": 0.08633769121661317, "kl": 0.029027049142314354, "context": 0.10085121357587923}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.10016396641731262, "loss_scale": 4.0, "retries": 0, "seconds": 123.6, "max_memory_gb": 18.98, "ce": 0.002635877138345677, "kl": 0.07778649407327975, "context": 0.04152912359131733}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.14145655930042267, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 18.98, "ce": 0.011882355953275692, "kl": 0.012073329289705725, "context": 0.017919021178386174}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 1.0349540710449219, "loss_scale": 4.0, "retries": 0, "seconds": 124.8, "max_memory_gb": 18.98, "ce": 0.03583223115128931, "kl": 0.20815578964538872, "context": 0.13991012645419687}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 1.4871704578399658, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 18.98, "ce": 0.3452576547861099, "kl": 0.10355844802688807, "context": 0.3970368728041649}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.024403948336839676, "loss_scale": 4.0, "retries": 0, "seconds": 125.8, "max_memory_gb": 18.98, "ce": 0.0008483973306283588, "kl": 0.003287613219981722, "context": 0.0024922038828663062}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.9192941188812256, "loss_scale": 4.0, "retries": 0, "seconds": 126.3, "max_memory_gb": 18.98, "ce": 0.02335078642136068, "kl": 0.6352659323683838, "context": 0.3409837462604628}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.3144189119338989, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 18.98, "ce": 0.028346118662739173, "kl": 0.025070299510844052, "context": 0.040881268214434385}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.16786018013954163, "loss_scale": 4.0, "retries": 0, "seconds": 127.3, "max_memory_gb": 18.98, "ce": 0.027028330136090517, "kl": 0.0296074912766926, "context": 0.04183207580354065}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.03469132259488106, "loss_scale": 4.0, "retries": 0, "seconds": 127.8, "max_memory_gb": 18.98, "ce": 0.004101354010344949, "kl": 0.00305124836449977, "context": 0.0056269781780429184}
|
runs/single_internvl/context/L02_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L02_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.2.self_attn.o_proj",
|
| 35 |
+
"model.language_model.layers.2.mlp.up_proj",
|
| 36 |
+
"model.language_model.layers.2.self_attn.q_proj",
|
| 37 |
+
"model.language_model.layers.2.mlp.gate_proj",
|
| 38 |
+
"model.language_model.layers.2.self_attn.k_proj",
|
| 39 |
+
"model.language_model.layers.2.mlp.down_proj",
|
| 40 |
+
"model.language_model.layers.2.self_attn.v_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L02_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 2,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L02_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.5113595724105835, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 18.86, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.8831896185874939, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 18.87, "ce": 0.6385493846828467, "kl": 0.4661056873178495, "context": 0.8716021942018415}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 2.078836441040039, "loss_scale": 4.0, "retries": 0, "seconds": 2.5, "max_memory_gb": 18.87, "ce": 1.098431184887886, "kl": 0.06860820762813091, "context": 1.1327353119850159}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.0982985496520996, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 18.87, "ce": 0.45131311725708656, "kl": 0.42913408672393416, "context": 0.6658801622470492}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.02149192802608013, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 18.87, "ce": 0.009262646888601012, "kl": 0.008212978355004452, "context": 0.013369136417168193}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.48654404282569885, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 18.87, "ce": 0.2871151866565924, "kl": 0.07454044656697079, "context": 0.32438541192095727}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 1.766396403312683, "loss_scale": 4.0, "retries": 0, "seconds": 4.6, "max_memory_gb": 18.87, "ce": 0.8232347047942312, "kl": 0.12348244012480336, "context": 0.8849759378954332}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 1.636108160018921, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 18.88, "ce": 0.7816616557538509, "kl": 0.28901919117197394, "context": 0.9261712692677975}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 1.1591521501541138, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 18.88, "ce": 0.5146011114120483, "kl": 0.13601952316821553, "context": 0.582610871642828}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.19274839758872986, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 18.88, "ce": 0.05225200984568801, "kl": 0.04544100305065513, "context": 0.07497251301538199}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 1.297598958015442, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 18.88, "ce": 0.6870616157175391, "kl": 0.004987321702628833, "context": 0.6895552827118081}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 0.9506701231002808, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 18.88, "ce": 0.4598000148243955, "kl": 0.002664682920112682, "context": 0.4611323652370629}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 0.72252357006073, "loss_scale": 4.0, "retries": 0, "seconds": 7.7, "max_memory_gb": 18.88, "ce": 0.4186271364305867, "kl": 0.019770268265347113, "context": 0.42851228077779524}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 1.6273829936981201, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 18.88, "ce": 0.7322311699390411, "kl": 0.9434448080137372, "context": 1.203953579068184}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 1.5432528257369995, "loss_scale": 4.0, "retries": 0, "seconds": 8.7, "max_memory_gb": 18.88, "ce": 0.4826182944880202, "kl": 0.46426316538975243, "context": 0.7147498846334202}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.47789692878723145, "loss_scale": 4.0, "retries": 0, "seconds": 9.2, "max_memory_gb": 18.88, "ce": 0.1989589036165853, "kl": 0.0009021912501339102, "context": 0.19941000494145555}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.595705509185791, "loss_scale": 4.0, "retries": 0, "seconds": 9.7, "max_memory_gb": 18.88, "ce": 0.30816595719079487, "kl": 0.8294909682299476, "context": 0.7229114273213781}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.5658928155899048, "loss_scale": 4.0, "retries": 0, "seconds": 10.2, "max_memory_gb": 18.88, "ce": 0.49361010815482587, "kl": 0.2102842673921259, "context": 0.5987522543000523}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.74367094039917, "loss_scale": 4.0, "retries": 0, "seconds": 10.7, "max_memory_gb": 18.88, "ce": 0.6955608278512955, "kl": 0.17086793365888298, "context": 0.7809948027133942}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.027533747255802155, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 18.88, "ce": 0.0023556412934340187, "kl": 0.020668803106673295, "context": 0.012690042600297602}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.542750597000122, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 18.88, "ce": 0.31264629607085226, "kl": 0.04866100795334205, "context": 0.33697679630859056}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 1.4042367935180664, "loss_scale": 4.0, "retries": 0, "seconds": 12.3, "max_memory_gb": 18.88, "ce": 0.3388922617014032, "kl": 0.032960911088594, "context": 0.35537270466738846}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 1.8449172973632812, "loss_scale": 4.0, "retries": 0, "seconds": 12.8, "max_memory_gb": 18.88, "ce": 0.5174039897365219, "kl": 0.98827925608839, "context": 1.011543623280886}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 1.3545864820480347, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 18.88, "ce": 0.2655661329627037, "kl": 0.30270162876695395, "context": 0.4169169422239065}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.7266093492507935, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 18.88, "ce": 0.6325299715535948, "kl": 0.20507778623141348, "context": 0.7350688688457012}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 1.6026840209960938, "loss_scale": 4.0, "retries": 0, "seconds": 14.4, "max_memory_gb": 18.88, "ce": 0.2632291540503502, "kl": 0.7033011235762388, "context": 0.6148797050118446}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 1.7736295461654663, "loss_scale": 4.0, "retries": 0, "seconds": 14.9, "max_memory_gb": 18.88, "ce": 0.3415953044386697, "kl": 0.41885100388753926, "context": 0.5510207996849203}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.8158969283103943, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 18.88, "ce": 0.5915569439530373, "kl": 0.9430697424104437, "context": 1.063091829419136}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 1.4336439371109009, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 18.88, "ce": 0.22879723040387034, "kl": 0.3679667663818691, "context": 0.41278060898184776}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 1.4258280992507935, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 18.88, "ce": 0.29597065763300634, "kl": 0.012094407606127788, "context": 0.3020178636688797}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 2.368364095687866, "loss_scale": 4.0, "retries": 0, "seconds": 16.9, "max_memory_gb": 18.88, "ce": 0.5646668076515198, "kl": 0.15558304451406002, "context": 0.6424583308398724}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 2.2849440574645996, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 18.88, "ce": 0.33852413296699524, "kl": 0.0174285288376268, "context": 0.347238402813673}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.8515433073043823, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 18.88, "ce": 0.19782700389623642, "kl": 0.06652530818246305, "context": 0.23108965856954455}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.9592751860618591, "loss_scale": 4.0, "retries": 0, "seconds": 18.5, "max_memory_gb": 18.88, "ce": 0.12500090804132924, "kl": 0.7711627234875778, "context": 0.5105822618534148}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.9967995285987854, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 18.88, "ce": 0.1461316129862098, "kl": 0.0040863378962967545, "context": 0.1481747828074731}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.7660173177719116, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 18.88, "ce": 0.19938988986541517, "kl": 0.057358146746992134, "context": 0.22806896406109445}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.8075708150863647, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 18.88, "ce": 0.30030008978792466, "kl": 0.7799924541031942, "context": 0.6902963075553998}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.810559093952179, "loss_scale": 4.0, "retries": 0, "seconds": 20.5, "max_memory_gb": 18.88, "ce": 0.10551806539297104, "kl": 0.5072835236787796, "context": 0.359159829095006}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.356835275888443, "loss_scale": 4.0, "retries": 0, "seconds": 21.0, "max_memory_gb": 18.88, "ce": 0.16724744548264425, "kl": 0.012548611019155942, "context": 0.17352174752159044}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.6043879985809326, "loss_scale": 4.0, "retries": 0, "seconds": 21.5, "max_memory_gb": 18.88, "ce": 0.07092333998298272, "kl": 0.012480068719014525, "context": 0.0771633745462168}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.1334683895111084, "loss_scale": 4.0, "retries": 0, "seconds": 22.0, "max_memory_gb": 18.88, "ce": 0.40875118284020573, "kl": 0.12902081792708486, "context": 0.4732615896500647}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.27106907963752747, "loss_scale": 4.0, "retries": 0, "seconds": 22.5, "max_memory_gb": 18.88, "ce": 0.021656896347849397, "kl": 0.0006079156755731674, "context": 0.021960854315693723}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.5822651386260986, "loss_scale": 4.0, "retries": 0, "seconds": 23.0, "max_memory_gb": 18.88, "ce": 0.03808628905972, "kl": 0.09920682742449571, "context": 0.08768970459641423}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 1.9198886156082153, "loss_scale": 4.0, "retries": 0, "seconds": 23.6, "max_memory_gb": 18.88, "ce": 0.6349793334593414, "kl": 1.403952966342331, "context": 1.3369558203557972}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 0.8576054573059082, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 18.88, "ce": 0.9093665395630524, "kl": 0.24546256712346803, "context": 1.032097838964546}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.4399333596229553, "loss_scale": 4.0, "retries": 0, "seconds": 24.6, "max_memory_gb": 18.88, "ce": 0.11865309439599514, "kl": 0.289541226811707, "context": 0.26342371106147766}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.4626289904117584, "loss_scale": 4.0, "retries": 0, "seconds": 25.1, "max_memory_gb": 18.88, "ce": 0.02978264380362816, "kl": 0.047433964093215764, "context": 0.053499626548727974}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.7467386722564697, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 18.88, "ce": 0.09227684864890762, "kl": 0.009112413903494598, "context": 0.09683305257931352}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.4085307717323303, "loss_scale": 4.0, "retries": 0, "seconds": 26.1, "max_memory_gb": 18.88, "ce": 0.041734760539839044, "kl": 0.3851880495541309, "context": 0.23432878904350218}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.6423295736312866, "loss_scale": 4.0, "retries": 0, "seconds": 26.6, "max_memory_gb": 18.88, "ce": 0.15438325429568067, "kl": 0.008005480543943122, "context": 0.15838599305425305}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.9470500946044922, "loss_scale": 4.0, "retries": 0, "seconds": 27.1, "max_memory_gb": 18.88, "ce": 0.6095561306447053, "kl": 0.019245454194560807, "context": 0.6191788721771445}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 1.2939679622650146, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 18.88, "ce": 0.15465629933169112, "kl": 0.27574429637752473, "context": 0.29252844490110874}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.47400233149528503, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 18.88, "ce": 0.1832674810430035, "kl": 0.4304545698687434, "context": 0.39849476516246796}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5721714496612549, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 18.88, "ce": 0.05620607116725296, "kl": 0.3530711789499037, "context": 0.23274165973998606}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.05784311145544052, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 18.88, "ce": 0.015214839891996235, "kl": 0.015100616104973597, "context": 0.022765147834434174}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.28254464268684387, "loss_scale": 4.0, "retries": 0, "seconds": 29.6, "max_memory_gb": 18.88, "ce": 0.03510688792448491, "kl": 0.30697250948287547, "context": 0.18859314545989037}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.07182689756155014, "loss_scale": 4.0, "retries": 0, "seconds": 30.1, "max_memory_gb": 18.88, "ce": 0.011962271517404588, "kl": 0.0193986533640782, "context": 0.021661598206264898}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.3119354248046875, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 18.88, "ce": 0.040557609987445176, "kl": 0.0023375425448080023, "context": 0.04172637957526604}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.17477233707904816, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 18.88, "ce": 0.024076633897493593, "kl": 0.06579507562855724, "context": 0.05697416889597662}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.11249559372663498, "loss_scale": 4.0, "retries": 0, "seconds": 31.7, "max_memory_gb": 18.88, "ce": 0.07213514648901764, "kl": 0.06269878198509105, "context": 0.10348453384358436}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.5968500375747681, "loss_scale": 4.0, "retries": 0, "seconds": 32.2, "max_memory_gb": 18.88, "ce": 0.029185163090005517, "kl": 1.0186271648854017, "context": 0.5384987406432629}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.33136048913002014, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 18.88, "ce": 0.039533545554149896, "kl": 0.010085488698678091, "context": 0.04457628936506808}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.14737972617149353, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 18.88, "ce": 0.035687374882400036, "kl": 0.12787054923683172, "context": 0.09962264867499471}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.12392888218164444, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 18.88, "ce": 0.020047694677487016, "kl": 0.01259436592226848, "context": 0.026344877667725086}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.5253763198852539, "loss_scale": 4.0, "retries": 0, "seconds": 34.2, "max_memory_gb": 18.88, "ce": 0.19899121019989252, "kl": 0.17438765824772418, "context": 0.2861850350163877}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.44296810030937195, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 18.88, "ce": 0.15271974743484407, "kl": 0.06921341986736707, "context": 0.187326457397603}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.4484291672706604, "loss_scale": 4.0, "retries": 0, "seconds": 35.3, "max_memory_gb": 18.88, "ce": 0.027913753787288442, "kl": 0.04339889367111027, "context": 0.04961320152506232}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.19014142453670502, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 18.88, "ce": 0.012002645266875334, "kl": 0.06207210784123163, "context": 0.0430386992302374}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.5069074034690857, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 18.88, "ce": 0.2735484786098823, "kl": 0.044543919852003455, "context": 0.2958204501774162}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.10731574147939682, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 18.88, "ce": 0.006269634410273284, "kl": 0.00037294940011634026, "context": 0.006456108967540786}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.29614412784576416, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 18.88, "ce": 0.02368951761036442, "kl": 0.013877169491024688, "context": 0.030628101598267676}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.23012785613536835, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 18.88, "ce": 0.024218989070504904, "kl": 0.4495849722297862, "context": 0.2490114774554968}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.1915099620819092, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 18.88, "ce": 0.41515193930536043, "kl": 0.12604301941610174, "context": 0.4781734415482788}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.4188486337661743, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 18.88, "ce": 0.09296812855336611, "kl": 0.1373348102424643, "context": 0.1616355327432757}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 0.7426994442939758, "loss_scale": 4.0, "retries": 0, "seconds": 39.4, "max_memory_gb": 18.88, "ce": 0.4447729145613266, "kl": 1.0575919058173895, "context": 0.9735688953660429}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 0.4913827180862427, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 18.88, "ce": 0.01899548279470764, "kl": 1.8341394618619233, "context": 0.9360651975148357}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.04694395512342453, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 18.88, "ce": 0.007853632338083116, "kl": 0.012030034918552701, "context": 0.013868649446521886}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.08969906717538834, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 18.88, "ce": 0.002572662808233872, "kl": 0.0740471308818087, "context": 0.039596229209564626}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.5547751188278198, "loss_scale": 4.0, "retries": 0, "seconds": 41.4, "max_memory_gb": 18.88, "ce": 0.042697395448158204, "kl": 0.07000628994273939, "context": 0.07770054228967638}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.15122023224830627, "loss_scale": 4.0, "retries": 0, "seconds": 41.9, "max_memory_gb": 18.88, "ce": 0.005733423635319923, "kl": 0.08658641122467259, "context": 0.04902662778886224}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.4814637005329132, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 18.88, "ce": 0.10485120955854654, "kl": 0.09545669483486563, "context": 0.1525795531924814}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.27430132031440735, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 18.88, "ce": 0.02691927074920386, "kl": 0.005224459422606742, "context": 0.02953150009852834}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.3892926573753357, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 18.88, "ce": 0.17205384958651848, "kl": 0.8328409044479486, "context": 0.5884743120404892}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.21825510263442993, "loss_scale": 4.0, "retries": 0, "seconds": 43.9, "max_memory_gb": 18.88, "ce": 0.020734932255436433, "kl": 0.05396751138755462, "context": 0.04771868899888432}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 1.7392094135284424, "loss_scale": 4.0, "retries": 0, "seconds": 44.4, "max_memory_gb": 18.88, "ce": 0.047781276643490855, "kl": 0.061957826385492965, "context": 0.07876018797298912}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.484867811203003, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 18.88, "ce": 0.28635678469436243, "kl": 0.5904164458625019, "context": 0.5815649775322527}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.742273211479187, "loss_scale": 4.0, "retries": 0, "seconds": 45.4, "max_memory_gb": 18.88, "ce": 0.008402338848100044, "kl": 0.770933729596436, "context": 0.393869211897254}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.11680491268634796, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 18.88, "ce": 0.005680304045426965, "kl": 0.03254066309538928, "context": 0.021950635334860635}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.07942890375852585, "loss_scale": 4.0, "retries": 0, "seconds": 46.5, "max_memory_gb": 18.88, "ce": 0.010610836661726353, "kl": 0.008340989138559962, "context": 0.014781331576159573}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.17913760244846344, "loss_scale": 4.0, "retries": 0, "seconds": 47.0, "max_memory_gb": 18.88, "ce": 0.014799913231399842, "kl": 0.016041817216319032, "context": 0.02282082161400467}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 0.43138787150382996, "loss_scale": 4.0, "retries": 0, "seconds": 47.5, "max_memory_gb": 18.88, "ce": 0.15117375550653378, "kl": 0.7199798083784117, "context": 0.5111636657493364}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.5078376531600952, "loss_scale": 4.0, "retries": 0, "seconds": 48.0, "max_memory_gb": 18.88, "ce": 0.47582230185435037, "kl": 0.5463971026929357, "context": 0.7490208822891873}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.043196439743042, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 18.88, "ce": 0.059661963663529605, "kl": 0.5290146787010599, "context": 0.3241693072486669}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 2.3947033882141113, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 18.88, "ce": 0.343331994721666, "kl": 0.1709398437233176, "context": 0.42880193097516894}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.1919069290161133, "loss_scale": 4.0, "retries": 0, "seconds": 49.5, "max_memory_gb": 18.88, "ce": 0.6912346010212786, "kl": 0.026654368499293923, "context": 0.7045617704279721}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 0.6814757585525513, "loss_scale": 4.0, "retries": 0, "seconds": 50.1, "max_memory_gb": 18.88, "ce": 0.5363397034088848, "kl": 0.005619933130219579, "context": 0.5391496940283105}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 2.214322328567505, "loss_scale": 4.0, "retries": 0, "seconds": 50.6, "max_memory_gb": 18.88, "ce": 0.028740245397784747, "kl": 0.5820843671681359, "context": 0.31978243216872215}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.08354225009679794, "loss_scale": 4.0, "retries": 0, "seconds": 51.1, "max_memory_gb": 18.88, "ce": 0.0035890029794245493, "kl": 0.09831708979345422, "context": 0.05274754625679634}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.589219331741333, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 18.88, "ce": 0.03543608151994704, "kl": 0.01903488566949818, "context": 0.04495352364938299}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.5695346593856812, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 18.88, "ce": 0.02914485803921707, "kl": 0.8091576422593789, "context": 0.43372367951087654}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.2582665979862213, "loss_scale": 4.0, "retries": 0, "seconds": 52.6, "max_memory_gb": 18.88, "ce": 0.0042288092081435025, "kl": 0.09200364985736087, "context": 0.050230632623424754}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.12736155092716217, "loss_scale": 4.0, "retries": 0, "seconds": 53.1, "max_memory_gb": 18.88, "ce": 0.04879177262773737, "kl": 0.03769864980131388, "context": 0.0676410966552794}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.20410892367362976, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 18.88, "ce": 0.01742412327803322, "kl": 0.02163885850313818, "context": 0.028243553460924886}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.030742142349481583, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 18.88, "ce": 0.002433383921015775, "kl": 0.0050562516059926566, "context": 0.00496150970866438}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.2778078317642212, "loss_scale": 4.0, "retries": 0, "seconds": 54.6, "max_memory_gb": 18.88, "ce": 0.06916242942679673, "kl": 0.3825884227408096, "context": 0.26045664260163903}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.08988302946090698, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 18.88, "ce": 0.005508097314304905, "kl": 0.005412839488599275, "context": 0.008214517118176445}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 1.699154257774353, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 18.88, "ce": 0.8679920309878071, "kl": 0.08564963402000103, "context": 0.9108168372658838}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.10021886229515076, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 18.88, "ce": 0.009820480889175087, "kl": 0.0014367111162982837, "context": 0.010538836653722683}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.32888296246528625, "loss_scale": 4.0, "retries": 0, "seconds": 56.7, "max_memory_gb": 18.88, "ce": 0.04322106213658117, "kl": 0.37913041561841965, "context": 0.23278627637773752}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 4.989300727844238, "loss_scale": 4.0, "retries": 0, "seconds": 57.2, "max_memory_gb": 18.88, "ce": 0.29900934058241546, "kl": 0.2963506575906649, "context": 0.44718467025086284}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.01362613309174776, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 18.88, "ce": 0.001107942804310369, "kl": 0.0013606439406430582, "context": 0.001788264784408966}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.022301629185676575, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 18.88, "ce": 0.0026211283693555743, "kl": 0.0004548519204945478, "context": 0.0028485543953138404}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.8071959614753723, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 18.88, "ce": 0.007810546609107405, "kl": 0.5211830346379429, "context": 0.26840206095948815}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.183284878730774, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 18.88, "ce": 0.015037076198495924, "kl": 1.2496597010176629, "context": 0.639866933459416}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.1717205047607422, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 18.88, "ce": 0.011683077862471691, "kl": 0.054788958295830525, "context": 0.0390775568648678}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 1.1080090999603271, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 18.88, "ce": 0.21326091537957836, "kl": 0.6330828563598487, "context": 0.529802343560732}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 2.0682930946350098, "loss_scale": 4.0, "retries": 0, "seconds": 60.7, "max_memory_gb": 18.89, "ce": 0.18685158744483488, "kl": 0.3030004665142769, "context": 0.3383518207047018}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.3343595266342163, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 18.89, "ce": 0.17128222900282708, "kl": 0.020991614033619044, "context": 0.1817780334586132}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.16238318383693695, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 18.89, "ce": 0.015038180366900633, "kl": 0.015255734262950682, "context": 0.02266604714350251}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.0973008871078491, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 18.89, "ce": 0.8185217813006602, "kl": 0.20451467654493172, "context": 0.9207791136577725}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.2579769194126129, "loss_scale": 4.0, "retries": 0, "seconds": 62.8, "max_memory_gb": 18.89, "ce": 0.05070473282785315, "kl": 0.04336546685863141, "context": 0.07238746857092337}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.13044314086437225, "loss_scale": 4.0, "retries": 0, "seconds": 63.3, "max_memory_gb": 18.89, "ce": 0.03751759706847224, "kl": 0.013418399934380432, "context": 0.04422679770505056}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.2901575565338135, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 18.89, "ce": 0.02908824064070359, "kl": 0.022619874769588932, "context": 0.040398177632596344}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.3625659942626953, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 18.89, "ce": 0.6997214895673096, "kl": 0.7350226845228462, "context": 1.0672328538494185}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 1.279762625694275, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 18.89, "ce": 0.13248213832412148, "kl": 0.15677112736739218, "context": 0.21086771271075122}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.3904930353164673, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 18.89, "ce": 0.46761762350797653, "kl": 0.1625472162850201, "context": 0.548891231417656}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.06018422171473503, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 18.89, "ce": 0.012815514479370904, "kl": 0.001472092917538248, "context": 0.01355156117824663}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.556426763534546, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 18.89, "ce": 0.4877530694066081, "kl": 1.625829963650176, "context": 1.3006680069956928}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.17675979435443878, "loss_scale": 4.0, "retries": 0, "seconds": 66.9, "max_memory_gb": 18.89, "ce": 0.07872130069881678, "kl": 0.0365673495689407, "context": 0.09700497658923268}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.1440631002187729, "loss_scale": 4.0, "retries": 0, "seconds": 67.4, "max_memory_gb": 18.89, "ce": 0.039983502560062334, "kl": 0.3169354952069625, "context": 0.19845124921994284}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 1.761919617652893, "loss_scale": 4.0, "retries": 0, "seconds": 67.9, "max_memory_gb": 18.89, "ce": 0.11020715062477393, "kl": 0.26966858439845964, "context": 0.24504144294769503}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.25378644466400146, "loss_scale": 4.0, "retries": 0, "seconds": 68.5, "max_memory_gb": 18.89, "ce": 0.17155381355405552, "kl": 0.03408046506137907, "context": 0.18859405097464332}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.9350085258483887, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 18.89, "ce": 0.3463568896113429, "kl": 0.19533413369208574, "context": 0.44402395375072956}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.7425650358200073, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 18.89, "ce": 0.2550849492827183, "kl": 0.005537800359299183, "context": 0.257853858778617}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.11762218922376633, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 18.89, "ce": 0.034445243974914774, "kl": 0.10000827729527373, "context": 0.08444938459433615}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 4.4566874504089355, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 18.89, "ce": 0.29240538328303955, "kl": 0.3567829923995305, "context": 0.4707968789880397}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.05169380456209183, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 18.89, "ce": 0.0073819988683681, "kl": 0.00775223936943803, "context": 0.011258118320256472}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 2.0606894493103027, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 18.89, "ce": 0.10334185883402824, "kl": 0.23173288628458977, "context": 0.2192082954570651}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.7088186144828796, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 18.89, "ce": 0.0025176682211167645, "kl": 0.4668662438343745, "context": 0.23595079587539658}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.0392739437520504, "loss_scale": 4.0, "retries": 0, "seconds": 72.8, "max_memory_gb": 18.89, "ce": 0.0050070770284946775, "kl": 0.0048849135637283325, "context": 0.007449533601175062}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.028299445286393166, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 18.89, "ce": 0.0031285111108445562, "kl": 0.005875022747204639, "context": 0.0060660225281026214}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.9832183122634888, "loss_scale": 4.0, "retries": 0, "seconds": 73.8, "max_memory_gb": 18.89, "ce": 0.0557661214043037, "kl": 0.03547204367350787, "context": 0.07350214314647019}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.2569880485534668, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 18.89, "ce": 0.7052900893959304, "kl": 0.04767127914510638, "context": 0.7291257363212935}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.8552082180976868, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 18.89, "ce": 0.1334397722966969, "kl": 0.20779064309317619, "context": 0.2373350909911096}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.04527544230222702, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 18.89, "ce": 0.0023920859275676776, "kl": 0.010452494388744071, "context": 0.007618332867423305}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.06806769967079163, "loss_scale": 4.0, "retries": 0, "seconds": 75.8, "max_memory_gb": 18.89, "ce": 0.00571758730802685, "kl": 0.013143622018105816, "context": 0.012289398582652211}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 0.5965805053710938, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 18.89, "ce": 0.46920313287046156, "kl": 0.6272097787295934, "context": 0.7828080110011797}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.07706161588430405, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 18.89, "ce": 0.0059839392050662354, "kl": 0.012459494504582835, "context": 0.012213686055474682}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.2682195007801056, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 18.89, "ce": 0.03231970823253505, "kl": 0.013301085869898088, "context": 0.038970251160208136}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.13815590739250183, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 18.89, "ce": 0.051556661317590624, "kl": 0.023378448211587965, "context": 0.06324588553979993}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.07738377153873444, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 18.89, "ce": 0.009928006597874628, "kl": 0.015562560093371758, "context": 0.01770928711266606}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.26473984122276306, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 18.89, "ce": 0.012553656335512642, "kl": 0.7146060263295269, "context": 0.3698566610892158}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.29259851574897766, "loss_scale": 4.0, "retries": 0, "seconds": 79.5, "max_memory_gb": 18.89, "ce": 0.009743491694280237, "kl": 0.49015998849472453, "context": 0.25482348175046354}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 5.170650482177734, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 18.89, "ce": 0.22635695175267756, "kl": 0.10096408089157194, "context": 0.2768389950506389}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.32605645060539246, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 18.89, "ce": 0.20195462819719978, "kl": 0.010813109495302342, "context": 0.20736118391050695}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 4.149329662322998, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 18.89, "ce": 0.011585809259486268, "kl": 0.5309517379698718, "context": 0.2770616656735001}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 1.8956719636917114, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 18.89, "ce": 0.17430788991623558, "kl": 0.1103549388317333, "context": 0.22948536317562684}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.3056911528110504, "loss_scale": 4.0, "retries": 0, "seconds": 82.1, "max_memory_gb": 18.89, "ce": 0.16214629114074341, "kl": 0.0978023222196498, "context": 0.21104745879711118}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.4842010736465454, "loss_scale": 4.0, "retries": 0, "seconds": 82.6, "max_memory_gb": 18.9, "ce": 0.003474015389656415, "kl": 0.08684656226250809, "context": 0.04689729785604868}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.9468673467636108, "loss_scale": 4.0, "retries": 0, "seconds": 83.1, "max_memory_gb": 18.9, "ce": 0.023742396791931242, "kl": 0.401237522426527, "context": 0.22436116728931665}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 1.2423341274261475, "loss_scale": 4.0, "retries": 0, "seconds": 83.6, "max_memory_gb": 18.9, "ce": 0.0988069017566886, "kl": 0.11994909667328102, "context": 0.15878144637099467}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.0983953475952148, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 18.9, "ce": 0.1415381482383964, "kl": 0.27228345389448805, "context": 0.27767989011408645}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 1.5998307466506958, "loss_scale": 4.0, "retries": 0, "seconds": 84.7, "max_memory_gb": 18.9, "ce": 0.2556429873802699, "kl": 0.9354742137948051, "context": 0.7233801156980917}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 4.494747161865234, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 18.9, "ce": 0.1696689094096655, "kl": 0.3397353502805345, "context": 0.3395365693140775}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.05098913237452507, "loss_scale": 4.0, "retries": 0, "seconds": 85.7, "max_memory_gb": 18.9, "ce": 0.0022457891191152157, "kl": 0.021904388286202447, "context": 0.013197982971178135}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.7256860733032227, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 18.9, "ce": 0.24846602516481653, "kl": 0.14364255111286184, "context": 0.32028729512239806}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.21737374365329742, "loss_scale": 4.0, "retries": 0, "seconds": 86.8, "max_memory_gb": 18.9, "ce": 0.03116546181263402, "kl": 0.007315540962736122, "context": 0.03482323215575889}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.09733942151069641, "loss_scale": 4.0, "retries": 0, "seconds": 87.3, "max_memory_gb": 18.9, "ce": 0.032030245653004386, "kl": 0.0036252559220883995, "context": 0.033842874778201804}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.1581867933273315, "loss_scale": 4.0, "retries": 0, "seconds": 87.8, "max_memory_gb": 18.9, "ce": 0.10644135042093694, "kl": 1.629457637667656, "context": 0.9211701676249504}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.06776829808950424, "loss_scale": 4.0, "retries": 0, "seconds": 88.3, "max_memory_gb": 18.9, "ce": 0.0030756278301851125, "kl": 0.001478969948948361, "context": 0.003815112911979668}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.4686625599861145, "loss_scale": 4.0, "retries": 0, "seconds": 88.8, "max_memory_gb": 18.9, "ce": 0.03458639364998817, "kl": 0.30187852478411514, "context": 0.18552565843128832}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.10762770473957062, "loss_scale": 4.0, "retries": 0, "seconds": 89.3, "max_memory_gb": 18.9, "ce": 0.015806299401447177, "kl": 0.014514032460283488, "context": 0.02306331554427743}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.12612314522266388, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 18.9, "ce": 0.05629992007743567, "kl": 0.03400942051666789, "context": 0.07330462976824492}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.23436784744262695, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 18.9, "ce": 0.053333914202085, "kl": 0.032538042353408514, "context": 0.0696029325863492}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 3.0537896156311035, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 18.9, "ce": 0.688330648932606, "kl": 0.10564704734133556, "context": 0.7411541873589158}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.03643615543842316, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 18.9, "ce": 0.009367422142531723, "kl": 0.003070872124226298, "context": 0.010902858019107953}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.38081833720207214, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 18.9, "ce": 0.008503638699039584, "kl": 0.9456400641356595, "context": 0.48132365263882093}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.8678735494613647, "loss_scale": 4.0, "retries": 0, "seconds": 92.4, "max_memory_gb": 18.9, "ce": 0.10888481841539033, "kl": 0.13936850534082623, "context": 0.17856907201348804}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.5732279419898987, "loss_scale": 4.0, "retries": 0, "seconds": 92.9, "max_memory_gb": 18.9, "ce": 0.3259128004319791, "kl": 0.040466211015882436, "context": 0.34614590035562287}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.08089287579059601, "loss_scale": 4.0, "retries": 0, "seconds": 93.4, "max_memory_gb": 18.9, "ce": 0.02950503909960389, "kl": 0.022444151458330452, "context": 0.040727114712353796}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.1227366179227829, "loss_scale": 4.0, "retries": 0, "seconds": 93.9, "max_memory_gb": 18.9, "ce": 0.0079719195200596, "kl": 0.11091881444735918, "context": 0.06343132734764367}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 2.447655200958252, "loss_scale": 4.0, "retries": 0, "seconds": 94.4, "max_memory_gb": 18.9, "ce": 0.04582215138361789, "kl": 0.2883826795132336, "context": 0.19001348927849904}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.016019307076931, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 18.9, "ce": 0.001701760187643231, "kl": 0.00050809513936656, "context": 0.0019558077274268726}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.06497994065284729, "loss_scale": 4.0, "retries": 0, "seconds": 95.4, "max_memory_gb": 18.9, "ce": 0.003004587681061821, "kl": 0.0027335571285220794, "context": 0.004371366310806479}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.19403710961341858, "loss_scale": 4.0, "retries": 0, "seconds": 95.9, "max_memory_gb": 18.9, "ce": 0.020152282435446978, "kl": 0.12106990907341242, "context": 0.08068723953329027}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 6.1361565589904785, "loss_scale": 4.0, "retries": 0, "seconds": 96.5, "max_memory_gb": 18.9, "ce": 0.24550153430027422, "kl": 0.5575770901123178, "context": 0.5242900644843758}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.7373682260513306, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 18.9, "ce": 0.1279454934410751, "kl": 0.5467671644873917, "context": 0.4013290833681822}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.1781491041183472, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 18.9, "ce": 0.40258247221663623, "kl": 0.018301024581319325, "context": 0.4117329886994412}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 3.4264349937438965, "loss_scale": 4.0, "retries": 0, "seconds": 98.0, "max_memory_gb": 18.9, "ce": 0.5200550878071226, "kl": 0.15553018060745671, "context": 0.5978201744146645}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.0687575563788414, "loss_scale": 4.0, "retries": 0, "seconds": 98.5, "max_memory_gb": 18.9, "ce": 0.0036271603748900816, "kl": 0.03679524958715774, "context": 0.022024784877430648}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 3.6814522743225098, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 18.9, "ce": 0.18557510995196935, "kl": 0.3009194073524668, "context": 0.3360348075748334}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.4721180200576782, "loss_scale": 4.0, "retries": 0, "seconds": 99.5, "max_memory_gb": 18.9, "ce": 0.49999095776001923, "kl": 0.14598525171459187, "context": 0.5729835733654909}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.1059667095541954, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 18.9, "ce": 0.007939679305309255, "kl": 0.017605403892730465, "context": 0.01674238123723626}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.2537337839603424, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 18.9, "ce": 0.02568903347128071, "kl": 0.020205609431286575, "context": 0.035791838192380965}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.02434499002993107, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 18.9, "ce": 0.00746212423837278, "kl": 0.0058623973054636735, "context": 0.010393322969321162}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.5026471614837646, "loss_scale": 4.0, "retries": 0, "seconds": 101.6, "max_memory_gb": 18.9, "ce": 0.0011579650708881672, "kl": 0.63811155817627, "context": 0.3202137384255366}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 0.8039165735244751, "loss_scale": 4.0, "retries": 0, "seconds": 102.1, "max_memory_gb": 18.9, "ce": 0.0016732350341044366, "kl": 0.8866619597756653, "context": 0.44500421508564614}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 1.338873267173767, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 18.9, "ce": 0.06287257609074004, "kl": 0.020249730499926955, "context": 0.07299743941985071}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.60829097032547, "loss_scale": 4.0, "retries": 0, "seconds": 103.1, "max_memory_gb": 18.9, "ce": 0.1887878956113127, "kl": 0.22932188454615243, "context": 0.3034488447374315}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 1.6656337976455688, "loss_scale": 4.0, "retries": 0, "seconds": 103.6, "max_memory_gb": 18.9, "ce": 0.4980893843603553, "kl": 0.04889349005679833, "context": 0.5225361396296648}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.7379074096679688, "loss_scale": 4.0, "retries": 0, "seconds": 104.1, "max_memory_gb": 18.9, "ce": 0.06446563342615264, "kl": 0.15137608818986337, "context": 0.14015367755200714}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.02618861012160778, "loss_scale": 4.0, "retries": 0, "seconds": 104.6, "max_memory_gb": 18.9, "ce": 0.001129636897076125, "kl": 0.002711456433701187, "context": 0.002485365228039882}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.5212881565093994, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 18.9, "ce": 0.015413295579492114, "kl": 0.036313691933173686, "context": 0.03357014202629216}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.10553821176290512, "loss_scale": 4.0, "retries": 0, "seconds": 105.6, "max_memory_gb": 18.9, "ce": 0.029062301968224347, "kl": 0.013561733008828014, "context": 0.03584316815249622}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.19816187024116516, "loss_scale": 4.0, "retries": 0, "seconds": 106.1, "max_memory_gb": 18.9, "ce": 0.0012313841652940027, "kl": 0.6471413736217073, "context": 0.3248020609753439}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.03645668178796768, "loss_scale": 4.0, "retries": 0, "seconds": 106.7, "max_memory_gb": 18.9, "ce": 0.005637440283408068, "kl": 0.0027462861227149915, "context": 0.00701058359209128}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.10849815607070923, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 18.9, "ce": 0.05235523470400949, "kl": 0.07746154823689722, "context": 0.09108600884792395}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.04621731862425804, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 18.9, "ce": 0.003396777119633043, "kl": 0.0019310610587126575, "context": 0.004362307648989372}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.5562772154808044, "loss_scale": 4.0, "retries": 0, "seconds": 108.2, "max_memory_gb": 18.9, "ce": 0.13017253625730518, "kl": 0.17936423387072864, "context": 0.2198546524887206}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 0.833697497844696, "loss_scale": 4.0, "retries": 0, "seconds": 108.7, "max_memory_gb": 18.9, "ce": 0.49371422678950694, "kl": 0.1720695651265487, "context": 0.5797490223621935}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.26130619645118713, "loss_scale": 4.0, "retries": 0, "seconds": 109.2, "max_memory_gb": 18.9, "ce": 0.06848285777959973, "kl": 0.21473319749929942, "context": 0.17584945261478424}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 0.30888688564300537, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 18.9, "ce": 0.0130164111033082, "kl": 1.1525760027579963, "context": 0.5893044057302177}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.38497477769851685, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 18.9, "ce": 0.21829724741110113, "kl": 0.07414165265799966, "context": 0.255368069956603}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.5274122357368469, "loss_scale": 4.0, "retries": 0, "seconds": 110.8, "max_memory_gb": 18.9, "ce": 0.48676197799795773, "kl": 0.20962539844913408, "context": 0.5915746574319201}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.0787267684936523, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 18.9, "ce": 0.6619613019174722, "kl": 0.02875064543411554, "context": 0.6763366320847126}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.5038290619850159, "loss_scale": 4.0, "retries": 0, "seconds": 111.8, "max_memory_gb": 18.9, "ce": 0.023289493927222793, "kl": 0.1702736260012898, "context": 0.10842630879415083}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.015605837106704712, "loss_scale": 4.0, "retries": 0, "seconds": 112.3, "max_memory_gb": 18.9, "ce": 0.0007389109377982095, "kl": 0.0036901253242831444, "context": 0.0025839737318165135}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.6746407747268677, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 18.9, "ce": 0.1232754229567945, "kl": 0.316683204844594, "context": 0.28161703050136566}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.7316732406616211, "loss_scale": 4.0, "retries": 0, "seconds": 113.3, "max_memory_gb": 18.9, "ce": 0.00603720545223041, "kl": 0.18420029794651782, "context": 0.09813735759234987}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.09516873955726624, "loss_scale": 4.0, "retries": 0, "seconds": 113.8, "max_memory_gb": 18.9, "ce": 0.053841171986277914, "kl": 0.006749897277245509, "context": 0.05721612249362806}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.27868616580963135, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 18.9, "ce": 0.0035211716894991696, "kl": 0.6531330618890934, "context": 0.33008770167361945}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.1526394784450531, "loss_scale": 4.0, "retries": 0, "seconds": 114.8, "max_memory_gb": 18.9, "ce": 0.03262870054459199, "kl": 0.07715457774060042, "context": 0.07120599130576011}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 0.050615373998880386, "loss_scale": 4.0, "retries": 0, "seconds": 115.4, "max_memory_gb": 18.9, "ce": 0.0049331340997014195, "kl": 0.0004086665067006834, "context": 0.005137467393069528}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.2917841970920563, "loss_scale": 4.0, "retries": 0, "seconds": 115.9, "max_memory_gb": 18.9, "ce": 0.007510998490033671, "kl": 0.27060433051701693, "context": 0.14281316456617787}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.16860336065292358, "loss_scale": 4.0, "retries": 0, "seconds": 116.4, "max_memory_gb": 18.9, "ce": 0.0029891380545450374, "kl": 0.29247627810400445, "context": 0.14922727211524034}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 1.7404640913009644, "loss_scale": 4.0, "retries": 0, "seconds": 116.9, "max_memory_gb": 18.9, "ce": 0.20475996096502058, "kl": 0.09697077388409525, "context": 0.2532453438907396}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.32006022334098816, "loss_scale": 4.0, "retries": 0, "seconds": 117.4, "max_memory_gb": 18.9, "ce": 0.03414975168470846, "kl": 0.0064719965139374835, "context": 0.03738574889393931}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.5106392502784729, "loss_scale": 4.0, "retries": 0, "seconds": 117.9, "max_memory_gb": 18.9, "ce": 0.001849978321843082, "kl": 0.5540874825383071, "context": 0.27889372278878}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.36453601717948914, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 18.9, "ce": 0.059651681813193136, "kl": 0.010095420939023825, "context": 0.06469939041198813}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 2.380955934524536, "loss_scale": 4.0, "retries": 0, "seconds": 118.9, "max_memory_gb": 18.9, "ce": 0.46124749607406557, "kl": 0.6756113603769336, "context": 0.7990531483228551}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.1453208476305008, "loss_scale": 4.0, "retries": 0, "seconds": 119.4, "max_memory_gb": 18.9, "ce": 0.021606644202620373, "kl": 0.14553453597181942, "context": 0.09437391134270001}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.11891420185565948, "loss_scale": 4.0, "retries": 0, "seconds": 119.9, "max_memory_gb": 18.9, "ce": 0.018341325980145484, "kl": 0.020755715551786125, "context": 0.02871918270830065}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.4520490765571594, "loss_scale": 4.0, "retries": 0, "seconds": 120.4, "max_memory_gb": 18.9, "ce": 0.008189158223103732, "kl": 0.17566660372540355, "context": 0.09602246090071276}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.46901869773864746, "loss_scale": 4.0, "retries": 0, "seconds": 120.9, "max_memory_gb": 18.9, "ce": 0.015337582713982556, "kl": 0.15264665048744064, "context": 0.09166090885992162}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.251568078994751, "loss_scale": 4.0, "retries": 0, "seconds": 121.4, "max_memory_gb": 18.9, "ce": 0.36951275294995867, "kl": 0.0682515601401974, "context": 0.4036385478611919}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.038255076855421066, "loss_scale": 4.0, "retries": 0, "seconds": 121.9, "max_memory_gb": 18.9, "ce": 0.005662986498464306, "kl": 0.014127179521892685, "context": 0.012726576289423974}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.5643749833106995, "loss_scale": 4.0, "retries": 0, "seconds": 122.4, "max_memory_gb": 18.9, "ce": 0.0852058190612297, "kl": 0.053516092220434075, "context": 0.11196386168012396}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.18103115260601044, "loss_scale": 4.0, "retries": 0, "seconds": 122.9, "max_memory_gb": 18.9, "ce": 0.024069887993391603, "kl": 0.0132474588790501, "context": 0.030693617125507444}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.6106793880462646, "loss_scale": 4.0, "retries": 0, "seconds": 123.4, "max_memory_gb": 18.9, "ce": 0.10470721342426259, "kl": 0.03572518959117588, "context": 0.12256980442907661}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.19120323657989502, "loss_scale": 4.0, "retries": 0, "seconds": 123.9, "max_memory_gb": 18.9, "ce": 0.1144380296318559, "kl": 0.045730036867098534, "context": 0.13730305003991816}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 0.4426332414150238, "loss_scale": 4.0, "retries": 0, "seconds": 124.5, "max_memory_gb": 18.9, "ce": 0.07245563185279025, "kl": 0.034429699491738575, "context": 0.08967048206432082}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.09602326899766922, "loss_scale": 4.0, "retries": 0, "seconds": 125.0, "max_memory_gb": 18.9, "ce": 0.0025339525327581214, "kl": 0.08439960592886564, "context": 0.04473375596171536}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.11480620503425598, "loss_scale": 4.0, "retries": 0, "seconds": 125.5, "max_memory_gb": 18.9, "ce": 0.008107792506052647, "kl": 0.009992014744057087, "context": 0.01310379985079635}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 1.725233554840088, "loss_scale": 4.0, "retries": 0, "seconds": 126.0, "max_memory_gb": 18.9, "ce": 0.05714576208265498, "kl": 0.20412181282154052, "context": 0.15920666477177292}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 1.9835269451141357, "loss_scale": 4.0, "retries": 0, "seconds": 126.5, "max_memory_gb": 18.9, "ce": 0.36059736367315054, "kl": 0.06511531746946275, "context": 0.3931550197303295}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.023404503241181374, "loss_scale": 4.0, "retries": 0, "seconds": 127.0, "max_memory_gb": 18.9, "ce": 0.0006721935096720699, "kl": 0.0025010072049553855, "context": 0.0019226971271564253}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 1.845417857170105, "loss_scale": 4.0, "retries": 0, "seconds": 127.5, "max_memory_gb": 18.9, "ce": 0.05182857988165779, "kl": 0.6563408559773052, "context": 0.3799989995895885}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.7408720850944519, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 18.9, "ce": 0.027412584131525364, "kl": 0.06665971304755658, "context": 0.06074243946932256}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.17499426007270813, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 18.9, "ce": 0.0254914466640912, "kl": 0.031207602529320866, "context": 0.04109524900559336}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.03326103463768959, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 18.9, "ce": 0.0037128091644262895, "kl": 0.0027201432258152636, "context": 0.005072880630905274}
|
runs/single_internvl/context/L07_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L07_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.7.self_attn.q_proj",
|
| 35 |
+
"model.language_model.layers.7.self_attn.o_proj",
|
| 36 |
+
"model.language_model.layers.7.mlp.gate_proj",
|
| 37 |
+
"model.language_model.layers.7.mlp.up_proj",
|
| 38 |
+
"model.language_model.layers.7.self_attn.v_proj",
|
| 39 |
+
"model.language_model.layers.7.mlp.down_proj",
|
| 40 |
+
"model.language_model.layers.7.self_attn.k_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L07_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 7,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L07_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.8093783259391785, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 18.45, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 1.1118857860565186, "loss_scale": 4.0, "retries": 0, "seconds": 2.1, "max_memory_gb": 18.46, "ce": 0.6394089035893558, "kl": 0.46732023550248414, "context": 0.8730690349329961}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 3.2022171020507812, "loss_scale": 4.0, "retries": 0, "seconds": 2.5, "max_memory_gb": 18.46, "ce": 1.0957384780049324, "kl": 0.06678311806172132, "context": 1.129130057990551}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.6975140571594238, "loss_scale": 4.0, "retries": 0, "seconds": 3.0, "max_memory_gb": 18.46, "ce": 0.44870465776330093, "kl": 0.4310474635644823, "context": 0.6642283752880758}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.03777274489402771, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 18.46, "ce": 0.009266586348530836, "kl": 0.00829822193554719, "context": 0.01341569719807012}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.8053039312362671, "loss_scale": 4.0, "retries": 0, "seconds": 4.0, "max_memory_gb": 18.46, "ce": 0.28601379091560375, "kl": 0.07441438977548387, "context": 0.3232209905690979}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 2.581650733947754, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 18.46, "ce": 0.7955253474065103, "kl": 0.12379652759568671, "context": 0.8574236180720618}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 2.575434446334839, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 18.47, "ce": 0.7383018843829632, "kl": 0.2866892742458731, "context": 0.8816465176641941}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 1.8268576860427856, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 18.47, "ce": 0.48064903542399406, "kl": 0.13802096381550655, "context": 0.5496595203876495}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.31950241327285767, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 18.47, "ce": 0.04567332015722059, "kl": 0.04456141625996679, "context": 0.06795402627903968}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 2.13383150100708, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 18.47, "ce": 0.6065322450667736, "kl": 0.004701573754118726, "context": 0.6088830335138482}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.6697543859481812, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 18.47, "ce": 0.3901472469569853, "kl": 0.0030975250083429273, "context": 0.39169600673812965}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.2018824815750122, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 18.47, "ce": 0.35798608777986374, "kl": 0.018559971811555442, "context": 0.36726607569653424}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 2.4499354362487793, "loss_scale": 4.0, "retries": 0, "seconds": 8.3, "max_memory_gb": 18.47, "ce": 0.5771607309579849, "kl": 0.9421110753901303, "context": 1.0482162982225418}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.056537628173828, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 18.47, "ce": 0.33719271530935657, "kl": 0.44883409085105086, "context": 0.5616097756433192}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.7096481919288635, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 18.47, "ce": 0.145615695269953, "kl": 0.001062145550349669, "context": 0.1461467708068085}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.8859085440635681, "loss_scale": 4.0, "retries": 0, "seconds": 9.8, "max_memory_gb": 18.47, "ce": 0.2287517465883866, "kl": 0.8331494017038494, "context": 0.6453264351002872}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.9893585443496704, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 18.47, "ce": 0.24913179549912456, "kl": 0.2164157670995337, "context": 0.3573396782448981}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.642822504043579, "loss_scale": 4.0, "retries": 0, "seconds": 10.8, "max_memory_gb": 18.47, "ce": 0.26016734540462494, "kl": 0.18805995909497142, "context": 0.35419732332229614}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.05733498930931091, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 18.47, "ce": 0.0029466383766703075, "kl": 0.020194954404360033, "context": 0.013044115505181253}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.172355055809021, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 18.47, "ce": 0.07792943335152813, "kl": 0.02476819167713984, "context": 0.0903135278076661}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.9743797779083252, "loss_scale": 4.0, "retries": 0, "seconds": 12.3, "max_memory_gb": 18.47, "ce": 0.1020221168291755, "kl": 0.03452502120126155, "context": 0.11928463150979951}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 1.8847167491912842, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 18.47, "ce": 0.2852041699597976, "kl": 0.9630412208393864, "context": 0.7667247889357895}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 1.0312761068344116, "loss_scale": 4.0, "retries": 0, "seconds": 13.4, "max_memory_gb": 18.47, "ce": 0.07200741674751043, "kl": 0.1308149395044893, "context": 0.1374148833565414}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.0321141481399536, "loss_scale": 4.0, "retries": 0, "seconds": 13.9, "max_memory_gb": 18.47, "ce": 0.3655803151195869, "kl": 0.14621682092547417, "context": 0.4386887401342392}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 0.45271408557891846, "loss_scale": 4.0, "retries": 0, "seconds": 14.4, "max_memory_gb": 18.47, "ce": 0.04713961202651262, "kl": 0.6653379362542182, "context": 0.37980858981609344}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 0.5400599241256714, "loss_scale": 4.0, "retries": 0, "seconds": 14.9, "max_memory_gb": 18.47, "ce": 0.10073327025747858, "kl": 0.34753570626708097, "context": 0.2745011308143148}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.7054867744445801, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 18.47, "ce": 0.49164121877402067, "kl": 0.9566515039186925, "context": 0.969966946169734}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.37614408135414124, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 18.47, "ce": 0.04852468287572265, "kl": 0.39422000414924696, "context": 0.24563468131236732}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.5022361874580383, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 18.47, "ce": 0.057547370855900226, "kl": 0.0017324871332675684, "context": 0.05841361268721812}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.7347604036331177, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 18.47, "ce": 0.21313571278005838, "kl": 0.09969231765717268, "context": 0.2629818692803383}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.3916105628013611, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 18.47, "ce": 0.027991937240585685, "kl": 0.012173057330073789, "context": 0.03407846577465534}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.36052554845809937, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 18.47, "ce": 0.04952889960259199, "kl": 0.026749102980829775, "context": 0.06290345080196857}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.47559359669685364, "loss_scale": 4.0, "retries": 0, "seconds": 18.5, "max_memory_gb": 18.47, "ce": 0.009776634899026249, "kl": 0.7948973570796625, "context": 0.4072253269441717}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.15000222623348236, "loss_scale": 4.0, "retries": 0, "seconds": 19.0, "max_memory_gb": 18.47, "ce": 0.009216489386744797, "kl": 0.002456795466969197, "context": 0.010444886982440948}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.2670605480670929, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 18.47, "ce": 0.09199556682142429, "kl": 0.05818757432280108, "context": 0.12108935706783086}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.500725507736206, "loss_scale": 4.0, "retries": 0, "seconds": 20.1, "max_memory_gb": 18.47, "ce": 0.14659166455385275, "kl": 0.802193233394064, "context": 0.5476882910588756}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.3101888597011566, "loss_scale": 4.0, "retries": 0, "seconds": 20.6, "max_memory_gb": 18.47, "ce": 0.023721180390566587, "kl": 0.5037924367934465, "context": 0.27561740204691887}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.4437806308269501, "loss_scale": 4.0, "retries": 0, "seconds": 21.1, "max_memory_gb": 18.47, "ce": 0.11287622080999427, "kl": 0.011158273009641562, "context": 0.11845535911561456}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.13431020081043243, "loss_scale": 4.0, "retries": 0, "seconds": 21.6, "max_memory_gb": 18.47, "ce": 0.010678554186597466, "kl": 0.008047391718719155, "context": 0.014702249580295756}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.1641845703125, "loss_scale": 4.0, "retries": 0, "seconds": 22.0, "max_memory_gb": 18.47, "ce": 0.392167154350318, "kl": 0.12804345809854567, "context": 0.45618888223543763}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.02337644062936306, "loss_scale": 4.0, "retries": 0, "seconds": 22.5, "max_memory_gb": 18.47, "ce": 0.00304308521663188, "kl": 0.0004328434697526973, "context": 0.003259506876929663}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.04946286231279373, "loss_scale": 4.0, "retries": 0, "seconds": 23.0, "max_memory_gb": 18.47, "ce": 0.0026147825774387456, "kl": 0.0032583025604253635, "context": 0.004243933784891851}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 2.9122231006622314, "loss_scale": 4.0, "retries": 0, "seconds": 23.5, "max_memory_gb": 18.47, "ce": 0.5569813381443964, "kl": 1.4045178340747952, "context": 1.259240238883649}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.198068618774414, "loss_scale": 4.0, "retries": 0, "seconds": 24.0, "max_memory_gb": 18.47, "ce": 0.8597203533281572, "kl": 0.2382812823052518, "context": 0.9788610105460975}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.5408453941345215, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 18.47, "ce": 0.08630128490040079, "kl": 0.2700957402703352, "context": 0.22134915925562382}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.06763318181037903, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 18.47, "ce": 0.005557872689678334, "kl": 0.014055873092729598, "context": 0.012585808843141422}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.11076442152261734, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 18.47, "ce": 0.005141554705915041, "kl": 0.0031191364396363497, "context": 0.006701123085804284}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.3132251501083374, "loss_scale": 4.0, "retries": 0, "seconds": 25.8, "max_memory_gb": 18.47, "ce": 0.01125720732670743, "kl": 0.3336299625659649, "context": 0.1780721851173439}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.4029938280582428, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 18.47, "ce": 0.09431529464927735, "kl": 0.008365330199012533, "context": 0.09849795841000741}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 46.715660095214844, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 18.47, "ce": 0.6076984116989479, "kl": 0.021586551603377302, "context": 0.6184916753936704}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.6578145027160645, "loss_scale": 4.0, "retries": 0, "seconds": 27.3, "max_memory_gb": 18.47, "ce": 0.02442845405312255, "kl": 0.26572012179531157, "context": 0.15728851174935699}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.7715049982070923, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 18.47, "ce": 0.17113778955535963, "kl": 0.3719246219843626, "context": 0.3571001123636961}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5414842963218689, "loss_scale": 4.0, "retries": 0, "seconds": 28.3, "max_memory_gb": 18.47, "ce": 0.04047188819095027, "kl": 0.336333505998482, "context": 0.2086386407027021}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.09809395670890808, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 18.47, "ce": 0.016492908267537132, "kl": 0.013809049303119991, "context": 0.02339743234915659}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.2216639518737793, "loss_scale": 4.0, "retries": 0, "seconds": 29.3, "max_memory_gb": 18.47, "ce": 0.017594198463484645, "kl": 0.27544746291823685, "context": 0.15531793469563127}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.0746966153383255, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 18.47, "ce": 0.009144820862275083, "kl": 0.018905126868048683, "context": 0.01859738428174751}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.11966653168201447, "loss_scale": 4.0, "retries": 0, "seconds": 30.3, "max_memory_gb": 18.47, "ce": 0.005958992751402548, "kl": 0.0012491545649027103, "context": 0.006583570098882774}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.22818322479724884, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 18.47, "ce": 0.03305682296922896, "kl": 0.058545295119984075, "context": 0.062329469627002254}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.2093343436717987, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 18.47, "ce": 0.08145129279000685, "kl": 0.07202067828620784, "context": 0.11746162798954174}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.6857801675796509, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 18.47, "ce": 0.019649143650894985, "kl": 0.9021352854178986, "context": 0.470716788462596}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.16536346077919006, "loss_scale": 4.0, "retries": 0, "seconds": 32.3, "max_memory_gb": 18.47, "ce": 0.026928312188829295, "kl": 0.00955107274967304, "context": 0.03170384804252535}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.19693627953529358, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 18.47, "ce": 0.0353465077932924, "kl": 0.12439606052066665, "context": 0.0975445406511426}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.10437505692243576, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 18.47, "ce": 0.013355272996705025, "kl": 0.01088431925745681, "context": 0.018797432770952582}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.5225779414176941, "loss_scale": 4.0, "retries": 0, "seconds": 33.8, "max_memory_gb": 18.47, "ce": 0.18457683152519166, "kl": 0.15613333188230172, "context": 0.2626434895209968}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.7622052431106567, "loss_scale": 4.0, "retries": 0, "seconds": 34.4, "max_memory_gb": 18.47, "ce": 0.14555010496314935, "kl": 0.0686090927062537, "context": 0.1798546513745123}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.0734722763299942, "loss_scale": 4.0, "retries": 0, "seconds": 34.9, "max_memory_gb": 18.47, "ce": 0.004862985544605181, "kl": 0.008131194816087373, "context": 0.008928582887165248}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.15053582191467285, "loss_scale": 4.0, "retries": 0, "seconds": 35.4, "max_memory_gb": 18.47, "ce": 0.005566675354202744, "kl": 0.0426467159340973, "context": 0.02689003272462287}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.6685784459114075, "loss_scale": 4.0, "retries": 0, "seconds": 35.9, "max_memory_gb": 18.47, "ce": 0.2949503767013084, "kl": 0.02357760170707479, "context": 0.3067391673102975}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.022214923053979874, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 18.47, "ce": 0.0012319293891778216, "kl": 0.00026808765596797457, "context": 0.0013659731921507046}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.053524039685726166, "loss_scale": 4.0, "retries": 0, "seconds": 37.0, "max_memory_gb": 18.47, "ce": 0.007356039413934923, "kl": 0.013959219300886616, "context": 0.014335648818814661}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.2875429689884186, "loss_scale": 4.0, "retries": 0, "seconds": 37.5, "max_memory_gb": 18.47, "ce": 0.013828091352479532, "kl": 0.42454172583529726, "context": 0.2260989473434165}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.151853322982788, "loss_scale": 4.0, "retries": 0, "seconds": 37.9, "max_memory_gb": 18.47, "ce": 0.3751508732111688, "kl": 0.10019476104025671, "context": 0.4252482630145096}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.3047366738319397, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 18.47, "ce": 0.06375961776620898, "kl": 0.1182545273550204, "context": 0.12288688155240379}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.0176622867584229, "loss_scale": 4.0, "retries": 0, "seconds": 39.0, "max_memory_gb": 18.47, "ce": 0.4315510602755239, "kl": 0.9882559906691313, "context": 0.9256790648214519}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 0.8547210693359375, "loss_scale": 4.0, "retries": 0, "seconds": 39.5, "max_memory_gb": 18.47, "ce": 0.017175776963995304, "kl": 1.752604336681543, "context": 0.893477970326785}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.05206604301929474, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 18.47, "ce": 0.007572426351543982, "kl": 0.010777594541423241, "context": 0.012961223503225483}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.13143490254878998, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 18.47, "ce": 0.0015796128136571497, "kl": 0.08042647333058994, "context": 0.0417928489332553}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.22757183015346527, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 18.47, "ce": 0.015348420693044318, "kl": 0.02774356715144677, "context": 0.029220205171441194}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.15620863437652588, "loss_scale": 4.0, "retries": 0, "seconds": 41.4, "max_memory_gb": 18.47, "ce": 0.00182208668502426, "kl": 0.08379256839454285, "context": 0.04371837242888432}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.47642049193382263, "loss_scale": 4.0, "retries": 0, "seconds": 41.9, "max_memory_gb": 18.47, "ce": 0.09069060790352523, "kl": 0.08515562966931611, "context": 0.13326842186506838}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.11674942821264267, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 18.47, "ce": 0.009795729944016784, "kl": 0.002212842434346385, "context": 0.01090215120348148}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.5172737240791321, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 18.47, "ce": 0.15927229617955163, "kl": 0.7932196388283046, "context": 0.555882131884573}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.3327857553958893, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 18.47, "ce": 0.015116443001716107, "kl": 0.05847907246197792, "context": 0.04435597935025726}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 1.302587628364563, "loss_scale": 4.0, "retries": 0, "seconds": 43.9, "max_memory_gb": 18.47, "ce": 0.04602062771846249, "kl": 0.03839159014345839, "context": 0.06521642465327204}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.5218003988265991, "loss_scale": 4.0, "retries": 0, "seconds": 44.4, "max_memory_gb": 18.47, "ce": 0.23213459562975913, "kl": 0.48972089996095747, "context": 0.4769950299523771}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.7985764145851135, "loss_scale": 4.0, "retries": 0, "seconds": 44.9, "max_memory_gb": 18.47, "ce": 0.00636492388730403, "kl": 0.6809392673894763, "context": 0.34683455154299736}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.08496019244194031, "loss_scale": 4.0, "retries": 0, "seconds": 45.4, "max_memory_gb": 18.47, "ce": 0.00858509776298888, "kl": 0.02513904649524079, "context": 0.02115462124311307}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.08434867113828659, "loss_scale": 4.0, "retries": 0, "seconds": 45.9, "max_memory_gb": 18.47, "ce": 0.014910582203583544, "kl": 0.011859096877742559, "context": 0.020840131162913167}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.11765921860933304, "loss_scale": 4.0, "retries": 0, "seconds": 46.4, "max_memory_gb": 18.47, "ce": 0.009347681741928682, "kl": 0.008190628082957119, "context": 0.013442995608784258}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 0.8421795964241028, "loss_scale": 4.0, "retries": 0, "seconds": 46.9, "max_memory_gb": 18.47, "ce": 0.2099076027952833, "kl": 0.5645861868761131, "context": 0.4922006857523229}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.673797369003296, "loss_scale": 4.0, "retries": 0, "seconds": 47.4, "max_memory_gb": 18.47, "ce": 0.34511200603083125, "kl": 0.45970759821739193, "context": 0.5749658352906408}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.110667109489441, "loss_scale": 4.0, "retries": 0, "seconds": 47.9, "max_memory_gb": 18.47, "ce": 0.0614320739696268, "kl": 0.4623624675696192, "context": 0.2926133073924575}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.5186080932617188, "loss_scale": 4.0, "retries": 0, "seconds": 48.4, "max_memory_gb": 18.47, "ce": 0.19886438973480836, "kl": 0.09673322566231946, "context": 0.24723100251867436}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.5014947652816772, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 18.47, "ce": 0.6098197237733984, "kl": 0.024554834235459566, "context": 0.6220971364527941}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 0.9440236687660217, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 18.47, "ce": 0.47235715232091025, "kl": 0.006093075862736441, "context": 0.47540369984926656}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 2.8847146034240723, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 18.47, "ce": 0.020778329839231446, "kl": 0.3843791548279114, "context": 0.2129679098725319}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.11759790033102036, "loss_scale": 4.0, "retries": 0, "seconds": 50.5, "max_memory_gb": 18.47, "ce": 0.0030730943217349704, "kl": 0.09931073142433888, "context": 0.05272845933723147}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.22539950907230377, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 18.47, "ce": 0.021890117584007385, "kl": 0.012798676445527235, "context": 0.028289454650803236}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.5116224884986877, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 18.47, "ce": 0.03713291484746151, "kl": 0.6546516873713699, "context": 0.3644587595190387}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.31528812646865845, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 18.47, "ce": 0.005323405712260865, "kl": 0.08077613069326617, "context": 0.04571147006936371}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.21977216005325317, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 18.47, "ce": 0.06077445950359106, "kl": 0.03138377610594034, "context": 0.07646634825505316}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.13578081130981445, "loss_scale": 4.0, "retries": 0, "seconds": 53.0, "max_memory_gb": 18.47, "ce": 0.01157485730436747, "kl": 0.014721977823683119, "context": 0.018935846332169604}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.0392204225063324, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 18.47, "ce": 0.0033053398510674015, "kl": 0.0033978837018366903, "context": 0.005004281585570425}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.28268158435821533, "loss_scale": 4.0, "retries": 0, "seconds": 53.9, "max_memory_gb": 18.47, "ce": 0.07328182342462242, "kl": 0.36673070245888084, "context": 0.2566471677273512}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.03147194907069206, "loss_scale": 4.0, "retries": 0, "seconds": 54.4, "max_memory_gb": 18.47, "ce": 0.003556175288395025, "kl": 0.001976060142624192, "context": 0.004544205381534994}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.3548102378845215, "loss_scale": 4.0, "retries": 0, "seconds": 54.9, "max_memory_gb": 18.47, "ce": 0.5499194997137238, "kl": 0.07555806817140365, "context": 0.5876985245740798}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.4284462332725525, "loss_scale": 4.0, "retries": 0, "seconds": 55.4, "max_memory_gb": 18.47, "ce": 0.025452820671489462, "kl": 0.00640662715079543, "context": 0.028656133574259002}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.7668432593345642, "loss_scale": 4.0, "retries": 0, "seconds": 55.9, "max_memory_gb": 18.47, "ce": 0.0587937074014917, "kl": 0.3448252473026514, "context": 0.23120632953941822}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.824815034866333, "loss_scale": 4.0, "retries": 0, "seconds": 56.4, "max_memory_gb": 18.47, "ce": 0.1815209488850087, "kl": 0.3166869154665619, "context": 0.33986440440639853}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.019714536145329475, "loss_scale": 4.0, "retries": 0, "seconds": 57.0, "max_memory_gb": 18.47, "ce": 0.0016304849705193192, "kl": 0.0017540326689413632, "context": 0.002507501296349801}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.07419957220554352, "loss_scale": 4.0, "retries": 0, "seconds": 57.5, "max_memory_gb": 18.47, "ce": 0.005536388081964105, "kl": 0.0005434619197330903, "context": 0.0058081190509255975}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.3960195779800415, "loss_scale": 4.0, "retries": 0, "seconds": 58.1, "max_memory_gb": 18.47, "ce": 0.011603174498304725, "kl": 0.3744211355224252, "context": 0.19881374877877533}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.6573652029037476, "loss_scale": 4.0, "retries": 0, "seconds": 58.6, "max_memory_gb": 18.47, "ce": 0.018109203840140253, "kl": 1.2804102139198221, "context": 0.658314295753371}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.12562158703804016, "loss_scale": 4.0, "retries": 0, "seconds": 59.1, "max_memory_gb": 18.47, "ce": 0.009458713509957306, "kl": 0.04864099810947664, "context": 0.03377921204082668}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 1.1433476209640503, "loss_scale": 4.0, "retries": 0, "seconds": 59.6, "max_memory_gb": 18.47, "ce": 0.15349731870082906, "kl": 0.5192712670450419, "context": 0.41313296713269665}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 1.641991376876831, "loss_scale": 4.0, "retries": 0, "seconds": 60.1, "max_memory_gb": 18.48, "ce": 0.076166234590346, "kl": 0.12657512477017008, "context": 0.13945379697543103}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.42933139204978943, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 18.48, "ce": 0.1761109207291156, "kl": 0.011590738147788215, "context": 0.18190628502634354}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.1631118804216385, "loss_scale": 4.0, "retries": 0, "seconds": 61.1, "max_memory_gb": 18.48, "ce": 0.01579053325986024, "kl": 0.016972878896012844, "context": 0.024276972137158737}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.520583152770996, "loss_scale": 4.0, "retries": 0, "seconds": 61.7, "max_memory_gb": 18.48, "ce": 0.6038629712711554, "kl": 0.19555335966288112, "context": 0.701639641833026}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.25486665964126587, "loss_scale": 4.0, "retries": 0, "seconds": 62.2, "max_memory_gb": 18.48, "ce": 0.051018117263083695, "kl": 0.039202583764563315, "context": 0.0706194086742471}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.3011314570903778, "loss_scale": 4.0, "retries": 0, "seconds": 62.7, "max_memory_gb": 18.48, "ce": 0.06282250296862912, "kl": 0.01569112250035687, "context": 0.07066806503280532}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.20321784913539886, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 18.48, "ce": 0.03803774400148541, "kl": 0.022566229396034032, "context": 0.04932085843756795}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.5511012077331543, "loss_scale": 4.0, "retries": 0, "seconds": 63.7, "max_memory_gb": 18.48, "ce": 0.67026973614702, "kl": 0.6315400077619415, "context": 0.9860397577576805}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.3709392845630646, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 18.48, "ce": 0.10706060426309705, "kl": 0.09530479708337225, "context": 0.15471300238277763}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.909188985824585, "loss_scale": 4.0, "retries": 0, "seconds": 64.7, "max_memory_gb": 18.48, "ce": 0.5171339027583599, "kl": 0.1567954858765006, "context": 0.5955316387116909}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.11333930492401123, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 18.48, "ce": 0.015565007590339519, "kl": 0.0018970690962305525, "context": 0.016513541813765187}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 2.200598955154419, "loss_scale": 4.0, "retries": 0, "seconds": 65.7, "max_memory_gb": 18.48, "ce": 0.29963945248164237, "kl": 1.237276355837821, "context": 0.9182776233647019}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.23478715121746063, "loss_scale": 4.0, "retries": 0, "seconds": 66.2, "max_memory_gb": 18.48, "ce": 0.07120763976126909, "kl": 0.03165793813968776, "context": 0.08703660877654329}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.2094045728445053, "loss_scale": 4.0, "retries": 0, "seconds": 66.7, "max_memory_gb": 18.48, "ce": 0.04055800673086196, "kl": 0.2729901996863191, "context": 0.17705310211749747}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 2.056617498397827, "loss_scale": 4.0, "retries": 0, "seconds": 67.2, "max_memory_gb": 18.48, "ce": 0.21293925555073656, "kl": 0.360022222797852, "context": 0.3929503664840013}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.3060912489891052, "loss_scale": 4.0, "retries": 0, "seconds": 67.7, "max_memory_gb": 18.48, "ce": 0.16004177899594652, "kl": 0.029432747036480578, "context": 0.1747581571689807}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.5598915219306946, "loss_scale": 4.0, "retries": 0, "seconds": 68.3, "max_memory_gb": 18.48, "ce": 0.3408675689424854, "kl": 0.20099297957494855, "context": 0.44136404106393456}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.5206038951873779, "loss_scale": 4.0, "retries": 0, "seconds": 68.8, "max_memory_gb": 18.48, "ce": 0.22546121165032673, "kl": 0.005678557866843903, "context": 0.22830048639661982}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.11897749453783035, "loss_scale": 4.0, "retries": 0, "seconds": 69.2, "max_memory_gb": 18.48, "ce": 0.02147461156710051, "kl": 0.10277448336637462, "context": 0.07286185125121847}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 2.968893527984619, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 18.48, "ce": 0.15398396574892104, "kl": 0.1928658450488001, "context": 0.2504168879240751}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.13729828596115112, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 18.48, "ce": 0.0073618115711724386, "kl": 0.019591158416005783, "context": 0.01715739065548405}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 1.714805006980896, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 18.48, "ce": 0.07896457915194333, "kl": 0.1552924821153283, "context": 0.15661082230508327}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.901220977306366, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 18.48, "ce": 0.004033021301438566, "kl": 0.33815488692016515, "context": 0.17311046359100146}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.05156850442290306, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 18.48, "ce": 0.004324786365032196, "kl": 0.0025793560926103964, "context": 0.005614464404061437}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.03313900902867317, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 18.48, "ce": 0.004514145766734146, "kl": 0.005085815952043049, "context": 0.007057053619064391}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.7064884901046753, "loss_scale": 4.0, "retries": 0, "seconds": 72.7, "max_memory_gb": 18.48, "ce": 0.048353448015404865, "kl": 0.02033207262866199, "context": 0.058519483427517116}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.9200443029403687, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 18.48, "ce": 0.5190599969537288, "kl": 0.06423202747100731, "context": 0.5511759969667764}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.5954647660255432, "loss_scale": 4.0, "retries": 0, "seconds": 73.8, "max_memory_gb": 18.48, "ce": 0.08246875752229244, "kl": 0.20241073536453769, "context": 0.18367412383668125}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.04557415470480919, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 18.48, "ce": 0.0018670876743271947, "kl": 0.009922206142164214, "context": 0.006828190489613917}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.08847998827695847, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 18.48, "ce": 0.0075811108108609915, "kl": 0.023227424737342517, "context": 0.019194823398720473}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 0.8320990204811096, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 18.48, "ce": 0.4815437728902907, "kl": 0.5561051784679876, "context": 0.7595963802887127}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.26666635274887085, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 18.48, "ce": 0.007391088784970634, "kl": 0.011212533107027411, "context": 0.012997355366678676}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.19573387503623962, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 18.48, "ce": 0.01642787962919101, "kl": 0.007548341367510147, "context": 0.020202050800435245}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.24380525946617126, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 18.48, "ce": 0.06791000009980053, "kl": 0.02571185422129929, "context": 0.08076592744328082}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.2544405162334442, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 18.48, "ce": 0.017481092712841928, "kl": 0.028456589043344138, "context": 0.03170938816765556}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.40415701270103455, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 18.48, "ce": 0.007578434218885377, "kl": 0.5964575107809651, "context": 0.30580719927093014}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.33012881875038147, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 18.48, "ce": 0.01234734195531928, "kl": 0.38990626876875467, "context": 0.20730047494362225}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 1.8082482814788818, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 18.48, "ce": 0.10562248883070424, "kl": 0.09251302533084527, "context": 0.15187900385353714}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.35463082790374756, "loss_scale": 4.0, "retries": 0, "seconds": 79.4, "max_memory_gb": 18.48, "ce": 0.19704864605773764, "kl": 0.010845071366929915, "context": 0.2024711868380109}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.882461667060852, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 18.48, "ce": 0.012594512467330787, "kl": 0.05573157912112947, "context": 0.04046030106837861}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.7909230589866638, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 18.48, "ce": 0.04473133571445942, "kl": 0.04146143932302948, "context": 0.06546205515041947}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.38760679960250854, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 18.48, "ce": 0.1724336081861111, "kl": 0.07214322595245903, "context": 0.20850522216642275}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.6634434461593628, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 18.48, "ce": 0.00503391501842998, "kl": 0.09123549549258314, "context": 0.05065166356507689}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.7087337374687195, "loss_scale": 4.0, "retries": 0, "seconds": 82.0, "max_memory_gb": 18.48, "ce": 0.024078057322185487, "kl": 0.3684059272054583, "context": 0.20828101428924128}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.39991867542266846, "loss_scale": 4.0, "retries": 0, "seconds": 82.5, "max_memory_gb": 18.48, "ce": 0.07022868914646097, "kl": 0.08408656355823041, "context": 0.11227197095286101}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.1059890985488892, "loss_scale": 4.0, "retries": 0, "seconds": 83.0, "max_memory_gb": 18.48, "ce": 0.08983841762528755, "kl": 0.16737416217802092, "context": 0.17352549868519418}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.0623507499694824, "loss_scale": 4.0, "retries": 0, "seconds": 83.5, "max_memory_gb": 18.48, "ce": 0.23777518869610503, "kl": 0.7693347124441061, "context": 0.6224425446707755}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 5.481932163238525, "loss_scale": 4.0, "retries": 0, "seconds": 84.0, "max_memory_gb": 18.48, "ce": 0.07998892768227961, "kl": 0.14725943698431365, "context": 0.15361865211161785}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.0833820179104805, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 18.48, "ce": 0.004915788544167299, "kl": 0.02803207685610687, "context": 0.018931827318738215}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.5972102880477905, "loss_scale": 4.0, "retries": 0, "seconds": 85.0, "max_memory_gb": 18.48, "ce": 0.4007855278905481, "kl": 0.14100711117498577, "context": 0.4712890723021701}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.19399553537368774, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 18.48, "ce": 0.024289147928357124, "kl": 0.0030636059236712754, "context": 0.025820950977504253}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.1141756922006607, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 18.48, "ce": 0.028090385341783985, "kl": 0.002329969407583121, "context": 0.029255369110614993}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.0624289512634277, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 18.48, "ce": 0.08839079475728795, "kl": 1.4561438858509064, "context": 0.8164627403020859}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.02815074287354946, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 18.48, "ce": 0.0017046406064764597, "kl": 0.0006126953485363629, "context": 0.0020109883043915033}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.7870268821716309, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 18.48, "ce": 0.023741157552649383, "kl": 0.26435705681797117, "context": 0.15591968037188053}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.11187440901994705, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 18.48, "ce": 0.016224606661126018, "kl": 0.017374750575982034, "context": 0.0249119809595868}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.16961349546909332, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 18.48, "ce": 0.04984113515820354, "kl": 0.02850122284507961, "context": 0.06409174669533968}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.21724176406860352, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 18.48, "ce": 0.03695670250090188, "kl": 0.03445829745533047, "context": 0.05418585029838141}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 2.2885775566101074, "loss_scale": 4.0, "retries": 0, "seconds": 89.6, "max_memory_gb": 18.48, "ce": 0.3626390206627548, "kl": 0.08505502610933036, "context": 0.405166526325047}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.05461809039115906, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 18.48, "ce": 0.009475256942096166, "kl": 0.002786136445138254, "context": 0.010868325014598668}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.6214926838874817, "loss_scale": 4.0, "retries": 0, "seconds": 90.6, "max_memory_gb": 18.48, "ce": 0.015435282271937467, "kl": 0.8021249046396406, "context": 0.41649773040262517}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.8385277390480042, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 18.48, "ce": 0.07241985358996317, "kl": 0.1369177869783016, "context": 0.14087874698452652}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.7279931902885437, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 18.48, "ce": 0.34662664375355234, "kl": 0.025103338151893695, "context": 0.3591783100782777}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.1678895652294159, "loss_scale": 4.0, "retries": 0, "seconds": 92.1, "max_memory_gb": 18.48, "ce": 0.039256186079001054, "kl": 0.023873618727520807, "context": 0.05119299543730449}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.1422039419412613, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 18.48, "ce": 0.00827909167855978, "kl": 0.08995465171756223, "context": 0.05325641774106771}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 1.289318323135376, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 18.48, "ce": 0.0227274437493179, "kl": 0.3774454868016619, "context": 0.21145019121468067}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.025013310834765434, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 18.48, "ce": 0.002547097467868298, "kl": 0.000693658585078083, "context": 0.002893926764954813}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.03234568610787392, "loss_scale": 4.0, "retries": 0, "seconds": 94.1, "max_memory_gb": 18.48, "ce": 0.00232083832815988, "kl": 0.003070948931053863, "context": 0.0038563127018278465}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.15873751044273376, "loss_scale": 4.0, "retries": 0, "seconds": 94.8, "max_memory_gb": 18.48, "ce": 0.009548407047986984, "kl": 0.09272348387821694, "context": 0.055910148948896676}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 2.3909823894500732, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 18.48, "ce": 0.1382546922777692, "kl": 0.33601366374387, "context": 0.30626152304466814}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.3072265684604645, "loss_scale": 4.0, "retries": 0, "seconds": 95.9, "max_memory_gb": 18.48, "ce": 0.08731738434289582, "kl": 0.48500519804656506, "context": 0.3298199992859736}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.5507251024246216, "loss_scale": 4.0, "retries": 0, "seconds": 96.5, "max_memory_gb": 18.48, "ce": 0.27495803195779445, "kl": 0.014098167657721206, "context": 0.28200710740929935}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.4446018934249878, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 18.48, "ce": 0.3674656613729894, "kl": 0.29938407510053366, "context": 0.5171577092260122}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.098702073097229, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 18.48, "ce": 0.004331600328441709, "kl": 0.04101794329471886, "context": 0.024840571684762836}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 4.276054859161377, "loss_scale": 4.0, "retries": 0, "seconds": 98.1, "max_memory_gb": 18.48, "ce": 0.27299035679607186, "kl": 0.2906342129949735, "context": 0.41830746359482873}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.0149080753326416, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 18.48, "ce": 0.4552980269945692, "kl": 0.09031485590094235, "context": 0.5004554530896712}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.07238470762968063, "loss_scale": 4.0, "retries": 0, "seconds": 99.3, "max_memory_gb": 18.48, "ce": 0.004788070593349403, "kl": 0.007698922227064031, "context": 0.008637531707790913}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.3743046820163727, "loss_scale": 4.0, "retries": 0, "seconds": 99.9, "max_memory_gb": 18.48, "ce": 0.034316368168219924, "kl": 0.033495755167678, "context": 0.05106424575205892}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.06260156631469727, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 18.48, "ce": 0.010247645775962155, "kl": 0.00505516659359273, "context": 0.012775229159160517}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.4795375466346741, "loss_scale": 4.0, "retries": 0, "seconds": 101.2, "max_memory_gb": 18.48, "ce": 0.001172544683868182, "kl": 0.46261500919285936, "context": 0.232480055963606}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.0205079317092896, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 18.48, "ce": 0.00388504151487723, "kl": 0.8052141382363516, "context": 0.4064920967211947}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.3217482268810272, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 18.48, "ce": 0.052459574602835346, "kl": 0.0095133573977364, "context": 0.05721625275327824}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.5000986456871033, "loss_scale": 4.0, "retries": 0, "seconds": 102.7, "max_memory_gb": 18.48, "ce": 0.1528206932125613, "kl": 0.18004974109499017, "context": 0.2428455641202163}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 1.2796674966812134, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 18.48, "ce": 0.3462075819261372, "kl": 0.043599960874416865, "context": 0.368007569340989}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.8780553340911865, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 18.48, "ce": 0.0834197456279071, "kl": 0.14039808964298572, "context": 0.153618790442124}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.040122877806425095, "loss_scale": 4.0, "retries": 0, "seconds": 104.4, "max_memory_gb": 18.48, "ce": 0.0016855299609233043, "kl": 0.002875573779760998, "context": 0.0031233167655955185}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 1.0846788883209229, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 18.48, "ce": 0.044176101495395415, "kl": 0.08100443438161165, "context": 0.08467832248425111}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.1762850433588028, "loss_scale": 4.0, "retries": 0, "seconds": 105.4, "max_memory_gb": 18.48, "ce": 0.03783526920597069, "kl": 0.013019267586059868, "context": 0.04434490104904398}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.327686071395874, "loss_scale": 4.0, "retries": 0, "seconds": 105.9, "max_memory_gb": 18.48, "ce": 0.00420107546960935, "kl": 0.5703759284806438, "context": 0.28938904823735356}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.05591459199786186, "loss_scale": 4.0, "retries": 0, "seconds": 106.5, "max_memory_gb": 18.48, "ce": 0.007303096277610166, "kl": 0.0035363587658139295, "context": 0.009071275864698691}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.16052646934986115, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 18.48, "ce": 0.06158375780796632, "kl": 0.06271295726764947, "context": 0.0929402363835834}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.04686427116394043, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 18.48, "ce": 0.004879402556980494, "kl": 0.0019712028515641578, "context": 0.005865004161023535}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.7455464005470276, "loss_scale": 4.0, "retries": 0, "seconds": 108.4, "max_memory_gb": 18.48, "ce": 0.15362401772290468, "kl": 0.12469581266850582, "context": 0.2159719312330708}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 2.718151092529297, "loss_scale": 4.0, "retries": 0, "seconds": 109.0, "max_memory_gb": 18.48, "ce": 0.5118235645477398, "kl": 0.1541785546214669, "context": 0.5889128260387224}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.28620991110801697, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 18.48, "ce": 0.12642978789517656, "kl": 0.13243705336935818, "context": 0.192648321040906}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 0.64717698097229, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 18.48, "ce": 0.037527224281802773, "kl": 1.0371684014244238, "context": 0.5561114144511521}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.5169382095336914, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 18.48, "ce": 0.21109288508887403, "kl": 0.05257501650339691, "context": 0.23738039321324322}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.6128838062286377, "loss_scale": 4.0, "retries": 0, "seconds": 111.5, "max_memory_gb": 18.48, "ce": 0.41018887985046604, "kl": 0.18282414553686976, "context": 0.5016009469109122}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.3910348415374756, "loss_scale": 4.0, "retries": 0, "seconds": 112.0, "max_memory_gb": 18.48, "ce": 0.647135672665172, "kl": 0.025282936079577212, "context": 0.6597771421293146}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.2637961208820343, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 18.48, "ce": 0.013972432992886752, "kl": 0.12051725946366787, "context": 0.0742310638888739}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.028820227831602097, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 18.48, "ce": 0.0014809537678956985, "kl": 0.0027204033694943064, "context": 0.002841155437636189}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.6101322770118713, "loss_scale": 4.0, "retries": 0, "seconds": 113.7, "max_memory_gb": 18.48, "ce": 0.12720654136501253, "kl": 0.32248619198799133, "context": 0.28844963386654854}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.6405938267707825, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 18.48, "ce": 0.0088790397712728, "kl": 0.1375268844421953, "context": 0.07764248494640924}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.1925240159034729, "loss_scale": 4.0, "retries": 0, "seconds": 114.9, "max_memory_gb": 18.48, "ce": 0.06276771388365887, "kl": 0.008363311424545827, "context": 0.06694937051361194}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.3475606143474579, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 18.48, "ce": 0.006520731549244374, "kl": 0.5460272276650358, "context": 0.27953434400842525}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.2101351022720337, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 18.48, "ce": 0.02677522636804497, "kl": 0.08141629529404781, "context": 0.06748337263707072}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 0.6011906862258911, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 18.48, "ce": 0.03204498174454784, "kl": 0.004513380124990363, "context": 0.034301672043511644}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.28707122802734375, "loss_scale": 4.0, "retries": 0, "seconds": 117.3, "max_memory_gb": 18.48, "ce": 0.016778119636001065, "kl": 0.21726807267077675, "context": 0.12541215957025997}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.22778065502643585, "loss_scale": 4.0, "retries": 0, "seconds": 117.9, "max_memory_gb": 18.48, "ce": 0.004564971182844602, "kl": 0.23401332221692428, "context": 0.1215716328297276}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 1.1402720212936401, "loss_scale": 4.0, "retries": 0, "seconds": 118.5, "max_memory_gb": 18.48, "ce": 0.060270661633694544, "kl": 0.04909426836820785, "context": 0.08481779383146204}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.24906586110591888, "loss_scale": 4.0, "retries": 0, "seconds": 119.1, "max_memory_gb": 18.48, "ce": 0.021511676959562465, "kl": 0.005170211436052341, "context": 0.024096781791740796}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.4489878714084625, "loss_scale": 4.0, "retries": 0, "seconds": 119.7, "max_memory_gb": 18.48, "ce": 0.00347842117480468, "kl": 0.3321473251271527, "context": 0.16955207925639115}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.35518115758895874, "loss_scale": 4.0, "retries": 0, "seconds": 120.3, "max_memory_gb": 18.48, "ce": 0.0560305784092634, "kl": 0.01641975612119495, "context": 0.06424045750463847}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.6344941854476929, "loss_scale": 4.0, "retries": 0, "seconds": 120.8, "max_memory_gb": 18.48, "ce": 0.5366228933780803, "kl": 0.4324381945771165, "context": 0.7528420169837773}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.1975729912519455, "loss_scale": 4.0, "retries": 0, "seconds": 121.4, "max_memory_gb": 18.48, "ce": 0.0290237508379505, "kl": 0.13796918481239118, "context": 0.09800834184716223}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.12636221945285797, "loss_scale": 4.0, "retries": 0, "seconds": 122.0, "max_memory_gb": 18.48, "ce": 0.02408856590045616, "kl": 0.011263537511695176, "context": 0.029720335558522493}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.3503357470035553, "loss_scale": 4.0, "retries": 0, "seconds": 122.6, "max_memory_gb": 18.48, "ce": 0.008281099260784686, "kl": 0.12123468739446253, "context": 0.06889844150282443}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.4642844796180725, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 18.48, "ce": 0.011010012807673775, "kl": 0.09874151970871026, "context": 0.06038077351695392}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.3751190900802612, "loss_scale": 4.0, "retries": 0, "seconds": 123.8, "max_memory_gb": 18.48, "ce": 0.3610643131978577, "kl": 0.0269700711742189, "context": 0.3745493347378215}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.06759192794561386, "loss_scale": 4.0, "retries": 0, "seconds": 124.4, "max_memory_gb": 18.48, "ce": 0.008868780307238922, "kl": 0.017959263150260085, "context": 0.017848411764134653}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.44072625041007996, "loss_scale": 4.0, "retries": 0, "seconds": 125.0, "max_memory_gb": 18.48, "ce": 0.07132689845457207, "kl": 0.030052232012167224, "context": 0.08635301428148523}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.3275434374809265, "loss_scale": 4.0, "retries": 0, "seconds": 125.6, "max_memory_gb": 18.48, "ce": 0.04027975129429251, "kl": 0.011146640716106049, "context": 0.04585307210800238}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 1.0911821126937866, "loss_scale": 4.0, "retries": 0, "seconds": 126.2, "max_memory_gb": 18.48, "ce": 0.10632484682719223, "kl": 0.03748423696379177, "context": 0.12506696159834974}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.22269682586193085, "loss_scale": 4.0, "retries": 0, "seconds": 126.8, "max_memory_gb": 18.48, "ce": 0.09390964270278346, "kl": 0.042945090663124574, "context": 0.11538219268550165}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 0.944218635559082, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 18.48, "ce": 0.10326486550911795, "kl": 0.0236517277371604, "context": 0.11509073310298845}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.11738599836826324, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 18.48, "ce": 0.00514639164248365, "kl": 0.06342997194406053, "context": 0.03686137877230067}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.06359083205461502, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 18.48, "ce": 0.004894513578619808, "kl": 0.005397318949690089, "context": 0.007593173126224428}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.5815021991729736, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 18.48, "ce": 0.04306828533299267, "kl": 0.18768466060282663, "context": 0.13691061505232938}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 1.6499794721603394, "loss_scale": 4.0, "retries": 0, "seconds": 129.5, "max_memory_gb": 18.48, "ce": 0.3805263619287871, "kl": 0.09261178353335708, "context": 0.4268322443822399}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.021813513711094856, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 18.48, "ce": 0.0012082168614142574, "kl": 0.0008224753655667882, "context": 0.001619454546016641}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.40018129348754883, "loss_scale": 4.0, "retries": 0, "seconds": 130.5, "max_memory_gb": 18.48, "ce": 0.006180022421176545, "kl": 0.4863664470394724, "context": 0.24936325060843956}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.1977994441986084, "loss_scale": 4.0, "retries": 0, "seconds": 131.0, "max_memory_gb": 18.48, "ce": 0.031648120551835746, "kl": 0.015182808972895145, "context": 0.039239524165168405}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.29526418447494507, "loss_scale": 4.0, "retries": 0, "seconds": 131.5, "max_memory_gb": 18.48, "ce": 0.03710743598639965, "kl": 0.044983799307374284, "context": 0.05959933763369918}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.05052896961569786, "loss_scale": 4.0, "retries": 0, "seconds": 132.0, "max_memory_gb": 18.48, "ce": 0.005944758144323714, "kl": 0.003194412376615219, "context": 0.007541964034317061}
|
runs/single_internvl/context/L09_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L09_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.9.self_attn.k_proj",
|
| 35 |
+
"model.language_model.layers.9.mlp.up_proj",
|
| 36 |
+
"model.language_model.layers.9.self_attn.q_proj",
|
| 37 |
+
"model.language_model.layers.9.self_attn.v_proj",
|
| 38 |
+
"model.language_model.layers.9.mlp.down_proj",
|
| 39 |
+
"model.language_model.layers.9.self_attn.o_proj",
|
| 40 |
+
"model.language_model.layers.9.mlp.gate_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L09_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 9,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L09_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.8271442651748657, "loss_scale": 4.0, "retries": 0, "seconds": 1.6, "max_memory_gb": 18.29, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 1.0194190740585327, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 18.3, "ce": 0.6370677337035886, "kl": 0.4671201559073097, "context": 0.8706278188183205}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 3.2324845790863037, "loss_scale": 4.0, "retries": 0, "seconds": 2.5, "max_memory_gb": 18.3, "ce": 1.0975014567375183, "kl": 0.0678083086386323, "context": 1.131405621767044}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.7054495811462402, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 18.3, "ce": 0.44897803019557614, "kl": 0.43064634570419, "context": 0.6643011873020441}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.03915682062506676, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 18.3, "ce": 0.009200214844895527, "kl": 0.00823867807048373, "context": 0.013319553938345052}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.815216064453125, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 18.3, "ce": 0.28358335938537493, "kl": 0.07536936150427209, "context": 0.3212680385040585}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 2.7183189392089844, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 18.3, "ce": 0.8022173918434419, "kl": 0.12636772719349665, "context": 0.8654012361148489}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 2.67185640335083, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 18.31, "ce": 0.7379385884851217, "kl": 0.28558468469418585, "context": 0.8807309456169605}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 1.8745927810668945, "loss_scale": 4.0, "retries": 0, "seconds": 5.6, "max_memory_gb": 18.31, "ce": 0.4797918666154146, "kl": 0.13590747630223632, "context": 0.5477456115186214}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.317951500415802, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 18.31, "ce": 0.044975467906624544, "kl": 0.04510957468301058, "context": 0.06753025599755347}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 2.280679225921631, "loss_scale": 4.0, "retries": 0, "seconds": 6.6, "max_memory_gb": 18.31, "ce": 0.5959243912948295, "kl": 0.004643247368221637, "context": 0.5982459975057282}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.6529954671859741, "loss_scale": 4.0, "retries": 0, "seconds": 7.1, "max_memory_gb": 18.31, "ce": 0.3827545993444801, "kl": 0.00319813551095649, "context": 0.3843536680233228}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.248725414276123, "loss_scale": 4.0, "retries": 0, "seconds": 7.6, "max_memory_gb": 18.31, "ce": 0.3522583464255149, "kl": 0.017397199781044037, "context": 0.36095693572133314}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 2.579228401184082, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 18.31, "ce": 0.5509330034255981, "kl": 0.9350451952777803, "context": 1.018455609679222}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.0880556106567383, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 18.31, "ce": 0.31610093341691936, "kl": 0.44536169442289975, "context": 0.5387817787511722}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.759891927242279, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 18.31, "ce": 0.12891674086131388, "kl": 0.0012284926854135847, "context": 0.12953098172874888}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.8608388900756836, "loss_scale": 4.0, "retries": 0, "seconds": 9.5, "max_memory_gb": 18.31, "ce": 0.2140708527294919, "kl": 0.8308871428889688, "context": 0.6295144269824959}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.788520336151123, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 18.31, "ce": 0.19873233897669706, "kl": 0.21671955438068835, "context": 0.3070921176113188}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.068955421447754, "loss_scale": 4.0, "retries": 0, "seconds": 10.5, "max_memory_gb": 18.31, "ce": 0.1890895813703537, "kl": 0.18551677279174328, "context": 0.2818479724228382}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.0597178116440773, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 18.31, "ce": 0.003075766884649056, "kl": 0.01989615909224085, "context": 0.013023846626310842}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 0.6678013205528259, "loss_scale": 4.0, "retries": 0, "seconds": 11.5, "max_memory_gb": 18.31, "ce": 0.04003822491631581, "kl": 0.02181251803813211, "context": 0.05094448488125636}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.7086774706840515, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 18.31, "ce": 0.07449555552011589, "kl": 0.030170661382726394, "context": 0.08958088318468072}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 1.7407026290893555, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 18.31, "ce": 0.2526570201262075, "kl": 0.944913686374548, "context": 0.7251138626734246}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 0.7018343806266785, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 18.31, "ce": 0.04862230969592929, "kl": 0.10348979581613094, "context": 0.10036720940843225}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 0.8349769115447998, "loss_scale": 4.0, "retries": 0, "seconds": 13.5, "max_memory_gb": 18.31, "ce": 0.3307002711262612, "kl": 0.14427863294258714, "context": 0.4028395861387253}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 0.3110015094280243, "loss_scale": 4.0, "retries": 0, "seconds": 14.0, "max_memory_gb": 18.31, "ce": 0.0335920515935868, "kl": 0.660844745580107, "context": 0.3640144132077694}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 0.33323943614959717, "loss_scale": 4.0, "retries": 0, "seconds": 14.5, "max_memory_gb": 18.31, "ce": 0.08659846185764764, "kl": 0.3315321889349434, "context": 0.2523645634064451}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.7230067849159241, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 18.31, "ce": 0.47781292255967855, "kl": 0.9552468643523753, "context": 0.9554363619536161}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.49240514636039734, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 18.31, "ce": 0.04400980763603002, "kl": 0.37875144119607285, "context": 0.2333855254109949}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.357898086309433, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 18.31, "ce": 0.03463703545094177, "kl": 0.0013297466712174355, "context": 0.0353019089470763}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.6178283095359802, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 18.31, "ce": 0.19598457869142294, "kl": 0.09698183147702366, "context": 0.24447548761963844}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.2658081650733948, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 18.31, "ce": 0.01752031990326941, "kl": 0.010145730506337713, "context": 0.022593185771256685}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.2661970555782318, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 18.31, "ce": 0.04412325518205762, "kl": 0.02139992080628872, "context": 0.05482321698218584}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.4194871187210083, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 18.31, "ce": 0.006565787218278274, "kl": 0.7858887769825742, "context": 0.39951018022111384}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.08731282502412796, "loss_scale": 4.0, "retries": 0, "seconds": 18.4, "max_memory_gb": 18.31, "ce": 0.005038796691223979, "kl": 0.0018180009569732647, "context": 0.005947797151748091}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.25056618452072144, "loss_scale": 4.0, "retries": 0, "seconds": 18.9, "max_memory_gb": 18.31, "ce": 0.08494866860564798, "kl": 0.05330599605804309, "context": 0.11160166485933587}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.442525714635849, "loss_scale": 4.0, "retries": 0, "seconds": 19.4, "max_memory_gb": 18.31, "ce": 0.1357577591552399, "kl": 0.7879719984484836, "context": 0.5297437568660825}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.3238140344619751, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 18.31, "ce": 0.02170454617589712, "kl": 0.48757938639028, "context": 0.26549423253163695}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.509132444858551, "loss_scale": 4.0, "retries": 0, "seconds": 20.4, "max_memory_gb": 18.31, "ce": 0.11005407731136074, "kl": 0.010317796091840137, "context": 0.11521297291619703}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.12750421464443207, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 18.31, "ce": 0.010254015534883365, "kl": 0.008598726795753464, "context": 0.014553379034623504}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.2191094160079956, "loss_scale": 4.0, "retries": 0, "seconds": 21.4, "max_memory_gb": 18.31, "ce": 0.38792001543333754, "kl": 0.12753031472675502, "context": 0.45168516458943486}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.025241464376449585, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 18.31, "ce": 0.0034998498886125162, "kl": 0.00039334110624622554, "context": 0.0036965205217711627}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.049043942242860794, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 18.31, "ce": 0.002619688202685211, "kl": 0.0026478303771000355, "context": 0.003943603434890974}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 3.00893497467041, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 18.31, "ce": 0.5377192363957874, "kl": 1.3693463196468656, "context": 1.22239238320617}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.2965245246887207, "loss_scale": 4.0, "retries": 0, "seconds": 23.4, "max_memory_gb": 18.31, "ce": 0.8312955310684629, "kl": 0.23856686468207045, "context": 0.9505789534887299}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.440765917301178, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 18.31, "ce": 0.08233691944042221, "kl": 0.2625017975224182, "context": 0.21358781831804663}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.0677410289645195, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 18.31, "ce": 0.0059955336328130215, "kl": 0.01629141572630033, "context": 0.014141241437755525}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.08412934839725494, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 18.31, "ce": 0.0035189492045901716, "kl": 0.0032775604850030504, "context": 0.0051577293779700994}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.34749090671539307, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 18.31, "ce": 0.01230293919070391, "kl": 0.3123348592353068, "context": 0.1684703636856284}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.5234605669975281, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 18.31, "ce": 0.09955064213863807, "kl": 0.00804259459982859, "context": 0.10357194223615807}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.5784813165664673, "loss_scale": 4.0, "retries": 0, "seconds": 26.4, "max_memory_gb": 18.31, "ce": 0.5286470650771662, "kl": 0.018678359187106253, "context": 0.5379862242989475}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.6477727890014648, "loss_scale": 4.0, "retries": 0, "seconds": 26.9, "max_memory_gb": 18.31, "ce": 0.022461388492956758, "kl": 0.24887905095238239, "context": 0.14690091158263385}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.6637927889823914, "loss_scale": 4.0, "retries": 0, "seconds": 27.4, "max_memory_gb": 18.31, "ce": 0.1681324343662709, "kl": 0.37186820479109883, "context": 0.3540665372274816}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5234547257423401, "loss_scale": 4.0, "retries": 0, "seconds": 27.9, "max_memory_gb": 18.31, "ce": 0.039591327440575697, "kl": 0.3379093054754776, "context": 0.20854597924335394}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.11448168754577637, "loss_scale": 4.0, "retries": 0, "seconds": 28.4, "max_memory_gb": 18.31, "ce": 0.01837207085918635, "kl": 0.014111172952013362, "context": 0.025427657819818705}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.23100119829177856, "loss_scale": 4.0, "retries": 0, "seconds": 28.9, "max_memory_gb": 18.31, "ce": 0.020300200441852212, "kl": 0.25196727947331965, "context": 0.14628384495154023}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.09128088504076004, "loss_scale": 4.0, "retries": 0, "seconds": 29.4, "max_memory_gb": 18.31, "ce": 0.0099630608165171, "kl": 0.020286070295696845, "context": 0.02010609688295517}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.08040177077054977, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 18.31, "ce": 0.003375453467015177, "kl": 0.0008728118018552777, "context": 0.003811859409324825}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.22107486426830292, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 18.31, "ce": 0.03786852068151347, "kl": 0.061335934849921614, "context": 0.0685364871751517}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.24085482954978943, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 18.31, "ce": 0.08507500181440264, "kl": 0.07190882985014468, "context": 0.12102941703051329}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.6412568688392639, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 18.31, "ce": 0.021548072196310386, "kl": 0.8686348678093054, "context": 0.45586551426094957}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.1573086678981781, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 18.31, "ce": 0.03023563709575683, "kl": 0.009415628908755025, "context": 0.03494345012586564}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.2326931655406952, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 18.31, "ce": 0.04089385352563113, "kl": 0.12802209332585335, "context": 0.10490489937365055}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.10524874925613403, "loss_scale": 4.0, "retries": 0, "seconds": 33.1, "max_memory_gb": 18.31, "ce": 0.014312720159068704, "kl": 0.010959225706756115, "context": 0.019792332721408457}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.5247194170951843, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 18.31, "ce": 0.1858377407770604, "kl": 0.157864173699636, "context": 0.2647698314394802}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.7164639234542847, "loss_scale": 4.0, "retries": 0, "seconds": 34.1, "max_memory_gb": 18.31, "ce": 0.15131554120659985, "kl": 0.06751369391258777, "context": 0.1850723881337899}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.05772680416703224, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 18.31, "ce": 0.005335082896635868, "kl": 0.008159080247423844, "context": 0.009414623025804758}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.14815087616443634, "loss_scale": 4.0, "retries": 0, "seconds": 35.1, "max_memory_gb": 18.31, "ce": 0.0052886468401993625, "kl": 0.03838271252607228, "context": 0.024480002666678047}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.7072883248329163, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 18.31, "ce": 0.29753925022669137, "kl": 0.02242511764052324, "context": 0.30875180481234565}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.02482912503182888, "loss_scale": 4.0, "retries": 0, "seconds": 36.1, "max_memory_gb": 18.31, "ce": 0.0013508947158697993, "kl": 0.00032829580231918953, "context": 0.001515042626124341}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.06355972588062286, "loss_scale": 4.0, "retries": 0, "seconds": 36.6, "max_memory_gb": 18.31, "ce": 0.008627928433270426, "kl": 0.015657501382520422, "context": 0.01645667853881605}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.29173770546913147, "loss_scale": 4.0, "retries": 0, "seconds": 37.1, "max_memory_gb": 18.31, "ce": 0.015608827030519024, "kl": 0.40952062740689144, "context": 0.2203691330505535}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.103294849395752, "loss_scale": 4.0, "retries": 0, "seconds": 37.5, "max_memory_gb": 18.31, "ce": 0.35900196746661095, "kl": 0.09040127827938704, "context": 0.40420260293831234}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.2852693498134613, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 18.31, "ce": 0.06570670588280336, "kl": 0.11226970819006965, "context": 0.1218415599596483}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 2.1051464080810547, "loss_scale": 4.0, "retries": 0, "seconds": 38.5, "max_memory_gb": 18.31, "ce": 0.4293172266188776, "kl": 0.9435356995090842, "context": 0.9010850894264877}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 0.9508921504020691, "loss_scale": 4.0, "retries": 0, "seconds": 39.0, "max_memory_gb": 18.31, "ce": 0.017855786536529195, "kl": 1.6894398319127504, "context": 0.862575686420314}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.061677172780036926, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 18.31, "ce": 0.009079406954697333, "kl": 0.01153336325012333, "context": 0.01484608840837609}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.1496879905462265, "loss_scale": 4.0, "retries": 0, "seconds": 40.1, "max_memory_gb": 18.31, "ce": 0.00187073097913526, "kl": 0.09037069733312819, "context": 0.047056080162292346}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.22822611033916473, "loss_scale": 4.0, "retries": 0, "seconds": 40.6, "max_memory_gb": 18.31, "ce": 0.01495709349273966, "kl": 0.02559577004831226, "context": 0.027754977643780876}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.1653987318277359, "loss_scale": 4.0, "retries": 0, "seconds": 41.1, "max_memory_gb": 18.31, "ce": 0.0022222725056053605, "kl": 0.08628513733768273, "context": 0.04536484184427536}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.4409562945365906, "loss_scale": 4.0, "retries": 0, "seconds": 41.6, "max_memory_gb": 18.31, "ce": 0.09439568291418254, "kl": 0.08000158457434736, "context": 0.13439647888299078}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.07800929248332977, "loss_scale": 4.0, "retries": 0, "seconds": 42.1, "max_memory_gb": 18.31, "ce": 0.008717981429072097, "kl": 0.0021005420185247203, "context": 0.009768252552021295}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.5409626960754395, "loss_scale": 4.0, "retries": 0, "seconds": 42.7, "max_memory_gb": 18.31, "ce": 0.15981310090865009, "kl": 0.7408919536464964, "context": 0.5302590896026231}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.37875252962112427, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 18.31, "ce": 0.015677110078286205, "kl": 0.06023865819361163, "context": 0.045796439672812994}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 2.014613628387451, "loss_scale": 4.0, "retries": 0, "seconds": 43.7, "max_memory_gb": 18.31, "ce": 0.03627780273836834, "kl": 0.02656441303943069, "context": 0.04956001019172618}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.3812342882156372, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 18.31, "ce": 0.20767677365802228, "kl": 0.45914862665813416, "context": 0.4372510735411197}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.8873803615570068, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 18.31, "ce": 0.006786598765756935, "kl": 0.6768288677558303, "context": 0.3452010303735733}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.10907445847988129, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 18.31, "ce": 0.011958211309320177, "kl": 0.024966539517322417, "context": 0.024441480980385677}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.11461703479290009, "loss_scale": 4.0, "retries": 0, "seconds": 45.8, "max_memory_gb": 18.31, "ce": 0.017589248155672976, "kl": 0.01573455924517475, "context": 0.025456526906054933}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.08990652859210968, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 18.31, "ce": 0.010102794025442563, "kl": 0.005364817508962005, "context": 0.012785202823579311}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 1.0030909776687622, "loss_scale": 4.0, "retries": 0, "seconds": 46.9, "max_memory_gb": 18.31, "ce": 0.2547361430770252, "kl": 0.4817892905557528, "context": 0.49563080046209507}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.6664037704467773, "loss_scale": 4.0, "retries": 0, "seconds": 47.4, "max_memory_gb": 18.31, "ce": 0.3109652580715192, "kl": 0.4325068355174153, "context": 0.5272186764123035}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.1071345806121826, "loss_scale": 4.0, "retries": 0, "seconds": 48.0, "max_memory_gb": 18.31, "ce": 0.055781604372896254, "kl": 0.4202911021639011, "context": 0.2659271619631909}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.35893976688385, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 18.31, "ce": 0.16453051622374915, "kl": 0.08305572860626853, "context": 0.2060583880811464}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.5679717063903809, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 18.31, "ce": 0.5703151841880754, "kl": 0.02564563532359898, "context": 0.5831379895098507}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 0.895445704460144, "loss_scale": 4.0, "retries": 0, "seconds": 49.5, "max_memory_gb": 18.31, "ce": 0.42565906509844353, "kl": 0.007494138495530933, "context": 0.42940612958045676}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 2.331935405731201, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 18.31, "ce": 0.022300483309663832, "kl": 0.286358773795655, "context": 0.16547987703233957}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.13080203533172607, "loss_scale": 4.0, "retries": 0, "seconds": 50.6, "max_memory_gb": 18.31, "ce": 0.00347155389317777, "kl": 0.09984395488572773, "context": 0.05339353226736421}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.17755316197872162, "loss_scale": 4.0, "retries": 0, "seconds": 51.1, "max_memory_gb": 18.31, "ce": 0.021960176831271383, "kl": 0.011929655222047586, "context": 0.02792500412033405}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.5429468750953674, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 18.31, "ce": 0.04133227617421653, "kl": 0.6180298188373854, "context": 0.3503471881558653}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.30605149269104004, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 18.31, "ce": 0.006698833196423948, "kl": 0.06649263732833788, "context": 0.039945151132997125}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.2510412931442261, "loss_scale": 4.0, "retries": 0, "seconds": 52.7, "max_memory_gb": 18.31, "ce": 0.06549578881822526, "kl": 0.028672949643805623, "context": 0.07983226329088211}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.1410365253686905, "loss_scale": 4.0, "retries": 0, "seconds": 53.2, "max_memory_gb": 18.31, "ce": 0.011974562577961478, "kl": 0.014821669186858344, "context": 0.01938539757975377}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.04679744318127632, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 18.31, "ce": 0.004186011996353045, "kl": 0.0036669072233053157, "context": 0.006019465668941848}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.28803887963294983, "loss_scale": 4.0, "retries": 0, "seconds": 54.2, "max_memory_gb": 18.31, "ce": 0.06890927022323012, "kl": 0.35519483988173306, "context": 0.246506683412008}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.03341598063707352, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 18.31, "ce": 0.004076313154655509, "kl": 0.0020114476828894112, "context": 0.00508203697972931}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.1681859493255615, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 18.31, "ce": 0.4417700678022811, "kl": 0.08634567627649403, "context": 0.4849428880115738}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.48647066950798035, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 18.31, "ce": 0.0343145829574496, "kl": 0.00820200209432187, "context": 0.03841558207932394}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.6846002340316772, "loss_scale": 4.0, "retries": 0, "seconds": 56.2, "max_memory_gb": 18.31, "ce": 0.06137305760057643, "kl": 0.3228507824242115, "context": 0.22279844619333744}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.50822913646698, "loss_scale": 4.0, "retries": 0, "seconds": 56.7, "max_memory_gb": 18.31, "ce": 0.1374713177792728, "kl": 0.28585120174102485, "context": 0.28039692249149084}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.024111775681376457, "loss_scale": 4.0, "retries": 0, "seconds": 57.2, "max_memory_gb": 18.31, "ce": 0.0021621047926601022, "kl": 0.0017083865950553445, "context": 0.003016298040165566}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.10347596555948257, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 18.31, "ce": 0.007018106174655259, "kl": 0.0007215518126031384, "context": 0.0073788820882327855}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.38887935876846313, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 18.31, "ce": 0.012900965521112084, "kl": 0.3454453523736447, "context": 0.18562364811077714}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.7630459070205688, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 18.31, "ce": 0.018704996880842373, "kl": 1.1927293869084679, "context": 0.6150696892291307}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.1101866215467453, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 18.31, "ce": 0.008981350198155269, "kl": 0.045786924281856045, "context": 0.031874812266323715}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 0.9345053434371948, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 18.31, "ce": 0.1414639625363634, "kl": 0.45878546524545527, "context": 0.37085671009117505}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 1.257325530052185, "loss_scale": 4.0, "retries": 0, "seconds": 60.2, "max_memory_gb": 18.32, "ce": 0.05684905192174483, "kl": 0.09393464975437382, "context": 0.10381637676618993}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.4413956105709076, "loss_scale": 4.0, "retries": 0, "seconds": 60.7, "max_memory_gb": 18.32, "ce": 0.17424805247719632, "kl": 0.00947758858819725, "context": 0.17898684306419455}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.14378707110881805, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 18.32, "ce": 0.015232775811455213, "kl": 0.017918615620601486, "context": 0.024192083161324263}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.4914829730987549, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 18.32, "ce": 0.5377920826722402, "kl": 0.19317668967414647, "context": 0.6343804246280342}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.2904178202152252, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 18.32, "ce": 0.053207829125312855, "kl": 0.038793847226770595, "context": 0.07260475229122676}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.33162903785705566, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 18.32, "ce": 0.07311893119185697, "kl": 0.01699711628316436, "context": 0.08161749171995325}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.1969805657863617, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 18.32, "ce": 0.03933385293930769, "kl": 0.021546875475905836, "context": 0.050107293063774705}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.7286057472229004, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 18.32, "ce": 0.6492247777641751, "kl": 0.5645864425350737, "context": 0.9315179915865883}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.28831401467323303, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 18.32, "ce": 0.10890397147159092, "kl": 0.08470876279170625, "context": 0.15125835611252114}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 2.0238516330718994, "loss_scale": 4.0, "retries": 0, "seconds": 65.0, "max_memory_gb": 18.32, "ce": 0.5274408552795649, "kl": 0.14252504706382751, "context": 0.5987033825367689}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.10245393961668015, "loss_scale": 4.0, "retries": 0, "seconds": 65.5, "max_memory_gb": 18.32, "ce": 0.014558817027136683, "kl": 0.0013071886532998178, "context": 0.01521241175942123}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.8297501802444458, "loss_scale": 4.0, "retries": 0, "seconds": 66.0, "max_memory_gb": 18.32, "ce": 0.23247215754236095, "kl": 1.1092689763754606, "context": 0.7871066400548443}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.27146583795547485, "loss_scale": 4.0, "retries": 0, "seconds": 66.5, "max_memory_gb": 18.32, "ce": 0.07202746835537255, "kl": 0.03164801249658922, "context": 0.08785147545859218}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.22427256405353546, "loss_scale": 4.0, "retries": 0, "seconds": 67.0, "max_memory_gb": 18.32, "ce": 0.044771852204576135, "kl": 0.24904753312875982, "context": 0.16929561854340136}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 2.076586961746216, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 18.32, "ce": 0.2482953515718691, "kl": 0.37229548415052705, "context": 0.43444310809718445}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.34264352917671204, "loss_scale": 4.0, "retries": 0, "seconds": 68.1, "max_memory_gb": 18.32, "ce": 0.16285525303101167, "kl": 0.029845034678146476, "context": 0.17777776850562077}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.6427749991416931, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 18.32, "ce": 0.3505528871319257, "kl": 0.18939941423013806, "context": 0.44525258988142014}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.45090335607528687, "loss_scale": 4.0, "retries": 0, "seconds": 69.2, "max_memory_gb": 18.32, "ce": 0.2123092158435611, "kl": 0.0054588370049373225, "context": 0.21503863155521685}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.11884354799985886, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 18.32, "ce": 0.022731609467882663, "kl": 0.09974041191162542, "context": 0.07260181364836171}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 2.12007737159729, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 18.32, "ce": 0.12608389870729297, "kl": 0.1442023723211605, "context": 0.19818507798481733}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.207941472530365, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 18.32, "ce": 0.009792064927751198, "kl": 0.025216998779796995, "context": 0.022400564281269908}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 1.2699213027954102, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 18.32, "ce": 0.07596478005871177, "kl": 0.14272892195731401, "context": 0.14732923917472363}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 1.096739649772644, "loss_scale": 4.0, "retries": 0, "seconds": 71.8, "max_memory_gb": 18.32, "ce": 0.004828946679481305, "kl": 0.26618798770596186, "context": 0.13792294048471376}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.054367389529943466, "loss_scale": 4.0, "retries": 0, "seconds": 72.3, "max_memory_gb": 18.32, "ce": 0.0044682776206173, "kl": 0.002020289517531637, "context": 0.005478422506712377}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.04765349254012108, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 18.32, "ce": 0.006385273736668751, "kl": 0.005924487952142954, "context": 0.00934751785825938}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.5509640574455261, "loss_scale": 4.0, "retries": 0, "seconds": 73.4, "max_memory_gb": 18.32, "ce": 0.04319983260938898, "kl": 0.016914468782488257, "context": 0.05165706668049097}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.706475853919983, "loss_scale": 4.0, "retries": 0, "seconds": 73.9, "max_memory_gb": 18.32, "ce": 0.4345039306936087, "kl": 0.06621464301861124, "context": 0.4676112487068167}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.5445912480354309, "loss_scale": 4.0, "retries": 0, "seconds": 74.4, "max_memory_gb": 18.32, "ce": 0.08396710082888603, "kl": 0.17749332924722694, "context": 0.17271377064753324}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.0486426055431366, "loss_scale": 4.0, "retries": 0, "seconds": 74.9, "max_memory_gb": 18.32, "ce": 0.0020827301923418418, "kl": 0.009160864201930963, "context": 0.006663162392214872}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.1099620908498764, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 18.32, "ce": 0.008902106084860861, "kl": 0.025466759982009535, "context": 0.021635486104059964}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 0.9082925319671631, "loss_scale": 4.0, "retries": 0, "seconds": 76.0, "max_memory_gb": 18.32, "ce": 0.47280303329898743, "kl": 0.5343655392789515, "context": 0.7399858174612746}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.06484975665807724, "loss_scale": 4.0, "retries": 0, "seconds": 76.5, "max_memory_gb": 18.32, "ce": 0.005693223254638724, "kl": 0.012625720228243154, "context": 0.012006083386950195}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.15911516547203064, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 18.32, "ce": 0.014988549461122602, "kl": 0.006503829874418443, "context": 0.018240464210975915}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.2762005031108856, "loss_scale": 4.0, "retries": 0, "seconds": 77.5, "max_memory_gb": 18.32, "ce": 0.07002074451884255, "kl": 0.021550355362705886, "context": 0.0807959190569818}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.22790715098381042, "loss_scale": 4.0, "retries": 0, "seconds": 78.0, "max_memory_gb": 18.32, "ce": 0.01807371100585442, "kl": 0.026572008966468275, "context": 0.031359715496364515}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.46757057309150696, "loss_scale": 4.0, "retries": 0, "seconds": 78.5, "max_memory_gb": 18.32, "ce": 0.007226378686027601, "kl": 0.541598946353588, "context": 0.2780258469720138}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.36913588643074036, "loss_scale": 4.0, "retries": 0, "seconds": 79.0, "max_memory_gb": 18.32, "ce": 0.012728224461170612, "kl": 0.37946306485719106, "context": 0.2024597545605502}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 1.9851493835449219, "loss_scale": 4.0, "retries": 0, "seconds": 79.4, "max_memory_gb": 18.32, "ce": 0.08524354084511288, "kl": 0.08611329994164407, "context": 0.12830018845852464}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.39415082335472107, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 18.32, "ce": 0.1896178545539442, "kl": 0.010229275863821385, "context": 0.19473249623115407}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.8266386985778809, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 18.32, "ce": 0.011402905936847674, "kl": 0.049649627452026834, "context": 0.036227720709575806}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.48895263671875, "loss_scale": 4.0, "retries": 0, "seconds": 80.9, "max_memory_gb": 18.32, "ce": 0.03306111495476216, "kl": 0.03383333087549545, "context": 0.049977780785411596}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.3375145494937897, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 18.32, "ce": 0.18081623836769722, "kl": 0.07093927668756805, "context": 0.21628588330349885}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.6929412484169006, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 18.32, "ce": 0.006077875732444227, "kl": 0.09467799312551506, "context": 0.05341687402687967}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.558924674987793, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 18.32, "ce": 0.030957703362219036, "kl": 0.3437402455601841, "context": 0.2028278261423111}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.3582874834537506, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 18.32, "ce": 0.07374430830532219, "kl": 0.09094552803662737, "context": 0.11921707048895769}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 0.9946955442428589, "loss_scale": 4.0, "retries": 0, "seconds": 83.4, "max_memory_gb": 18.32, "ce": 0.07356584181252401, "kl": 0.13267254547099583, "context": 0.1399021219694987}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.043882369995117, "loss_scale": 4.0, "retries": 0, "seconds": 83.9, "max_memory_gb": 18.32, "ce": 0.22046136611606926, "kl": 0.6851350664510392, "context": 0.5630288964603096}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 4.664178371429443, "loss_scale": 4.0, "retries": 0, "seconds": 84.4, "max_memory_gb": 18.32, "ce": 0.08806918375194073, "kl": 0.12747575213506934, "context": 0.15180706145474687}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.08535894751548767, "loss_scale": 4.0, "retries": 0, "seconds": 85.0, "max_memory_gb": 18.32, "ce": 0.006539188834722154, "kl": 0.02726138438356429, "context": 0.02016988120158203}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.4631235599517822, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 18.32, "ce": 0.41539592354092747, "kl": 0.13450014259433374, "context": 0.4826460105832666}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.15342864394187927, "loss_scale": 4.0, "retries": 0, "seconds": 86.3, "max_memory_gb": 18.32, "ce": 0.02108615287579596, "kl": 0.003318505856441334, "context": 0.022745405323803425}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.12498246878385544, "loss_scale": 4.0, "retries": 0, "seconds": 86.8, "max_memory_gb": 18.32, "ce": 0.02858746850688476, "kl": 0.0021921703591942787, "context": 0.02968355457414873}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.0954875946044922, "loss_scale": 4.0, "retries": 0, "seconds": 87.3, "max_memory_gb": 18.32, "ce": 0.08188906841678545, "kl": 1.406306192278862, "context": 0.7850421518087387}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.03152351826429367, "loss_scale": 4.0, "retries": 0, "seconds": 87.8, "max_memory_gb": 18.32, "ce": 0.0021627194655593485, "kl": 0.0005765459645772353, "context": 0.002450992469675839}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.669542133808136, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 18.32, "ce": 0.023814264928660123, "kl": 0.2440278367139399, "context": 0.1458281825762242}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.13192112743854523, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 18.32, "ce": 0.01541077037109062, "kl": 0.0162338292284403, "context": 0.023527684912551194}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.2009027749300003, "loss_scale": 4.0, "retries": 0, "seconds": 89.5, "max_memory_gb": 18.32, "ce": 0.051115245500113815, "kl": 0.027026507697883062, "context": 0.06462850124808028}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.19806121289730072, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 18.32, "ce": 0.036475951070315205, "kl": 0.03471217744652222, "context": 0.05383204142708564}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 2.5790300369262695, "loss_scale": 4.0, "retries": 0, "seconds": 90.7, "max_memory_gb": 18.32, "ce": 0.2282873704098165, "kl": 0.0871333759278059, "context": 0.2718540634959936}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.06777597218751907, "loss_scale": 4.0, "retries": 0, "seconds": 91.3, "max_memory_gb": 18.32, "ce": 0.010046235911431722, "kl": 0.0027512679880601354, "context": 0.01142186984361615}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.6261573433876038, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 18.32, "ce": 0.021006907903938554, "kl": 0.7123800847548409, "context": 0.3771969493391225}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.8276799917221069, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 18.32, "ce": 0.0650369023787789, "kl": 0.13677083831862547, "context": 0.13342231791466475}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.7660202980041504, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 18.32, "ce": 0.3503832175920252, "kl": 0.024003041777177714, "context": 0.3623847432609182}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.2262033224105835, "loss_scale": 4.0, "retries": 0, "seconds": 93.3, "max_memory_gb": 18.32, "ce": 0.042553066115942784, "kl": 0.027848524870933034, "context": 0.05647732809302397}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.1759469211101532, "loss_scale": 4.0, "retries": 0, "seconds": 93.9, "max_memory_gb": 18.32, "ce": 0.00865196471568197, "kl": 0.08995552315900568, "context": 0.05362972675357014}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 2.0340685844421387, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 18.32, "ce": 0.02276845381129533, "kl": 0.3373914938838425, "context": 0.19146420565084554}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.028469692915678024, "loss_scale": 4.0, "retries": 0, "seconds": 95.0, "max_memory_gb": 18.32, "ce": 0.0028639693391596666, "kl": 0.0007628437815583311, "context": 0.0032453911990160123}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.043963946402072906, "loss_scale": 4.0, "retries": 0, "seconds": 95.5, "max_memory_gb": 18.32, "ce": 0.0029243247699923813, "kl": 0.0042506471945671365, "context": 0.005049648585554678}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.22163943946361542, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 18.32, "ce": 0.009569005895173177, "kl": 0.0911978408730647, "context": 0.0551679263880942}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.892106533050537, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 18.32, "ce": 0.09960941837562132, "kl": 0.25609112537495093, "context": 0.22765497320506256}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.3338058590888977, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 18.32, "ce": 0.08485686097992584, "kl": 0.47198047052370384, "context": 0.32084711140487343}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 0.8469110727310181, "loss_scale": 4.0, "retries": 0, "seconds": 97.8, "max_memory_gb": 18.32, "ce": 0.2536181266805215, "kl": 0.013596369162314659, "context": 0.26041631288899225}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 2.0594820976257324, "loss_scale": 4.0, "retries": 0, "seconds": 98.4, "max_memory_gb": 18.32, "ce": 0.33921142714098096, "kl": 0.34005062258802354, "context": 0.5092367337783799}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.10661167651414871, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 18.32, "ce": 0.004779007955221459, "kl": 0.03870864066993818, "context": 0.024133328115567565}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 3.748889923095703, "loss_scale": 4.0, "retries": 0, "seconds": 99.6, "max_memory_gb": 18.32, "ce": 0.281895003921818, "kl": 0.26668632698965666, "context": 0.41523816736298613}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 0.9218399524688721, "loss_scale": 4.0, "retries": 0, "seconds": 100.2, "max_memory_gb": 18.32, "ce": 0.4501953414583113, "kl": 0.07898770000974764, "context": 0.4896892016986385}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.06635238230228424, "loss_scale": 4.0, "retries": 0, "seconds": 100.7, "max_memory_gb": 18.32, "ce": 0.005328866038325941, "kl": 0.006209034671201152, "context": 0.008433383256488014}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.4182707965373993, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 18.32, "ce": 0.044339890679111704, "kl": 0.03804808350105304, "context": 0.06336393236415461}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.0952487662434578, "loss_scale": 4.0, "retries": 0, "seconds": 101.9, "max_memory_gb": 18.32, "ce": 0.013598955716588534, "kl": 0.0048748231238278095, "context": 0.01603636769868899}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.5404762029647827, "loss_scale": 4.0, "retries": 0, "seconds": 102.4, "max_memory_gb": 18.32, "ce": 0.0012784295322489925, "kl": 0.4532658958917182, "context": 0.22791137734020595}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.0653188228607178, "loss_scale": 4.0, "retries": 0, "seconds": 103.0, "max_memory_gb": 18.32, "ce": 0.007248063120641746, "kl": 0.7256390749753336, "context": 0.3700676026928704}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.31150034070014954, "loss_scale": 4.0, "retries": 0, "seconds": 103.6, "max_memory_gb": 18.32, "ce": 0.05658267832768615, "kl": 0.009387738668920065, "context": 0.061276546330191195}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.5542716383934021, "loss_scale": 4.0, "retries": 0, "seconds": 104.1, "max_memory_gb": 18.32, "ce": 0.14616044331341982, "kl": 0.168922874967393, "context": 0.23062187299365178}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.9514266848564148, "loss_scale": 4.0, "retries": 0, "seconds": 104.7, "max_memory_gb": 18.32, "ce": 0.3038477217196487, "kl": 0.04275200345000485, "context": 0.3252237191190943}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.8363171815872192, "loss_scale": 4.0, "retries": 0, "seconds": 105.3, "max_memory_gb": 18.32, "ce": 0.08419196825707331, "kl": 0.12579877062671585, "context": 0.1470913535449654}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.036321401596069336, "loss_scale": 4.0, "retries": 0, "seconds": 106.0, "max_memory_gb": 18.32, "ce": 0.0017547895604366204, "kl": 0.0034265310342789235, "context": 0.0034680550270422827}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 1.1788201332092285, "loss_scale": 4.0, "retries": 0, "seconds": 106.6, "max_memory_gb": 18.32, "ce": 0.035094179707812145, "kl": 0.06069554144050926, "context": 0.06544194859452546}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.18624043464660645, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 18.32, "ce": 0.037583860917948186, "kl": 0.012563459225930274, "context": 0.043865590705536306}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.37462764978408813, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 18.32, "ce": 0.004491896863328293, "kl": 0.5346619561460102, "context": 0.27182287839241326}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.07716964930295944, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 18.32, "ce": 0.00826396469892643, "kl": 0.0035285149419905792, "context": 0.010028222197433934}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.17314763367176056, "loss_scale": 4.0, "retries": 0, "seconds": 108.9, "max_memory_gb": 18.32, "ce": 0.06774423169554211, "kl": 0.0520109327044338, "context": 0.09374970174394548}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.05145974084734917, "loss_scale": 4.0, "retries": 0, "seconds": 109.5, "max_memory_gb": 18.32, "ce": 0.0051718110698857345, "kl": 0.0016895994704100303, "context": 0.006016610859660432}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.8586048483848572, "loss_scale": 4.0, "retries": 0, "seconds": 110.1, "max_memory_gb": 18.32, "ce": 0.17124988546129316, "kl": 0.11271541669702856, "context": 0.2276075899717398}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 2.151212453842163, "loss_scale": 4.0, "retries": 0, "seconds": 110.7, "max_memory_gb": 18.32, "ce": 0.5012006575489067, "kl": 0.14344467690534657, "context": 0.5729230239521712}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.33279332518577576, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 18.32, "ce": 0.13705339113948867, "kl": 0.11417012475430965, "context": 0.19413844775408506}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 2.7034249305725098, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 18.32, "ce": 0.05815726285800338, "kl": 1.0388916685187723, "context": 0.5776030949200504}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.5176781415939331, "loss_scale": 4.0, "retries": 0, "seconds": 112.5, "max_memory_gb": 18.32, "ce": 0.22414544023922645, "kl": 0.04793263312603813, "context": 0.24811175657669082}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.7593203783035278, "loss_scale": 4.0, "retries": 0, "seconds": 113.1, "max_memory_gb": 18.32, "ce": 0.4173375059835962, "kl": 0.17540782832656987, "context": 0.5050414172874298}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.616382360458374, "loss_scale": 4.0, "retries": 0, "seconds": 113.7, "max_memory_gb": 18.32, "ce": 0.6351635371902375, "kl": 0.02144004523302101, "context": 0.6458835719167837}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.2195647954940796, "loss_scale": 4.0, "retries": 0, "seconds": 114.3, "max_memory_gb": 18.32, "ce": 0.012587543722474948, "kl": 0.11340446515532676, "context": 0.06928977745701559}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.028939571231603622, "loss_scale": 4.0, "retries": 0, "seconds": 114.9, "max_memory_gb": 18.32, "ce": 0.0018326881545362994, "kl": 0.002521322116081137, "context": 0.003093349252594635}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.6299047470092773, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 18.32, "ce": 0.1379614882171154, "kl": 0.30430164001882076, "context": 0.29011230915784836}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.6244353652000427, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 18.32, "ce": 0.008934030323871411, "kl": 0.12486064338008873, "context": 0.07136435052962042}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.2068178802728653, "loss_scale": 4.0, "retries": 0, "seconds": 116.7, "max_memory_gb": 18.32, "ce": 0.05954356287838891, "kl": 0.007496145362893003, "context": 0.06329163374903146}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.3665604293346405, "loss_scale": 4.0, "retries": 0, "seconds": 117.3, "max_memory_gb": 18.32, "ce": 0.006715113850077614, "kl": 0.5280895679934474, "context": 0.27075990397133864}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.24672779440879822, "loss_scale": 4.0, "retries": 0, "seconds": 117.9, "max_memory_gb": 18.32, "ce": 0.029806770908180624, "kl": 0.0815273009189923, "context": 0.07057041925145313}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 1.298209547996521, "loss_scale": 4.0, "retries": 0, "seconds": 118.4, "max_memory_gb": 18.32, "ce": 0.07232539333199384, "kl": 0.006806317574955756, "context": 0.07572855197213357}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.49164706468582153, "loss_scale": 4.0, "retries": 0, "seconds": 118.9, "max_memory_gb": 18.32, "ce": 0.03061221557436511, "kl": 0.20563910141754604, "context": 0.13343176324269734}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.26813822984695435, "loss_scale": 4.0, "retries": 0, "seconds": 119.4, "max_memory_gb": 18.32, "ce": 0.005213821874349378, "kl": 0.21708939332165755, "context": 0.1137585150427185}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.6957666873931885, "loss_scale": 4.0, "retries": 0, "seconds": 119.8, "max_memory_gb": 18.32, "ce": 0.03985205735079944, "kl": 0.03757523198146373, "context": 0.05863967485493049}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.20748257637023926, "loss_scale": 4.0, "retries": 0, "seconds": 120.3, "max_memory_gb": 18.32, "ce": 0.01817698341983487, "kl": 0.006372778468175966, "context": 0.021363373103667982}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.4645305275917053, "loss_scale": 4.0, "retries": 0, "seconds": 120.8, "max_memory_gb": 18.32, "ce": 0.00376829403830925, "kl": 0.3009946657693945, "context": 0.15426562243374065}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.41877979040145874, "loss_scale": 4.0, "retries": 0, "seconds": 121.3, "max_memory_gb": 18.32, "ce": 0.06019767601537751, "kl": 0.01832512508133277, "context": 0.06936023786693113}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.691939115524292, "loss_scale": 4.0, "retries": 0, "seconds": 121.8, "max_memory_gb": 18.32, "ce": 0.5377778253605356, "kl": 0.35850373312132433, "context": 0.7170297135598958}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.3646712899208069, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 18.32, "ce": 0.037260949560732115, "kl": 0.14946302821772406, "context": 0.11199246181058697}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.16631165146827698, "loss_scale": 4.0, "retries": 0, "seconds": 122.7, "max_memory_gb": 18.32, "ce": 0.02645563954138197, "kl": 0.011759177665226161, "context": 0.032335229276213795}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.3173955976963043, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 18.32, "ce": 0.009739138185977936, "kl": 0.10379538463894278, "context": 0.061636831145733595}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.3355146646499634, "loss_scale": 4.0, "retries": 0, "seconds": 123.7, "max_memory_gb": 18.32, "ce": 0.011441822411143221, "kl": 0.055399888660758734, "context": 0.03914176487887744}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.3060599565505981, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 18.32, "ce": 0.3462743010313716, "kl": 0.01850422933603113, "context": 0.3555264195165364}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.08857408165931702, "loss_scale": 4.0, "retries": 0, "seconds": 124.7, "max_memory_gb": 18.32, "ce": 0.009962266783986706, "kl": 0.018094234179443447, "context": 0.019009384093806148}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.382929265499115, "loss_scale": 4.0, "retries": 0, "seconds": 125.1, "max_memory_gb": 18.32, "ce": 0.07211616149288602, "kl": 0.024178749628845253, "context": 0.08420553717587609}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.42208045721054077, "loss_scale": 4.0, "retries": 0, "seconds": 125.6, "max_memory_gb": 18.32, "ce": 0.03572762788098771, "kl": 0.009967416915969807, "context": 0.040711335284868255}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.3444626033306122, "loss_scale": 4.0, "retries": 0, "seconds": 126.1, "max_memory_gb": 18.32, "ce": 0.1081127087236382, "kl": 0.026000415411544964, "context": 0.12111291458131745}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.3190174698829651, "loss_scale": 4.0, "retries": 0, "seconds": 126.6, "max_memory_gb": 18.32, "ce": 0.09575906774261966, "kl": 0.04131793564738473, "context": 0.11641803581733257}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 0.9160267114639282, "loss_scale": 4.0, "retries": 0, "seconds": 127.1, "max_memory_gb": 18.32, "ce": 0.10300391793862218, "kl": 0.022326273101498373, "context": 0.11416705274314154}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.12464330345392227, "loss_scale": 4.0, "retries": 0, "seconds": 127.6, "max_memory_gb": 18.32, "ce": 0.005780153100204188, "kl": 0.060918145813047886, "context": 0.03623922460246831}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.06001122295856476, "loss_scale": 4.0, "retries": 0, "seconds": 128.1, "max_memory_gb": 18.32, "ce": 0.005218150559812784, "kl": 0.00478324037976563, "context": 0.007609770400449634}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.6015925407409668, "loss_scale": 4.0, "retries": 0, "seconds": 128.6, "max_memory_gb": 18.32, "ce": 0.04855867731384933, "kl": 0.19098267553818005, "context": 0.1440500175303896}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 2.558736801147461, "loss_scale": 4.0, "retries": 0, "seconds": 129.1, "max_memory_gb": 18.32, "ce": 0.4200516671407968, "kl": 0.049737941357307136, "context": 0.44492062949575484}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.019028982147574425, "loss_scale": 4.0, "retries": 0, "seconds": 129.6, "max_memory_gb": 18.32, "ce": 0.0012473551396396942, "kl": 0.000738473012461327, "context": 0.0016165915876626968}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.38743194937705994, "loss_scale": 4.0, "retries": 0, "seconds": 130.1, "max_memory_gb": 18.32, "ce": 0.006482511409558356, "kl": 0.4753480465733446, "context": 0.24415653699543327}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.20653633773326874, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 18.32, "ce": 0.030540917068719864, "kl": 0.011063692392781377, "context": 0.03607276326511055}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.3610062301158905, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 18.32, "ce": 0.0423017586581409, "kl": 0.049997084570350125, "context": 0.06730029871687293}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.047395769506692886, "loss_scale": 4.0, "retries": 0, "seconds": 131.6, "max_memory_gb": 18.32, "ce": 0.0061965717759449035, "kl": 0.00298009116522735, "context": 0.007686617551371455}
|
runs/single_internvl/context/L11_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L11_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.11.self_attn.q_proj",
|
| 35 |
+
"model.language_model.layers.11.self_attn.o_proj",
|
| 36 |
+
"model.language_model.layers.11.self_attn.k_proj",
|
| 37 |
+
"model.language_model.layers.11.mlp.gate_proj",
|
| 38 |
+
"model.language_model.layers.11.mlp.up_proj",
|
| 39 |
+
"model.language_model.layers.11.mlp.down_proj",
|
| 40 |
+
"model.language_model.layers.11.self_attn.v_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L11_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 11,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L11_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.9291117787361145, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 18.13, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 1.2146095037460327, "loss_scale": 4.0, "retries": 0, "seconds": 1.9, "max_memory_gb": 18.14, "ce": 0.6385369354175054, "kl": 0.46720077286204287, "context": 0.872137291567924}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 3.2859768867492676, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 18.14, "ce": 1.0982384979724884, "kl": 0.06742501817643642, "context": 1.1319509968161583}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.7314800024032593, "loss_scale": 4.0, "retries": 0, "seconds": 2.9, "max_memory_gb": 18.14, "ce": 0.4504057520243805, "kl": 0.4305011790897879, "context": 0.6656563331562211}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.04778214544057846, "loss_scale": 4.0, "retries": 0, "seconds": 3.4, "max_memory_gb": 18.14, "ce": 0.0093093293489801, "kl": 0.008332225967023987, "context": 0.013475442319759168}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.9157899618148804, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 18.14, "ce": 0.28440476044488605, "kl": 0.07494310860420228, "context": 0.3218763086333638}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 2.833906412124634, "loss_scale": 4.0, "retries": 0, "seconds": 4.4, "max_memory_gb": 18.14, "ce": 0.7944834506288316, "kl": 0.12474629241989987, "context": 0.8568565983514418}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 2.822571039199829, "loss_scale": 4.0, "retries": 0, "seconds": 4.9, "max_memory_gb": 18.15, "ce": 0.7360810004174709, "kl": 0.2855896190740168, "context": 0.8788758236914873}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 2.015159845352173, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 18.15, "ce": 0.473635358735919, "kl": 0.13433880507363938, "context": 0.5408047698438168}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.361081063747406, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 18.15, "ce": 0.04405079077696428, "kl": 0.04561266698874533, "context": 0.06685712636681274}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 2.348710536956787, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 18.15, "ce": 0.5854048133624019, "kl": 0.004641342862669262, "context": 0.5877254874285427}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.7136592864990234, "loss_scale": 4.0, "retries": 0, "seconds": 6.9, "max_memory_gb": 18.15, "ce": 0.3754042784912599, "kl": 0.003411983462228818, "context": 0.37711027907607786}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.2995357513427734, "loss_scale": 4.0, "retries": 0, "seconds": 7.4, "max_memory_gb": 18.15, "ce": 0.3426686481689103, "kl": 0.01761364178128133, "context": 0.35147545517247636}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 2.693561315536499, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 18.15, "ce": 0.5339736342430115, "kl": 0.936545180156827, "context": 1.002246230840683}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.100045919418335, "loss_scale": 4.0, "retries": 0, "seconds": 8.3, "max_memory_gb": 18.15, "ce": 0.29800998223231545, "kl": 0.4435098171679783, "context": 0.5197649094427561}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.7714201211929321, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 18.15, "ce": 0.11922081660304684, "kl": 0.0014665123312624928, "context": 0.1199540735615301}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.9302612543106079, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 18.15, "ce": 0.2026217080419883, "kl": 0.8302952987141907, "context": 0.6177693508798257}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.6706310510635376, "loss_scale": 4.0, "retries": 0, "seconds": 9.8, "max_memory_gb": 18.15, "ce": 0.16845112724695355, "kl": 0.21751208881323691, "context": 0.2772071774234064}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 1.8701218366622925, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 18.15, "ce": 0.1619601845741272, "kl": 0.18363978248089552, "context": 0.2537800818681717}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.06947007775306702, "loss_scale": 4.0, "retries": 0, "seconds": 10.8, "max_memory_gb": 18.15, "ce": 0.0028635507915169, "kl": 0.02024099533446133, "context": 0.012984048546059057}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 0.4905403256416321, "loss_scale": 4.0, "retries": 0, "seconds": 11.3, "max_memory_gb": 18.15, "ce": 0.02794297413129243, "kl": 0.019889475164745818, "context": 0.037887711714574834}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.5875646471977234, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 18.15, "ce": 0.06109267122519668, "kl": 0.024305802968228818, "context": 0.07324557026731782}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.360948324203491, "loss_scale": 4.0, "retries": 0, "seconds": 12.3, "max_memory_gb": 18.15, "ce": 0.23952790781004296, "kl": 0.9458142951457376, "context": 0.7124350688873164}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 0.5826642513275146, "loss_scale": 4.0, "retries": 0, "seconds": 12.8, "max_memory_gb": 18.15, "ce": 0.03765900246798992, "kl": 0.09360127756372094, "context": 0.0844596407841891}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 0.88252854347229, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 18.15, "ce": 0.3231272671473562, "kl": 0.14252867689356208, "context": 0.394391605630517}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 0.29429858922958374, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 18.15, "ce": 0.028894161339849234, "kl": 0.6534270849078894, "context": 0.3556076902896166}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 0.29467660188674927, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 18.15, "ce": 0.08123791154503124, "kl": 0.330468225374716, "context": 0.24647202234336874}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.8321999311447144, "loss_scale": 4.0, "retries": 0, "seconds": 14.8, "max_memory_gb": 18.15, "ce": 0.4775619732681662, "kl": 0.9533117769751698, "context": 0.954217866063118}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.49203595519065857, "loss_scale": 4.0, "retries": 0, "seconds": 15.3, "max_memory_gb": 18.15, "ce": 0.04384113394189626, "kl": 0.3799553078133613, "context": 0.23381878877989948}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.324317991733551, "loss_scale": 4.0, "retries": 0, "seconds": 15.8, "max_memory_gb": 18.15, "ce": 0.027416471697506495, "kl": 0.001247662873538502, "context": 0.02804030220977438}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.6024863123893738, "loss_scale": 4.0, "retries": 0, "seconds": 16.3, "max_memory_gb": 18.15, "ce": 0.18535410449840128, "kl": 0.09620195958996192, "context": 0.23345509357750416}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.2149583101272583, "loss_scale": 4.0, "retries": 0, "seconds": 16.8, "max_memory_gb": 18.15, "ce": 0.013136480120010674, "kl": 0.009218828759912867, "context": 0.017745894379913807}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.2947569489479065, "loss_scale": 4.0, "retries": 0, "seconds": 17.3, "max_memory_gb": 18.15, "ce": 0.04429748607799411, "kl": 0.020478466991335154, "context": 0.054536718875169754}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.5242428779602051, "loss_scale": 4.0, "retries": 0, "seconds": 17.8, "max_memory_gb": 18.15, "ce": 0.005159375308721792, "kl": 0.7852372446882327, "context": 0.39777801007949165}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.06784414499998093, "loss_scale": 4.0, "retries": 0, "seconds": 18.3, "max_memory_gb": 18.15, "ce": 0.0043062665208708495, "kl": 0.001698021695119678, "context": 0.005155277482117526}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.2684260606765747, "loss_scale": 4.0, "retries": 0, "seconds": 18.8, "max_memory_gb": 18.15, "ce": 0.08695755233929958, "kl": 0.052084985247347504, "context": 0.11300004489021376}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.4864974617958069, "loss_scale": 4.0, "retries": 0, "seconds": 19.3, "max_memory_gb": 18.15, "ce": 0.13671289055491798, "kl": 0.779988172929734, "context": 0.5267069581896067}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.30756163597106934, "loss_scale": 4.0, "retries": 0, "seconds": 19.7, "max_memory_gb": 18.15, "ce": 0.020800873870030046, "kl": 0.4772206577472389, "context": 0.2594112087972462}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.48471468687057495, "loss_scale": 4.0, "retries": 0, "seconds": 20.2, "max_memory_gb": 18.15, "ce": 0.1070868434762815, "kl": 0.010766851526568644, "context": 0.1124702668748796}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.11484257131814957, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 18.15, "ce": 0.008355108293471858, "kl": 0.008493409113725647, "context": 0.012601812661159784}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.4829702377319336, "loss_scale": 4.0, "retries": 0, "seconds": 21.2, "max_memory_gb": 18.15, "ce": 0.385015873296652, "kl": 0.12365006655454636, "context": 0.44684090465307236}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.021626770496368408, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 18.15, "ce": 0.003242572522140108, "kl": 0.00036962772628612584, "context": 0.0034273864584974945}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.03350977972149849, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 18.15, "ce": 0.0020211758310324512, "kl": 0.0022315212354442338, "context": 0.003136936473310925}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 3.443107843399048, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 18.15, "ce": 0.5276720335605205, "kl": 1.3351157978031551, "context": 1.1952299510594457}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.4581031799316406, "loss_scale": 4.0, "retries": 0, "seconds": 23.2, "max_memory_gb": 18.15, "ce": 0.8125936535361689, "kl": 0.24667916005273582, "context": 0.9359332337917294}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.5341800451278687, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 18.15, "ce": 0.08185041529941373, "kl": 0.2538272278616205, "context": 0.2087640311801806}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.06719125807285309, "loss_scale": 4.0, "retries": 0, "seconds": 24.4, "max_memory_gb": 18.15, "ce": 0.005469122686918126, "kl": 0.01582615589722991, "context": 0.013382200675550848}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.05683750659227371, "loss_scale": 4.0, "retries": 0, "seconds": 24.9, "max_memory_gb": 18.15, "ce": 0.0023746024089632556, "kl": 0.0033435368050049874, "context": 0.004046370799187571}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.33464810252189636, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 18.15, "ce": 0.013000300896237604, "kl": 0.3063874185476152, "context": 0.16619400266063167}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.4262702763080597, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 18.15, "ce": 0.08958555921708466, "kl": 0.008024155657039955, "context": 0.09359763995598769}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.7017297744750977, "loss_scale": 4.0, "retries": 0, "seconds": 26.4, "max_memory_gb": 18.15, "ce": 0.5240702132905426, "kl": 0.020803862390152972, "context": 0.5344721331930486}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.65752774477005, "loss_scale": 4.0, "retries": 0, "seconds": 26.9, "max_memory_gb": 18.15, "ce": 0.020795864576939493, "kl": 0.22530695877503604, "context": 0.13344934210181236}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.5713489651679993, "loss_scale": 4.0, "retries": 0, "seconds": 27.4, "max_memory_gb": 18.15, "ce": 0.14422618283424526, "kl": 0.4011994069442153, "context": 0.3448258969001472}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5661740303039551, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 18.15, "ce": 0.0381371616240358, "kl": 0.32378645128483186, "context": 0.20003039049333893}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.1165347546339035, "loss_scale": 4.0, "retries": 0, "seconds": 28.3, "max_memory_gb": 18.15, "ce": 0.017925121792359278, "kl": 0.013536558214468641, "context": 0.024693400191608816}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.24399791657924652, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 18.15, "ce": 0.021400877623818815, "kl": 0.25242811581119895, "context": 0.14761492959223688}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.08965124189853668, "loss_scale": 4.0, "retries": 0, "seconds": 29.3, "max_memory_gb": 18.15, "ce": 0.009936314447259065, "kl": 0.020183196211291943, "context": 0.02002791162522044}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.05846288055181503, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 18.15, "ce": 0.0022479108465631725, "kl": 0.0010213859621899246, "context": 0.0027586038049776107}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.23902785778045654, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 18.15, "ce": 0.03808850413770415, "kl": 0.06277236915775575, "context": 0.06947468873113394}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.2298155277967453, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 18.15, "ce": 0.0831354790716432, "kl": 0.06896891753422096, "context": 0.11761993798427284}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.6572760343551636, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 18.15, "ce": 0.02319277881179005, "kl": 0.8475354436795897, "context": 0.44696049464982934}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.311914324760437, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 18.15, "ce": 0.029605419942527078, "kl": 0.009449814933759626, "context": 0.034330328286159784}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.2377753108739853, "loss_scale": 4.0, "retries": 0, "seconds": 32.3, "max_memory_gb": 18.15, "ce": 0.039536086027510464, "kl": 0.1262578350433614, "context": 0.10266500432044268}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.09804918617010117, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 18.15, "ce": 0.013999658462125808, "kl": 0.011660254938760772, "context": 0.01982978609157726}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.6114555597305298, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 18.15, "ce": 0.18790228059515357, "kl": 0.1603372874087654, "context": 0.2680709243286401}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.6632749438285828, "loss_scale": 4.0, "retries": 0, "seconds": 33.8, "max_memory_gb": 18.15, "ce": 0.15592659945207288, "kl": 0.06541057590266064, "context": 0.18863188364912276}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.056615717709064484, "loss_scale": 4.0, "retries": 0, "seconds": 34.3, "max_memory_gb": 18.15, "ce": 0.0051587143243523315, "kl": 0.008212809192627901, "context": 0.009265118889743462}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.17257235944271088, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 18.15, "ce": 0.005041736258135643, "kl": 0.035829770255077165, "context": 0.02295662112373975}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.7452378273010254, "loss_scale": 4.0, "retries": 0, "seconds": 35.3, "max_memory_gb": 18.15, "ce": 0.29562281811377034, "kl": 0.02361737284809351, "context": 0.30743150983471423}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.01745838113129139, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 18.15, "ce": 0.0011864941770909354, "kl": 0.0002568953150330344, "context": 0.0013149418373359367}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.06813450902700424, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 18.15, "ce": 0.009014494378789095, "kl": 0.01571958819113206, "context": 0.01687428903824184}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.29901936650276184, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 18.15, "ce": 0.016919521964155138, "kl": 0.398119637160562, "context": 0.21597933681914583}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.3325587511062622, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 18.15, "ce": 0.3510820066949236, "kl": 0.09526643438039173, "context": 0.39871521829627454}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.3305263817310333, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 18.15, "ce": 0.0654347650715863, "kl": 0.11083396457070194, "context": 0.12085174732055748}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 3.949610948562622, "loss_scale": 4.0, "retries": 0, "seconds": 38.2, "max_memory_gb": 18.15, "ce": 0.4296244882280007, "kl": 0.9210941893979907, "context": 0.8901715772226453}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 0.9806334972381592, "loss_scale": 4.0, "retries": 0, "seconds": 38.7, "max_memory_gb": 18.15, "ce": 0.017849509204097558, "kl": 1.692798393254634, "context": 0.8642487394390628}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.06665092706680298, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 18.15, "ce": 0.010502472199732438, "kl": 0.01202208995619003, "context": 0.01651351701002568}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.1444871723651886, "loss_scale": 4.0, "retries": 0, "seconds": 39.7, "max_memory_gb": 18.15, "ce": 0.0019854423735523596, "kl": 0.08611758548795478, "context": 0.045044236598187126}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.2625989615917206, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 18.15, "ce": 0.013929876256952411, "kl": 0.022563430089576286, "context": 0.02521159130265005}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.17506003379821777, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 18.15, "ce": 0.00242383754812181, "kl": 0.08551266308404593, "context": 0.045180167457147036}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.4046684503555298, "loss_scale": 4.0, "retries": 0, "seconds": 41.2, "max_memory_gb": 18.15, "ce": 0.09080989542417228, "kl": 0.0695925914333202, "context": 0.1256061962340027}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.08161745220422745, "loss_scale": 4.0, "retries": 0, "seconds": 41.7, "max_memory_gb": 18.15, "ce": 0.008477979194140062, "kl": 0.0025808995123952627, "context": 0.009768428979441524}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.5586838722229004, "loss_scale": 4.0, "retries": 0, "seconds": 42.2, "max_memory_gb": 18.15, "ce": 0.1554780245933216, "kl": 0.7350764669463388, "context": 0.5230162503721658}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.3561215400695801, "loss_scale": 4.0, "retries": 0, "seconds": 42.8, "max_memory_gb": 18.15, "ce": 0.017690274498818326, "kl": 0.05537455174453498, "context": 0.04537754976627184}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 1.172987699508667, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 18.15, "ce": 0.03344687441631322, "kl": 0.023527804153250287, "context": 0.045210775564100913}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.8033097982406616, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 18.15, "ce": 0.18197088374290615, "kl": 0.42709056031890213, "context": 0.3955161648336798}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.8820270299911499, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 18.15, "ce": 0.007240172766614705, "kl": 0.6601191888330504, "context": 0.33729976369068027}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.11193723976612091, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 18.15, "ce": 0.012850095254179905, "kl": 0.02313915708532477, "context": 0.024419673323791358}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.1412448137998581, "loss_scale": 4.0, "retries": 0, "seconds": 45.3, "max_memory_gb": 18.15, "ce": 0.018527779569467384, "kl": 0.016373045604268555, "context": 0.02671430190639512}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.09433746337890625, "loss_scale": 4.0, "retries": 0, "seconds": 45.8, "max_memory_gb": 18.15, "ce": 0.009960487484931946, "kl": 0.005210027622524649, "context": 0.012565501208882779}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 1.252737045288086, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 18.15, "ce": 0.27399093215353787, "kl": 0.44835442002658965, "context": 0.4981681389035657}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 2.0004143714904785, "loss_scale": 4.0, "retries": 0, "seconds": 46.8, "max_memory_gb": 18.15, "ce": 0.28440687876354787, "kl": 0.42301125864287314, "context": 0.495912508420588}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.710572600364685, "loss_scale": 4.0, "retries": 0, "seconds": 47.3, "max_memory_gb": 18.15, "ce": 0.05860068608308211, "kl": 0.42423395872174297, "context": 0.2707176732365042}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.5209276676177979, "loss_scale": 4.0, "retries": 0, "seconds": 47.8, "max_memory_gb": 18.15, "ce": 0.12770467216614634, "kl": 0.061761005548760295, "context": 0.15858517307788134}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.6523054838180542, "loss_scale": 4.0, "retries": 0, "seconds": 48.3, "max_memory_gb": 18.15, "ce": 0.5695412301283795, "kl": 0.025873608188703656, "context": 0.5824780543334782}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.0071903467178345, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 18.15, "ce": 0.4149770545773208, "kl": 0.007090575378242647, "context": 0.4185223360545933}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 2.480346441268921, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 18.15, "ce": 0.021384341380326077, "kl": 0.19468086474807933, "context": 0.118724774569273}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.14298419654369354, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 18.15, "ce": 0.0036883839711663313, "kl": 0.10199777010711841, "context": 0.05468726947583491}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.17731617391109467, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 18.15, "ce": 0.018784405532642268, "kl": 0.010044523733085953, "context": 0.02380666652607033}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.6063717603683472, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 18.15, "ce": 0.0412164260778809, "kl": 0.5852357146068243, "context": 0.33383429373498075}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.34269091486930847, "loss_scale": 4.0, "retries": 0, "seconds": 51.4, "max_memory_gb": 18.15, "ce": 0.005318293129676022, "kl": 0.0685440749512054, "context": 0.039590332424268126}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.2769660949707031, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 18.15, "ce": 0.06967411190271378, "kl": 0.02986148581840098, "context": 0.08460485562682152}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.15261976420879364, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 18.15, "ce": 0.011053122929297388, "kl": 0.01223748371558031, "context": 0.01717186473251786}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.0587712898850441, "loss_scale": 4.0, "retries": 0, "seconds": 53.0, "max_memory_gb": 18.15, "ce": 0.005153313293703832, "kl": 0.0044940208390471525, "context": 0.007400323826004751}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.32417377829551697, "loss_scale": 4.0, "retries": 0, "seconds": 53.5, "max_memory_gb": 18.15, "ce": 0.058768728049471974, "kl": 0.36151379137299955, "context": 0.23952562594786286}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.037826742976903915, "loss_scale": 4.0, "retries": 0, "seconds": 54.0, "max_memory_gb": 18.15, "ce": 0.004639766528271139, "kl": 0.00227292776980903, "context": 0.005776230507763103}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.572845458984375, "loss_scale": 4.0, "retries": 0, "seconds": 54.6, "max_memory_gb": 18.15, "ce": 0.3671512344662915, "kl": 0.09287650836887451, "context": 0.41358947593107587}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.9179390668869019, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 18.15, "ce": 0.05531371074175695, "kl": 0.014164375390009809, "context": 0.0623958965879865}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.753691554069519, "loss_scale": 4.0, "retries": 0, "seconds": 55.6, "max_memory_gb": 18.15, "ce": 0.061365943402051926, "kl": 0.3247556295245886, "context": 0.22374376002699137}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.509399652481079, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 18.15, "ce": 0.10851941769942641, "kl": 0.25708426523488015, "context": 0.23706154758110642}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.026777809485793114, "loss_scale": 4.0, "retries": 0, "seconds": 56.6, "max_memory_gb": 18.15, "ce": 0.002313477758434601, "kl": 0.0018921472774309223, "context": 0.003259551405790262}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.14706629514694214, "loss_scale": 4.0, "retries": 0, "seconds": 57.1, "max_memory_gb": 18.15, "ce": 0.00915152719244361, "kl": 0.001110734709072858, "context": 0.00970689469249919}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.35810816287994385, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 18.15, "ce": 0.01307546824682504, "kl": 0.3201808414887637, "context": 0.17316588945686817}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.7111715078353882, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 18.15, "ce": 0.017358009878080338, "kl": 1.128528239321895, "context": 0.5816221159766428}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.11062951385974884, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 18.15, "ce": 0.008623579167760909, "kl": 0.044445188366808, "context": 0.030846173263853416}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 3.0539543628692627, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 18.15, "ce": 0.11249925999436527, "kl": 0.4396751314643552, "context": 0.3323368331184611}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 1.359500527381897, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 18.16, "ce": 0.04754152748500928, "kl": 0.07621566977468319, "context": 0.08564936608308926}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.5457756519317627, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 18.16, "ce": 0.18849364257766865, "kl": 0.008232342828705441, "context": 0.19260980852413923}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.15300253033638, "loss_scale": 4.0, "retries": 0, "seconds": 60.8, "max_memory_gb": 18.16, "ce": 0.014027058728970587, "kl": 0.01912536602731052, "context": 0.02358974233357003}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.9748860597610474, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 18.16, "ce": 0.5046571904676966, "kl": 0.1957237555179745, "context": 0.6025190713698976}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.3285699784755707, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 18.16, "ce": 0.053181276307441294, "kl": 0.03528011208982207, "context": 0.07082133329095086}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.47480788826942444, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 18.16, "ce": 0.08675171832146589, "kl": 0.01777930565731367, "context": 0.09564137272536755}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.1821429431438446, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 18.16, "ce": 0.03881792118772864, "kl": 0.02050077880267054, "context": 0.04906830983236432}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.9378947019577026, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 18.16, "ce": 0.6690018452645745, "kl": 0.5693810492412013, "context": 0.9536923978012055}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.35464945435523987, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 18.16, "ce": 0.11567087130970322, "kl": 0.08054358960362151, "context": 0.15594266608241014}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.8551528453826904, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 18.16, "ce": 0.5670040715485811, "kl": 0.1377873681485653, "context": 0.6358977425843477}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.11988420784473419, "loss_scale": 4.0, "retries": 0, "seconds": 64.9, "max_memory_gb": 18.16, "ce": 0.016605394921498373, "kl": 0.001317637761530932, "context": 0.017264214053284377}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 2.193629026412964, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 18.16, "ce": 0.19219108087418135, "kl": 1.1111277856398374, "context": 0.7477549706818536}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.2647772431373596, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 18.16, "ce": 0.0731520865811035, "kl": 0.031096065147721674, "context": 0.08870011731050909}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.192884624004364, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 18.16, "ce": 0.044592786696739495, "kl": 0.2447200210444862, "context": 0.16695280553540215}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 2.526207208633423, "loss_scale": 4.0, "retries": 0, "seconds": 67.0, "max_memory_gb": 18.16, "ce": 0.30523367496789433, "kl": 0.3977161234943196, "context": 0.504091736394912}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.34303018450737, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 18.16, "ce": 0.16779401829990093, "kl": 0.03080620406581147, "context": 0.18319711716321763}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.6927597522735596, "loss_scale": 4.0, "retries": 0, "seconds": 68.0, "max_memory_gb": 18.16, "ce": 0.3564182014670223, "kl": 0.18897814163938165, "context": 0.4509072727523744}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.531990110874176, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 18.16, "ce": 0.18832053714504582, "kl": 0.004820739694395115, "context": 0.19073090186429908}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.11950720846652985, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 18.16, "ce": 0.023128781176637858, "kl": 0.09778304179781117, "context": 0.07202030124608427}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 1.6024932861328125, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 18.16, "ce": 0.07936490944121033, "kl": 0.09866528268321417, "context": 0.12869755108840764}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.36228346824645996, "loss_scale": 4.0, "retries": 0, "seconds": 70.1, "max_memory_gb": 18.16, "ce": 0.011205258662812412, "kl": 0.031123968459723983, "context": 0.026767242423375137}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 1.187363862991333, "loss_scale": 4.0, "retries": 0, "seconds": 70.6, "max_memory_gb": 18.16, "ce": 0.07308550993911922, "kl": 0.1314606498926878, "context": 0.1388158341869712}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 1.2456223964691162, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 18.16, "ce": 0.005093655534437858, "kl": 0.19726368515375725, "context": 0.10372549413295928}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.05562524497509003, "loss_scale": 4.0, "retries": 0, "seconds": 71.7, "max_memory_gb": 18.16, "ce": 0.004926590685499832, "kl": 0.0020456954516703263, "context": 0.005949438374955207}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.05634591728448868, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 18.16, "ce": 0.007282567326910794, "kl": 0.0065080213971668854, "context": 0.010536578192841262}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.5550774335861206, "loss_scale": 4.0, "retries": 0, "seconds": 72.7, "max_memory_gb": 18.16, "ce": 0.0407154590357095, "kl": 0.01619102736003697, "context": 0.04881097318138927}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 2.6214711666107178, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 18.16, "ce": 0.4350802690969431, "kl": 0.0699265446382924, "context": 0.4700435439881403}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.5755144953727722, "loss_scale": 4.0, "retries": 0, "seconds": 73.8, "max_memory_gb": 18.16, "ce": 0.06141856103204191, "kl": 0.19674048433080316, "context": 0.15978880785405636}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.04520151764154434, "loss_scale": 4.0, "retries": 0, "seconds": 74.3, "max_memory_gb": 18.16, "ce": 0.00209605970303528, "kl": 0.008727176412776316, "context": 0.0064596478041494265}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.10646949708461761, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 18.16, "ce": 0.00923600874375552, "kl": 0.0217980863271805, "context": 0.020135051454417408}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.0268045663833618, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 18.16, "ce": 0.4889645984439994, "kl": 0.5199243706592824, "context": 0.7489267792843748}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.10247103869915009, "loss_scale": 4.0, "retries": 0, "seconds": 75.8, "max_memory_gb": 18.16, "ce": 0.006134760156783159, "kl": 0.011507052535307594, "context": 0.011888286542671267}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.15284842252731323, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 18.16, "ce": 0.011864277592394501, "kl": 0.004658267709601205, "context": 0.014193410868756473}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.2614138424396515, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 18.16, "ce": 0.06194893328938633, "kl": 0.024311499029863626, "context": 0.07410468289162964}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.2876218557357788, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 18.16, "ce": 0.02105469828529749, "kl": 0.028783877311070682, "context": 0.035446635069092736}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.48000994324684143, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 18.16, "ce": 0.007566517670056783, "kl": 0.5246122007147278, "context": 0.2698726211092435}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.3766905963420868, "loss_scale": 4.0, "retries": 0, "seconds": 78.6, "max_memory_gb": 18.16, "ce": 0.01396619663137244, "kl": 0.3703451588480675, "context": 0.19913878164516063}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 0.6332608461380005, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 18.16, "ce": 0.06149989413097501, "kl": 0.08032134745735675, "context": 0.10166057193418965}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.37203967571258545, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 18.16, "ce": 0.1837999059644062, "kl": 0.009653122410782089, "context": 0.1886264675667917}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.7445963025093079, "loss_scale": 4.0, "retries": 0, "seconds": 80.5, "max_memory_gb": 18.16, "ce": 0.012341526682575932, "kl": 0.03888283690639582, "context": 0.03178294519602787}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.3904459476470947, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 18.16, "ce": 0.02582500409334898, "kl": 0.028552964940899983, "context": 0.04010148660745472}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.43466249108314514, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 18.16, "ce": 0.19226834762957878, "kl": 0.06561337105813436, "context": 0.22507502953521907}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.4691566526889801, "loss_scale": 4.0, "retries": 0, "seconds": 82.2, "max_memory_gb": 18.16, "ce": 0.006627468857914209, "kl": 0.05176516744541004, "context": 0.03251005246420391}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.6593745350837708, "loss_scale": 4.0, "retries": 0, "seconds": 82.9, "max_memory_gb": 18.16, "ce": 0.03675386687973514, "kl": 0.33973436518135713, "context": 0.2066210462944582}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.45579877495765686, "loss_scale": 4.0, "retries": 0, "seconds": 83.5, "max_memory_gb": 18.16, "ce": 0.07794625172391534, "kl": 0.09538840128516313, "context": 0.1256404486775864}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.016558289527893, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 18.16, "ce": 0.06463748517853674, "kl": 0.11178696555725764, "context": 0.1205309679498896}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.2336862087249756, "loss_scale": 4.0, "retries": 0, "seconds": 84.8, "max_memory_gb": 18.16, "ce": 0.2347833207459189, "kl": 0.754215543391183, "context": 0.6118910773075186}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 3.158400297164917, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 18.16, "ce": 0.0803072792768944, "kl": 0.08795851663307985, "context": 0.1242865391104715}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.09860288351774216, "loss_scale": 4.0, "retries": 0, "seconds": 85.9, "max_memory_gb": 18.16, "ce": 0.006746628729160875, "kl": 0.026913940677331993, "context": 0.020203599124215543}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.7399766445159912, "loss_scale": 4.0, "retries": 0, "seconds": 86.5, "max_memory_gb": 18.16, "ce": 0.4488376588560641, "kl": 0.1265149204991758, "context": 0.5120951246935874}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.165523961186409, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 18.16, "ce": 0.022568637039512396, "kl": 0.0033477775141363963, "context": 0.02424252510536462}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.12527862191200256, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 18.16, "ce": 0.030177966327755712, "kl": 0.0018693027668632567, "context": 0.031112618424231187}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.136484980583191, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 18.16, "ce": 0.08755737525643781, "kl": 1.3892329186201096, "context": 0.7821738347411156}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.06107250228524208, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 18.16, "ce": 0.0030922925361664966, "kl": 0.0008444003215117846, "context": 0.0035144927678629756}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.6445450186729431, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 18.16, "ce": 0.024922322660131613, "kl": 0.24403326166793704, "context": 0.14693895558593795}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.120402030646801, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 18.16, "ce": 0.015395148424431682, "kl": 0.015193125087535009, "context": 0.022991711157374084}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.1905524879693985, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 18.16, "ce": 0.053643432387616485, "kl": 0.025733816913998453, "context": 0.06651034147944301}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.19833362102508545, "loss_scale": 4.0, "retries": 0, "seconds": 90.9, "max_memory_gb": 18.16, "ce": 0.03746494752704166, "kl": 0.03559879600175009, "context": 0.05526434366765898}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 2.7611629962921143, "loss_scale": 4.0, "retries": 0, "seconds": 91.5, "max_memory_gb": 18.16, "ce": 0.15539117623120546, "kl": 0.09876603540033102, "context": 0.20477419532835484}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.07303284108638763, "loss_scale": 4.0, "retries": 0, "seconds": 92.2, "max_memory_gb": 18.16, "ce": 0.01110946579137817, "kl": 0.0028300197154749185, "context": 0.012524475750979036}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.6768840551376343, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 18.16, "ce": 0.023624871086212806, "kl": 0.6844284205035365, "context": 0.3658390766795492}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.9679599404335022, "loss_scale": 4.0, "retries": 0, "seconds": 93.4, "max_memory_gb": 18.16, "ce": 0.05498152179643512, "kl": 0.10347235793597065, "context": 0.10671770083718002}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.8080229759216309, "loss_scale": 4.0, "retries": 0, "seconds": 94.1, "max_memory_gb": 18.16, "ce": 0.38240435368788894, "kl": 0.020387090349686332, "context": 0.39259790029609576}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.245510071516037, "loss_scale": 4.0, "retries": 0, "seconds": 94.7, "max_memory_gb": 18.16, "ce": 0.044531080158776604, "kl": 0.028149226011009887, "context": 0.05860569467768073}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.1663549244403839, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 18.16, "ce": 0.009159324341453612, "kl": 0.09257169294869527, "context": 0.05544516915688291}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.8649271130561829, "loss_scale": 4.0, "retries": 0, "seconds": 95.8, "max_memory_gb": 18.16, "ce": 0.0264545485842973, "kl": 0.3064566419343464, "context": 0.17968287094845437}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.03246263787150383, "loss_scale": 4.0, "retries": 0, "seconds": 96.5, "max_memory_gb": 18.16, "ce": 0.0034264913338120095, "kl": 0.0008851219317875803, "context": 0.0038690521687385626}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.05769718438386917, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 18.16, "ce": 0.003669514393550344, "kl": 0.005448766547488049, "context": 0.006393897667294368}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.19875091314315796, "loss_scale": 4.0, "retries": 0, "seconds": 97.7, "max_memory_gb": 18.16, "ce": 0.012117310456233099, "kl": 0.08580444120525499, "context": 0.055019530234858394}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 2.2527573108673096, "loss_scale": 4.0, "retries": 0, "seconds": 98.4, "max_memory_gb": 18.16, "ce": 0.09539636109548155, "kl": 0.23689167276097578, "context": 0.21384219663013937}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.31531596183776855, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 18.16, "ce": 0.0819029373669764, "kl": 0.4850091665284708, "context": 0.3244075206457637}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.1778465509414673, "loss_scale": 4.0, "retries": 0, "seconds": 99.7, "max_memory_gb": 18.16, "ce": 0.2268725860158156, "kl": 0.013335193555121805, "context": 0.2335401832642674}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.8151841163635254, "loss_scale": 4.0, "retries": 0, "seconds": 100.4, "max_memory_gb": 18.16, "ce": 0.3449337752535939, "kl": 0.38448736630380154, "context": 0.5371774567756802}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.09387233853340149, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 18.16, "ce": 0.005078710353700444, "kl": 0.03916587773710489, "context": 0.02466164992074482}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 3.3795812129974365, "loss_scale": 4.0, "retries": 0, "seconds": 101.6, "max_memory_gb": 18.16, "ce": 0.3354569557122886, "kl": 0.2900545084357873, "context": 0.48048420215491205}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 0.8634588122367859, "loss_scale": 4.0, "retries": 0, "seconds": 102.3, "max_memory_gb": 18.16, "ce": 0.450477282050997, "kl": 0.05583619609387824, "context": 0.4783953856676817}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.06392166763544083, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 18.16, "ce": 0.005502071631781291, "kl": 0.0054622546563223295, "context": 0.00823319896881003}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.625909686088562, "loss_scale": 4.0, "retries": 0, "seconds": 103.6, "max_memory_gb": 18.16, "ce": 0.04564471935736947, "kl": 0.04325247365341056, "context": 0.06727095623500645}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.10466702282428741, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 18.16, "ce": 0.014867123682051897, "kl": 0.004983745981007814, "context": 0.01735899693449028}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.5522909760475159, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 18.16, "ce": 0.001393627953802934, "kl": 0.44018437605791405, "context": 0.22148580998327816}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.0854878425598145, "loss_scale": 4.0, "retries": 0, "seconds": 105.6, "max_memory_gb": 18.16, "ce": 0.01059409385197796, "kl": 0.7083078075502272, "context": 0.3647480113140773}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.330223947763443, "loss_scale": 4.0, "retries": 0, "seconds": 106.2, "max_memory_gb": 18.16, "ce": 0.059626264483085833, "kl": 0.010963610373437405, "context": 0.06510807009181008}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.5528402328491211, "loss_scale": 4.0, "retries": 0, "seconds": 106.9, "max_memory_gb": 18.16, "ce": 0.1474127100664191, "kl": 0.16077489008603152, "context": 0.22780015133321285}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.7919898629188538, "loss_scale": 4.0, "retries": 0, "seconds": 107.5, "max_memory_gb": 18.16, "ce": 0.28080491413129494, "kl": 0.040287258474563714, "context": 0.3009485367801972}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.7278779745101929, "loss_scale": 4.0, "retries": 0, "seconds": 108.2, "max_memory_gb": 18.16, "ce": 0.08082071965327486, "kl": 0.11276446154079167, "context": 0.13720295784878545}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.044589608907699585, "loss_scale": 4.0, "retries": 0, "seconds": 108.9, "max_memory_gb": 18.16, "ce": 0.002232952656413545, "kl": 0.003336050789641831, "context": 0.00390097815397894}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.8003522157669067, "loss_scale": 4.0, "retries": 0, "seconds": 109.5, "max_memory_gb": 18.16, "ce": 0.022660467591776978, "kl": 0.03741606487892568, "context": 0.0413684991071932}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.191671222448349, "loss_scale": 4.0, "retries": 0, "seconds": 110.2, "max_memory_gb": 18.16, "ce": 0.03895501440274529, "kl": 0.01185367233119905, "context": 0.04488185024820268}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.3934735953807831, "loss_scale": 4.0, "retries": 0, "seconds": 110.8, "max_memory_gb": 18.16, "ce": 0.004632411524653435, "kl": 0.5347243833093671, "context": 0.2719945965218358}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.08756950497627258, "loss_scale": 4.0, "retries": 0, "seconds": 111.4, "max_memory_gb": 18.16, "ce": 0.009319663367932662, "kl": 0.0038752531345380703, "context": 0.011257289992499864}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.19778138399124146, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 18.16, "ce": 0.06877758330665529, "kl": 0.04651077228481881, "context": 0.09203297318890691}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.04080832377076149, "loss_scale": 4.0, "retries": 0, "seconds": 112.4, "max_memory_gb": 18.16, "ce": 0.00503996922634542, "kl": 0.0018084675866703037, "context": 0.005944203105173074}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.8914674520492554, "loss_scale": 4.0, "retries": 0, "seconds": 113.0, "max_memory_gb": 18.16, "ce": 0.18576994113391265, "kl": 0.10597444418090163, "context": 0.23875715950271115}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 0.9354951977729797, "loss_scale": 4.0, "retries": 0, "seconds": 113.5, "max_memory_gb": 18.16, "ce": 0.5194884926822851, "kl": 0.15485736144182738, "context": 0.596917161354213}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.36138802766799927, "loss_scale": 4.0, "retries": 0, "seconds": 114.0, "max_memory_gb": 18.16, "ce": 0.14771630411269143, "kl": 0.10074471388361417, "context": 0.19808866071980447}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 2.426981210708618, "loss_scale": 4.0, "retries": 0, "seconds": 114.5, "max_memory_gb": 18.16, "ce": 0.07346208690432832, "kl": 1.0592524367384613, "context": 0.6030883136554621}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.5065577626228333, "loss_scale": 4.0, "retries": 0, "seconds": 115.0, "max_memory_gb": 18.16, "ce": 0.2241905099072028, "kl": 0.04365069052801118, "context": 0.24601586040807888}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.7668130993843079, "loss_scale": 4.0, "retries": 0, "seconds": 115.6, "max_memory_gb": 18.16, "ce": 0.42778084040037356, "kl": 0.1731422187294811, "context": 0.5143519558187108}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.6302787065505981, "loss_scale": 4.0, "retries": 0, "seconds": 116.1, "max_memory_gb": 18.16, "ce": 0.6303804616618436, "kl": 0.022339635801955637, "context": 0.64155027490051}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.19681987166404724, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 18.16, "ce": 0.01218225629418157, "kl": 0.1055387411997799, "context": 0.06495162527426146}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.03386642783880234, "loss_scale": 4.0, "retries": 0, "seconds": 117.1, "max_memory_gb": 18.16, "ce": 0.0020038165530422702, "kl": 0.0033573467644600896, "context": 0.0036824898706981912}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.7575856447219849, "loss_scale": 4.0, "retries": 0, "seconds": 117.6, "max_memory_gb": 18.16, "ce": 0.16211352008394897, "kl": 0.28187140449881554, "context": 0.3030492179095745}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.9681352972984314, "loss_scale": 4.0, "retries": 0, "seconds": 118.2, "max_memory_gb": 18.16, "ce": 0.009112019601161592, "kl": 0.17450792000454385, "context": 0.09636598275392316}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.19180582463741302, "loss_scale": 4.0, "retries": 0, "seconds": 118.7, "max_memory_gb": 18.16, "ce": 0.060637763046543114, "kl": 0.007231757848785492, "context": 0.06425364503229503}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.3820079565048218, "loss_scale": 4.0, "retries": 0, "seconds": 119.2, "max_memory_gb": 18.16, "ce": 0.008356048463610932, "kl": 0.518702984401898, "context": 0.2677075387327932}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.24675315618515015, "loss_scale": 4.0, "retries": 0, "seconds": 119.7, "max_memory_gb": 18.16, "ce": 0.03466674132505432, "kl": 0.07627193564076151, "context": 0.07280270727642346}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 1.3674052953720093, "loss_scale": 4.0, "retries": 0, "seconds": 120.2, "max_memory_gb": 18.16, "ce": 0.08323781201761449, "kl": 0.006250570939300815, "context": 0.08636309692519717}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.7938270568847656, "loss_scale": 4.0, "retries": 0, "seconds": 120.7, "max_memory_gb": 18.16, "ce": 0.04413888382259756, "kl": 0.20205712973256595, "context": 0.14516744890715927}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.2484951913356781, "loss_scale": 4.0, "retries": 0, "seconds": 121.3, "max_memory_gb": 18.16, "ce": 0.0065091080468846485, "kl": 0.2034016040415736, "context": 0.10820990955107845}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.47446727752685547, "loss_scale": 4.0, "retries": 0, "seconds": 121.7, "max_memory_gb": 18.16, "ce": 0.023856698302552104, "kl": 0.028114961882238276, "context": 0.03791417842148803}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.20403781533241272, "loss_scale": 4.0, "retries": 0, "seconds": 122.2, "max_memory_gb": 18.16, "ce": 0.016959247921477072, "kl": 0.007786294838297181, "context": 0.020852395296969917}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.4342653155326843, "loss_scale": 4.0, "retries": 0, "seconds": 122.7, "max_memory_gb": 18.16, "ce": 0.003987133779446594, "kl": 0.29992785667127464, "context": 0.1539510638685897}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 1.1929080486297607, "loss_scale": 4.0, "retries": 0, "seconds": 123.2, "max_memory_gb": 18.16, "ce": 0.07510731220827438, "kl": 0.043493661164575315, "context": 0.09685414208797738}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.6187242269515991, "loss_scale": 4.0, "retries": 0, "seconds": 123.8, "max_memory_gb": 18.16, "ce": 0.5274349360406632, "kl": 0.3586580588016659, "context": 0.7067639703163877}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.6230618953704834, "loss_scale": 4.0, "retries": 0, "seconds": 124.3, "max_memory_gb": 18.16, "ce": 0.04540930224175099, "kl": 0.15963640427798964, "context": 0.12522750344942324}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.14963126182556152, "loss_scale": 4.0, "retries": 0, "seconds": 124.9, "max_memory_gb": 18.16, "ce": 0.027202629076782614, "kl": 0.011775925871916115, "context": 0.033090591023210436}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.25389304757118225, "loss_scale": 4.0, "retries": 0, "seconds": 125.4, "max_memory_gb": 18.16, "ce": 0.010113343363627791, "kl": 0.08441372902598232, "context": 0.052320207469165325}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.29434823989868164, "loss_scale": 4.0, "retries": 0, "seconds": 125.9, "max_memory_gb": 18.16, "ce": 0.011670611798763275, "kl": 0.03770634427564801, "context": 0.030523783920216374}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.343923807144165, "loss_scale": 4.0, "retries": 0, "seconds": 126.4, "max_memory_gb": 18.16, "ce": 0.3803454174340004, "kl": 0.016355427927919663, "context": 0.388523144996725}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.084680937230587, "loss_scale": 4.0, "retries": 0, "seconds": 126.9, "max_memory_gb": 18.16, "ce": 0.010279748217726592, "kl": 0.017215369567566086, "context": 0.018887432641349733}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.5091727375984192, "loss_scale": 4.0, "retries": 0, "seconds": 127.5, "max_memory_gb": 18.16, "ce": 0.06918482220498845, "kl": 0.023706546278162932, "context": 0.08103809168096632}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.34352460503578186, "loss_scale": 4.0, "retries": 0, "seconds": 128.0, "max_memory_gb": 18.16, "ce": 0.031561302035697736, "kl": 0.009022316033224342, "context": 0.03607246121100616}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.33480289578437805, "loss_scale": 4.0, "retries": 0, "seconds": 128.5, "max_memory_gb": 18.16, "ce": 0.10961479635443538, "kl": 0.021336240570235532, "context": 0.12028291652677581}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.3301413655281067, "loss_scale": 4.0, "retries": 0, "seconds": 129.0, "max_memory_gb": 18.16, "ce": 0.10360276100982446, "kl": 0.03892290751355176, "context": 0.12306421615357976}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 1.1262208223342896, "loss_scale": 4.0, "retries": 0, "seconds": 129.5, "max_memory_gb": 18.16, "ce": 0.1251646450473345, "kl": 0.026949096303724218, "context": 0.13863919592404272}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.1120605543255806, "loss_scale": 4.0, "retries": 0, "seconds": 130.0, "max_memory_gb": 18.16, "ce": 0.007063552904583048, "kl": 0.06028173987942864, "context": 0.03720442167832516}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.06022348627448082, "loss_scale": 4.0, "retries": 0, "seconds": 130.6, "max_memory_gb": 18.16, "ce": 0.005608267325442284, "kl": 0.004706614578026347, "context": 0.007961574476212263}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.9655460119247437, "loss_scale": 4.0, "retries": 0, "seconds": 131.1, "max_memory_gb": 18.16, "ce": 0.04138711076666368, "kl": 0.21312253546420834, "context": 0.14794837896624813}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 2.1830437183380127, "loss_scale": 4.0, "retries": 0, "seconds": 131.6, "max_memory_gb": 18.16, "ce": 0.43063946662005037, "kl": 0.04215784720145166, "context": 0.4517183913849294}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.01686519756913185, "loss_scale": 4.0, "retries": 0, "seconds": 132.1, "max_memory_gb": 18.16, "ce": 0.0015149969622143544, "kl": 0.000849447240398149, "context": 0.0019397205614950508}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.39668557047843933, "loss_scale": 4.0, "retries": 0, "seconds": 132.6, "max_memory_gb": 18.16, "ce": 0.006870272787637077, "kl": 0.4846877374511678, "context": 0.24921413451374974}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.19301418960094452, "loss_scale": 4.0, "retries": 0, "seconds": 133.1, "max_memory_gb": 18.16, "ce": 0.031124230765271932, "kl": 0.008216290618292987, "context": 0.035232376772910357}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.37099117040634155, "loss_scale": 4.0, "retries": 0, "seconds": 133.7, "max_memory_gb": 18.16, "ce": 0.04723994294181466, "kl": 0.05285527545493096, "context": 0.0736675807274878}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.05412944778800011, "loss_scale": 4.0, "retries": 0, "seconds": 134.2, "max_memory_gb": 18.16, "ce": 0.006934169796295464, "kl": 0.0034225238050566986, "context": 0.008645431458717212}
|
runs/single_internvl/context/L16_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L16_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.16.self_attn.k_proj",
|
| 35 |
+
"model.language_model.layers.16.self_attn.q_proj",
|
| 36 |
+
"model.language_model.layers.16.mlp.up_proj",
|
| 37 |
+
"model.language_model.layers.16.self_attn.v_proj",
|
| 38 |
+
"model.language_model.layers.16.self_attn.o_proj",
|
| 39 |
+
"model.language_model.layers.16.mlp.down_proj",
|
| 40 |
+
"model.language_model.layers.16.mlp.gate_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L16_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 16,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L16_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 1.2319722175598145, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 17.73, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 1.5816937685012817, "loss_scale": 4.0, "retries": 0, "seconds": 1.9, "max_memory_gb": 17.74, "ce": 0.6389555425630533, "kl": 0.4673658170898989, "context": 0.8726384816254722}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 4.529516696929932, "loss_scale": 4.0, "retries": 0, "seconds": 2.3, "max_memory_gb": 17.74, "ce": 1.0992508679628372, "kl": 0.06730219349265099, "context": 1.132901981472969}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 2.5088653564453125, "loss_scale": 4.0, "retries": 0, "seconds": 2.8, "max_memory_gb": 17.74, "ce": 0.447088462926331, "kl": 0.430662757366008, "context": 0.662419846674311}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.04643377661705017, "loss_scale": 4.0, "retries": 0, "seconds": 3.3, "max_memory_gb": 17.74, "ce": 0.009248094203940127, "kl": 0.008221633281209506, "context": 0.013358911048271693}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 1.2712005376815796, "loss_scale": 4.0, "retries": 0, "seconds": 3.8, "max_memory_gb": 17.74, "ce": 0.28200617697439156, "kl": 0.07438646068840171, "context": 0.3191994063236052}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 3.959188461303711, "loss_scale": 4.0, "retries": 0, "seconds": 4.2, "max_memory_gb": 17.74, "ce": 0.7718550093177328, "kl": 0.12527147167156727, "context": 0.8344907407299615}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 3.7620463371276855, "loss_scale": 4.0, "retries": 0, "seconds": 4.7, "max_memory_gb": 17.75, "ce": 0.7145880274474621, "kl": 0.28356534731574357, "context": 0.8563706986606121}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 2.8637924194335938, "loss_scale": 4.0, "retries": 0, "seconds": 5.2, "max_memory_gb": 17.75, "ce": 0.4437823947519064, "kl": 0.136436051921919, "context": 0.5120004303753376}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.44045084714889526, "loss_scale": 4.0, "retries": 0, "seconds": 5.7, "max_memory_gb": 17.75, "ce": 0.04012025259726215, "kl": 0.04502242314629257, "context": 0.06263146200217307}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 3.2226321697235107, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 17.75, "ce": 0.537233399154502, "kl": 0.004832660577449133, "context": 0.5396497370384168}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 2.238381862640381, "loss_scale": 4.0, "retries": 0, "seconds": 6.6, "max_memory_gb": 17.75, "ce": 0.32995051738180337, "kl": 0.0037641099802385725, "context": 0.3318325654927321}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.7204259634017944, "loss_scale": 4.0, "retries": 0, "seconds": 7.1, "max_memory_gb": 17.75, "ce": 0.30343703570542857, "kl": 0.014591269056836609, "context": 0.31073266006205813}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 3.2855827808380127, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 17.75, "ce": 0.42983679473400116, "kl": 0.9154488197527826, "context": 0.887561209499836}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.3335177898406982, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 17.75, "ce": 0.21873645295613642, "kl": 0.4332627390160724, "context": 0.4353678224497344}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.8139029145240784, "loss_scale": 4.0, "retries": 0, "seconds": 8.4, "max_memory_gb": 17.75, "ce": 0.08512376115686493, "kl": 0.0026352268607752194, "context": 0.08644137252122164}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.9985685348510742, "loss_scale": 4.0, "retries": 0, "seconds": 8.8, "max_memory_gb": 17.75, "ce": 0.16359343999647535, "kl": 0.8323929276084527, "context": 0.5797899084864184}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.0310770273208618, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 17.75, "ce": 0.07354483184462879, "kl": 0.2281871897575911, "context": 0.18763842617045157}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 1.0408846139907837, "loss_scale": 4.0, "retries": 0, "seconds": 9.7, "max_memory_gb": 17.75, "ce": 0.06581765785813332, "kl": 0.17552482686005533, "context": 0.15358007233589888}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.08073427528142929, "loss_scale": 4.0, "retries": 0, "seconds": 10.1, "max_memory_gb": 17.75, "ce": 0.0025511375097266864, "kl": 0.01957241916352359, "context": 0.012337346859567333}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 0.17704032361507416, "loss_scale": 4.0, "retries": 0, "seconds": 10.6, "max_memory_gb": 17.75, "ce": 0.007585319341160357, "kl": 0.01734199763268407, "context": 0.01625631831757346}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.30719903111457825, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 17.75, "ce": 0.0358782771800179, "kl": 0.017825100912887137, "context": 0.044790828455006704}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 3.949448347091675, "loss_scale": 4.0, "retries": 0, "seconds": 11.5, "max_memory_gb": 17.75, "ce": 0.19682922237370803, "kl": 0.9719000103363555, "context": 0.6827792298699933}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 0.3523868918418884, "loss_scale": 4.0, "retries": 0, "seconds": 11.9, "max_memory_gb": 17.75, "ce": 0.014857353060506284, "kl": 0.07416656205896288, "context": 0.05194063461385667}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.099323034286499, "loss_scale": 4.0, "retries": 0, "seconds": 12.3, "max_memory_gb": 17.75, "ce": 0.3129321330034145, "kl": 0.1373327684123069, "context": 0.38159852381795645}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 0.2828271985054016, "loss_scale": 4.0, "retries": 0, "seconds": 12.8, "max_memory_gb": 17.75, "ce": 0.020881773147266358, "kl": 0.647398776258342, "context": 0.34458114800509065}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 0.2982996106147766, "loss_scale": 4.0, "retries": 0, "seconds": 13.2, "max_memory_gb": 17.75, "ce": 0.08496072042544256, "kl": 0.3192923987780887, "context": 0.2446069226534746}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 1.2198131084442139, "loss_scale": 4.0, "retries": 0, "seconds": 13.6, "max_memory_gb": 17.75, "ce": 0.46289670816622674, "kl": 0.9748985916376114, "context": 0.950346008874476}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.5356914401054382, "loss_scale": 4.0, "retries": 0, "seconds": 14.1, "max_memory_gb": 17.75, "ce": 0.04399340273812413, "kl": 0.3817104079644196, "context": 0.23484860768076032}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.2844121754169464, "loss_scale": 4.0, "retries": 0, "seconds": 14.5, "max_memory_gb": 17.75, "ce": 0.015760199996293522, "kl": 0.0009044391872521373, "context": 0.016212419910516473}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.6890127062797546, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 17.75, "ce": 0.17778790299780667, "kl": 0.10806185624096543, "context": 0.23181882360950112}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.16904953122138977, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 17.75, "ce": 0.008160828787367791, "kl": 0.006549293422722258, "context": 0.011435475607868284}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.33346301317214966, "loss_scale": 4.0, "retries": 0, "seconds": 15.8, "max_memory_gb": 17.75, "ce": 0.044160736724734306, "kl": 0.017306996160186827, "context": 0.05281423311680555}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.6070664525032043, "loss_scale": 4.0, "retries": 0, "seconds": 16.3, "max_memory_gb": 17.75, "ce": 0.0037004765836172737, "kl": 0.7847345011796278, "context": 0.3960677367358585}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.04317091405391693, "loss_scale": 4.0, "retries": 0, "seconds": 16.7, "max_memory_gb": 17.75, "ce": 0.0024420841509709135, "kl": 0.0009611186792426452, "context": 0.002922643441706896}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.30805668234825134, "loss_scale": 4.0, "retries": 0, "seconds": 17.2, "max_memory_gb": 17.75, "ce": 0.07952551599009894, "kl": 0.05343276564963162, "context": 0.10624190038652159}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.635492205619812, "loss_scale": 4.0, "retries": 0, "seconds": 17.6, "max_memory_gb": 17.75, "ce": 0.12163863080786541, "kl": 0.793500219238922, "context": 0.5183887349558063}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.37430837750434875, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 17.75, "ce": 0.017454226268455386, "kl": 0.46769361472979654, "context": 0.25130103318952024}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.5293652415275574, "loss_scale": 4.0, "retries": 0, "seconds": 18.5, "max_memory_gb": 17.75, "ce": 0.10709668917843373, "kl": 0.011555174729437567, "context": 0.11287427612114698}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.12179223448038101, "loss_scale": 4.0, "retries": 0, "seconds": 18.9, "max_memory_gb": 17.75, "ce": 0.007602894795127213, "kl": 0.008062628083280288, "context": 0.01163420919328928}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.4430935382843018, "loss_scale": 4.0, "retries": 0, "seconds": 19.3, "max_memory_gb": 17.75, "ce": 0.37909196471446194, "kl": 0.11942226230166852, "context": 0.43880308186635375}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.02662881836295128, "loss_scale": 4.0, "retries": 0, "seconds": 19.8, "max_memory_gb": 17.75, "ce": 0.0034217201537103392, "kl": 0.00038804997711849865, "context": 0.003615745124989189}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.045526858419179916, "loss_scale": 4.0, "retries": 0, "seconds": 20.2, "max_memory_gb": 17.75, "ce": 0.0022446137663791887, "kl": 0.0021288305142661557, "context": 0.003309029038064182}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 5.104233264923096, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 17.75, "ce": 0.46780399506678805, "kl": 1.2835047396292794, "context": 1.109556359529961}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.8080356121063232, "loss_scale": 4.0, "retries": 0, "seconds": 21.1, "max_memory_gb": 17.75, "ce": 0.7670933515473735, "kl": 0.2691634900547797, "context": 0.9016750784066971}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.6320818066596985, "loss_scale": 4.0, "retries": 0, "seconds": 21.5, "max_memory_gb": 17.75, "ce": 0.06590125848015305, "kl": 0.25917781243333593, "context": 0.19549016596283764}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.08844560384750366, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 17.75, "ce": 0.005352779608074343, "kl": 0.015219655237160623, "context": 0.012962606706423685}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.051266804337501526, "loss_scale": 4.0, "retries": 0, "seconds": 22.4, "max_memory_gb": 17.75, "ce": 0.0016162213360075839, "kl": 0.0031306592836699565, "context": 0.003181551001034677}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.391717791557312, "loss_scale": 4.0, "retries": 0, "seconds": 22.8, "max_memory_gb": 17.75, "ce": 0.01352538014180027, "kl": 0.3026531023059533, "context": 0.16485193740663817}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.5219753384590149, "loss_scale": 4.0, "retries": 0, "seconds": 23.2, "max_memory_gb": 17.75, "ce": 0.09080663228451158, "kl": 0.00802992016542703, "context": 0.09482159321487416}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 2.3101372718811035, "loss_scale": 4.0, "retries": 0, "seconds": 23.7, "max_memory_gb": 17.75, "ce": 0.4425612008126336, "kl": 0.03311379184162888, "context": 0.4591180934694421}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.5332924723625183, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 17.75, "ce": 0.021229013276752084, "kl": 0.17538295057602227, "context": 0.10892049293033779}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.7617713212966919, "loss_scale": 4.0, "retries": 0, "seconds": 24.6, "max_memory_gb": 17.75, "ce": 0.09887583972886205, "kl": 0.4813906028866768, "context": 0.33957114396616817}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 1.1211837530136108, "loss_scale": 4.0, "retries": 0, "seconds": 25.0, "max_memory_gb": 17.75, "ce": 0.035028248879825696, "kl": 0.28597314207581803, "context": 0.17801482579670846}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.1428728848695755, "loss_scale": 4.0, "retries": 0, "seconds": 25.5, "max_memory_gb": 17.75, "ce": 0.018602756375912577, "kl": 0.014308106041198698, "context": 0.025756810326129198}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.29583948850631714, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 17.75, "ce": 0.02185801276937127, "kl": 0.25578948005568236, "context": 0.14975275774486363}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.12227892130613327, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 17.75, "ce": 0.010239342853310518, "kl": 0.020706203235022258, "context": 0.020592443586792797}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.035881407558918, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 17.75, "ce": 0.001304059973335825, "kl": 0.0007606188024169569, "context": 0.0016843693629198242}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.2853132486343384, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 17.75, "ce": 0.041735532257007435, "kl": 0.06670631683664396, "context": 0.07508869250887074}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.31162741780281067, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 17.75, "ce": 0.0843187996360939, "kl": 0.06899556837743148, "context": 0.11881658365018666}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.7945164442062378, "loss_scale": 4.0, "retries": 0, "seconds": 28.0, "max_memory_gb": 17.75, "ce": 0.02396601709187962, "kl": 0.8115472139070334, "context": 0.4297396107867826}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.21081724762916565, "loss_scale": 4.0, "retries": 0, "seconds": 28.5, "max_memory_gb": 17.75, "ce": 0.024591358305769973, "kl": 0.00864206990081584, "context": 0.028912392794154584}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.2865503132343292, "loss_scale": 4.0, "retries": 0, "seconds": 28.9, "max_memory_gb": 17.75, "ce": 0.04016635240986943, "kl": 0.12653535199933685, "context": 0.10343402717262506}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.12457866966724396, "loss_scale": 4.0, "retries": 0, "seconds": 29.4, "max_memory_gb": 17.75, "ce": 0.01510238234186545, "kl": 0.01126182044390589, "context": 0.020733292447403073}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.7419996857643127, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 17.75, "ce": 0.14589713071472943, "kl": 0.15358047129120678, "context": 0.2226873654872179}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.6293042898178101, "loss_scale": 4.0, "retries": 0, "seconds": 30.2, "max_memory_gb": 17.75, "ce": 0.1558044792795954, "kl": 0.06323601368580967, "context": 0.18742248610078605}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.06744783371686935, "loss_scale": 4.0, "retries": 0, "seconds": 30.7, "max_memory_gb": 17.75, "ce": 0.005310065491357818, "kl": 0.00873762292394531, "context": 0.009678876958787441}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.16388416290283203, "loss_scale": 4.0, "retries": 0, "seconds": 31.1, "max_memory_gb": 17.75, "ce": 0.005562044607358985, "kl": 0.027530822921107756, "context": 0.01932745586236706}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 1.026323914527893, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 17.75, "ce": 0.31745226922794245, "kl": 0.025388464906427544, "context": 0.3301464952819515}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.03424953296780586, "loss_scale": 4.0, "retries": 0, "seconds": 32.0, "max_memory_gb": 17.75, "ce": 0.0023130252957344055, "kl": 0.0003449123141763266, "context": 0.002485481498297304}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.09006053954362869, "loss_scale": 4.0, "retries": 0, "seconds": 32.4, "max_memory_gb": 17.75, "ce": 0.009005183335830225, "kl": 0.015038112105685286, "context": 0.01652423950145021}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.31693875789642334, "loss_scale": 4.0, "retries": 0, "seconds": 32.9, "max_memory_gb": 17.75, "ce": 0.015285825589671731, "kl": 0.36893973930273205, "context": 0.19975570059614256}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.4469068050384521, "loss_scale": 4.0, "retries": 0, "seconds": 33.3, "max_memory_gb": 17.75, "ce": 0.28444774417221197, "kl": 0.057368131823750446, "context": 0.3131318128507701}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.2938985824584961, "loss_scale": 4.0, "retries": 0, "seconds": 33.7, "max_memory_gb": 17.75, "ce": 0.06554116206098115, "kl": 0.09210888262896333, "context": 0.11159560332453111}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.0557628870010376, "loss_scale": 4.0, "retries": 0, "seconds": 34.2, "max_memory_gb": 17.75, "ce": 0.44182629125134554, "kl": 0.9357189498841763, "context": 0.9096857588738203}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.3138376474380493, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 17.75, "ce": 0.018233422364573926, "kl": 1.6337685361504555, "context": 0.8351176885189489}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.0966842994093895, "loss_scale": 4.0, "retries": 0, "seconds": 35.1, "max_memory_gb": 17.75, "ce": 0.012319152854615822, "kl": 0.013002211426282884, "context": 0.01882025843951851}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.18083380162715912, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 17.75, "ce": 0.002363760519074276, "kl": 0.08512501957011409, "context": 0.0449262717156671}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.29903605580329895, "loss_scale": 4.0, "retries": 0, "seconds": 35.9, "max_memory_gb": 17.75, "ce": 0.015212225578579819, "kl": 0.02127527664470108, "context": 0.0258498629409587}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.20541560649871826, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 17.75, "ce": 0.0029040575645922218, "kl": 0.08141858392525592, "context": 0.04361335069552297}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.44740450382232666, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 17.75, "ce": 0.09062073053792119, "kl": 0.0442758639837848, "context": 0.1127586601069197}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.06990393996238708, "loss_scale": 4.0, "retries": 0, "seconds": 37.2, "max_memory_gb": 17.75, "ce": 0.006790952058508992, "kl": 0.0019143676818202948, "context": 0.007748135772999376}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.6575981974601746, "loss_scale": 4.0, "retries": 0, "seconds": 37.6, "max_memory_gb": 17.75, "ce": 0.14910627651261166, "kl": 0.702242801176908, "context": 0.5002276778104715}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.3362799286842346, "loss_scale": 4.0, "retries": 0, "seconds": 38.1, "max_memory_gb": 17.75, "ce": 0.0190781129822426, "kl": 0.045475834657736414, "context": 0.0418160305780475}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 0.2857571840286255, "loss_scale": 4.0, "retries": 0, "seconds": 38.5, "max_memory_gb": 17.75, "ce": 0.008216898298996966, "kl": 0.00279701614986827, "context": 0.009615406600460119}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.5065709352493286, "loss_scale": 4.0, "retries": 0, "seconds": 38.9, "max_memory_gb": 17.75, "ce": 0.0740868552820757, "kl": 0.29838315036613494, "context": 0.2232784393709153}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 1.0144890546798706, "loss_scale": 4.0, "retries": 0, "seconds": 39.4, "max_memory_gb": 17.75, "ce": 0.008034655766095966, "kl": 0.644941094564274, "context": 0.3305051983334124}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.14360055327415466, "loss_scale": 4.0, "retries": 0, "seconds": 39.8, "max_memory_gb": 17.75, "ce": 0.014866362640532316, "kl": 0.017909220701312734, "context": 0.02382097296140273}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.1699550747871399, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 17.75, "ce": 0.01868223168185068, "kl": 0.012509164691437036, "context": 0.024936813104432076}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.09461043030023575, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 17.75, "ce": 0.009500251355348155, "kl": 0.004519101610640064, "context": 0.011759802233427763}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 1.9492747783660889, "loss_scale": 4.0, "retries": 0, "seconds": 41.1, "max_memory_gb": 17.75, "ce": 0.41326973207469564, "kl": 0.2824895757221384, "context": 0.5545145311043598}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 2.584890365600586, "loss_scale": 4.0, "retries": 0, "seconds": 41.6, "max_memory_gb": 17.75, "ce": 0.15758929020375945, "kl": 0.30656878459922154, "context": 0.3108736902795499}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.7253048419952393, "loss_scale": 4.0, "retries": 0, "seconds": 42.0, "max_memory_gb": 17.75, "ce": 0.0583919171767775, "kl": 0.2885966899630148, "context": 0.20269026415189728}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 0.8042491674423218, "loss_scale": 4.0, "retries": 0, "seconds": 42.4, "max_memory_gb": 17.75, "ce": 0.0448789456859231, "kl": 0.022187779868545476, "context": 0.05597283644601703}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 2.0449233055114746, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 17.75, "ce": 0.4036335559503641, "kl": 0.039437796105630696, "context": 0.423352456651628}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.3956749439239502, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 17.75, "ce": 0.32938294897030573, "kl": 0.00788552808080567, "context": 0.33332572237122804}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 0.3960878551006317, "loss_scale": 4.0, "retries": 0, "seconds": 43.8, "max_memory_gb": 17.75, "ce": 0.019558706786483526, "kl": 0.025327188952360302, "context": 0.03222230193205178}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.15226949751377106, "loss_scale": 4.0, "retries": 0, "seconds": 44.2, "max_memory_gb": 17.75, "ce": 0.004288563497539144, "kl": 0.09574983795755543, "context": 0.05216348201793153}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.09719705581665039, "loss_scale": 4.0, "retries": 0, "seconds": 44.6, "max_memory_gb": 17.75, "ce": 0.012668361494434066, "kl": 0.0065512673354533035, "context": 0.01594399515306577}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 1.2363700866699219, "loss_scale": 4.0, "retries": 0, "seconds": 45.1, "max_memory_gb": 17.75, "ce": 0.039270549721550196, "kl": 0.4225883500621421, "context": 0.2505647170182783}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.3373444080352783, "loss_scale": 4.0, "retries": 0, "seconds": 45.5, "max_memory_gb": 17.75, "ce": 0.005418326472863555, "kl": 0.06716069096000865, "context": 0.038998671690933406}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.31649914383888245, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 17.75, "ce": 0.06671954470220953, "kl": 0.029319228953681886, "context": 0.08137915609404445}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.15366138517856598, "loss_scale": 4.0, "retries": 0, "seconds": 46.4, "max_memory_gb": 17.75, "ce": 0.010742080696218181, "kl": 0.010016119476858876, "context": 0.015750141217722557}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.08618797361850739, "loss_scale": 4.0, "retries": 0, "seconds": 46.8, "max_memory_gb": 17.75, "ce": 0.007370755425654352, "kl": 0.004185648616839899, "context": 0.009463579743169248}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.6562780141830444, "loss_scale": 4.0, "retries": 0, "seconds": 47.2, "max_memory_gb": 17.75, "ce": 0.032570125767961144, "kl": 0.3463633800856769, "context": 0.20575182256288826}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.05286172777414322, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 17.75, "ce": 0.004875424288911745, "kl": 0.00257990901809535, "context": 0.0061653788143303245}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 1.150240421295166, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 17.75, "ce": 0.07683309087587986, "kl": 0.13521338615186096, "context": 0.1444397895829752}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 2.1999425888061523, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 17.75, "ce": 0.23351499378622975, "kl": 0.016712786240532296, "context": 0.2418713934821426}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 2.298095464706421, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 17.75, "ce": 0.1331716906861402, "kl": 0.4183718338608742, "context": 0.34235760662704706}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.2787846326828003, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 17.75, "ce": 0.049266109243035316, "kl": 0.15738862217403948, "context": 0.12796042347326875}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.02968861162662506, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 17.75, "ce": 0.0024493031087331474, "kl": 0.0019338455313118175, "context": 0.0034162257506977767}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.10420659184455872, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 17.75, "ce": 0.008333651581779122, "kl": 0.0007623517522006296, "context": 0.008714827301446348}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.4173636734485626, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 17.75, "ce": 0.013523124973289669, "kl": 0.28111930401064456, "context": 0.1540827788412571}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.690385103225708, "loss_scale": 4.0, "retries": 0, "seconds": 51.2, "max_memory_gb": 17.75, "ce": 0.01768481489853002, "kl": 0.8860530875681434, "context": 0.4607113694655709}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.3118142783641815, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 17.75, "ce": 0.010880564572289586, "kl": 0.049510176249896176, "context": 0.03563565236981958}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 0.6317144632339478, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 17.75, "ce": 0.06475471268640831, "kl": 0.33452976759735975, "context": 0.23201960395090282}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 0.24181212484836578, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 17.76, "ce": 0.008784893609117717, "kl": 0.009290046084061032, "context": 0.013429916551103815}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.8104220032691956, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 17.76, "ce": 0.2224977606383618, "kl": 0.006784358607546892, "context": 0.2258899469889002}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.16627328097820282, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 17.76, "ce": 0.009997440472943708, "kl": 0.020264274721057518, "context": 0.02012957795523107}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 2.4856951236724854, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 17.76, "ce": 0.24491052300436422, "kl": 0.2027280426118523, "context": 0.3462745387223549}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.44285064935684204, "loss_scale": 4.0, "retries": 0, "seconds": 54.3, "max_memory_gb": 17.76, "ce": 0.0668761705965153, "kl": 0.03343300955748418, "context": 0.08359267399646342}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.93439120054245, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 17.76, "ce": 0.17231654387433082, "kl": 0.019046180259465473, "context": 0.18183963476622012}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.2629188001155853, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 17.76, "ce": 0.044872561818920076, "kl": 0.02369675738736987, "context": 0.0567209409782663}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.7695590257644653, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 17.76, "ce": 0.6627518907771446, "kl": 0.43967319471994415, "context": 0.8825884890393354}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.6798504590988159, "loss_scale": 4.0, "retries": 0, "seconds": 56.0, "max_memory_gb": 17.76, "ce": 0.15063567302422598, "kl": 0.07837077902877354, "context": 0.18982107000192627}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 2.279172420501709, "loss_scale": 4.0, "retries": 0, "seconds": 56.4, "max_memory_gb": 17.76, "ce": 0.7265258114784956, "kl": 0.09790552221238613, "context": 0.7754785735160112}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.14486528933048248, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 17.76, "ce": 0.017357919074129313, "kl": 0.0008631257733213715, "context": 0.017789481658837758}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.505720615386963, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 17.76, "ce": 0.05704627172963228, "kl": 0.9300430353614502, "context": 0.5220677991164848}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.3332342803478241, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 17.76, "ce": 0.08252488600555807, "kl": 0.030831664582365192, "context": 0.09794071922078729}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.2114253044128418, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 17.76, "ce": 0.04848963691620156, "kl": 0.2571890303515829, "context": 0.17708415701054037}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 3.050330400466919, "loss_scale": 4.0, "retries": 0, "seconds": 58.6, "max_memory_gb": 17.76, "ce": 0.40865440876223147, "kl": 0.3746871383918915, "context": 0.5959979774197564}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.45698508620262146, "loss_scale": 4.0, "retries": 0, "seconds": 59.0, "max_memory_gb": 17.76, "ce": 0.17543839052086696, "kl": 0.031302967116062064, "context": 0.19108986665378325}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 1.0265750885009766, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 17.76, "ce": 0.35228395817102864, "kl": 0.1692846491932869, "context": 0.4369262820109725}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.5506671071052551, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 17.76, "ce": 0.12850329758657608, "kl": 0.004555227591936273, "context": 0.13078091185889207}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.15374143421649933, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 17.76, "ce": 0.023046563263051212, "kl": 0.09747541211254429, "context": 0.07178426731843501}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 1.761008381843567, "loss_scale": 4.0, "retries": 0, "seconds": 60.8, "max_memory_gb": 17.76, "ce": 0.04231582389911637, "kl": 0.172406757628778, "context": 0.12851920002140105}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.5376363396644592, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 17.76, "ce": 0.02902974814060144, "kl": 0.11299000553117367, "context": 0.0855247510044137}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 0.5066392421722412, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 17.76, "ce": 0.06548799620941281, "kl": 0.09427362587302923, "context": 0.11262480914592743}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.677640974521637, "loss_scale": 4.0, "retries": 0, "seconds": 62.1, "max_memory_gb": 17.76, "ce": 0.007446596850058995, "kl": 0.0803003905020887, "context": 0.047596792377589736}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.07066713273525238, "loss_scale": 4.0, "retries": 0, "seconds": 62.5, "max_memory_gb": 17.76, "ce": 0.0068557552876882255, "kl": 0.0023523191011918243, "context": 0.008031914883758873}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.13126251101493835, "loss_scale": 4.0, "retries": 0, "seconds": 63.0, "max_memory_gb": 17.76, "ce": 0.011931957735214382, "kl": 0.00779799569863826, "context": 0.015830955671845004}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.2721940279006958, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 17.76, "ce": 0.02861450330237858, "kl": 0.0096104888652917, "context": 0.03341974824434146}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 2.370868444442749, "loss_scale": 4.0, "retries": 0, "seconds": 63.9, "max_memory_gb": 17.76, "ce": 0.34497529693180695, "kl": 0.03027610812932835, "context": 0.36011335672810674}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.9684215784072876, "loss_scale": 4.0, "retries": 0, "seconds": 64.3, "max_memory_gb": 17.76, "ce": 0.02707296295557171, "kl": 0.1512584846932441, "context": 0.10270220716483891}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.05022377893328667, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 17.76, "ce": 0.002516647888114676, "kl": 0.007422116159887082, "context": 0.006227706035133451}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.15000703930854797, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 17.76, "ce": 0.014887655503116548, "kl": 0.014826383147010347, "context": 0.022300846758298576}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.2746680974960327, "loss_scale": 4.0, "retries": 0, "seconds": 65.7, "max_memory_gb": 17.76, "ce": 0.45395328871381935, "kl": 0.47205172292888165, "context": 0.6899791344039841}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.09196636825799942, "loss_scale": 4.0, "retries": 0, "seconds": 66.2, "max_memory_gb": 17.76, "ce": 0.007128110515623121, "kl": 0.01252257045416627, "context": 0.013389395826379769}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.18821239471435547, "loss_scale": 4.0, "retries": 0, "seconds": 66.7, "max_memory_gb": 17.76, "ce": 0.014128931827144697, "kl": 0.004563076081467443, "context": 0.01641047015436925}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.27690714597702026, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 17.76, "ce": 0.06207523384364322, "kl": 0.027053053840063512, "context": 0.07560176216065884}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.2454889565706253, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 17.76, "ce": 0.02301519001775887, "kl": 0.026929805702934573, "context": 0.03648009285097942}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.5818483829498291, "loss_scale": 4.0, "retries": 0, "seconds": 68.1, "max_memory_gb": 17.76, "ce": 0.008375917066587135, "kl": 0.5135838402247828, "context": 0.2651678509428166}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.46188628673553467, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 17.76, "ce": 0.014320038084406406, "kl": 0.3514323125264127, "context": 0.1900361889929627}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 0.23226375877857208, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 17.76, "ce": 0.04551958345109597, "kl": 0.06921016308479011, "context": 0.08012466120999306}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.4126468002796173, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 17.76, "ce": 0.17292694015486632, "kl": 0.009368867811645032, "context": 0.17761137077468447}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.4827142059803009, "loss_scale": 4.0, "retries": 0, "seconds": 70.0, "max_memory_gb": 17.76, "ce": 0.021438008378027007, "kl": 0.009318906705402696, "context": 0.02609746097732568}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.1993989795446396, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 17.76, "ce": 0.01849867159035057, "kl": 0.025982336490415037, "context": 0.03148983954451978}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.9820252060890198, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 17.76, "ce": 0.2360683292208705, "kl": 0.08208657670184039, "context": 0.27711160827311687}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.544734001159668, "loss_scale": 4.0, "retries": 0, "seconds": 71.3, "max_memory_gb": 17.76, "ce": 0.008403486746829003, "kl": 0.05591864851885475, "context": 0.03636281075887382}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.5151125192642212, "loss_scale": 4.0, "retries": 0, "seconds": 71.8, "max_memory_gb": 17.76, "ce": 0.06346541142556816, "kl": 0.26437708266894333, "context": 0.19565396028338}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.48266929388046265, "loss_scale": 4.0, "retries": 0, "seconds": 72.2, "max_memory_gb": 17.76, "ce": 0.10369655440445058, "kl": 0.06143395111166683, "context": 0.1344135290128179}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.219137191772461, "loss_scale": 4.0, "retries": 0, "seconds": 72.6, "max_memory_gb": 17.76, "ce": 0.07102174827014096, "kl": 0.10186438303207979, "context": 0.12195393606089056}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.255819320678711, "loss_scale": 4.0, "retries": 0, "seconds": 73.0, "max_memory_gb": 17.76, "ce": 0.22634620714234188, "kl": 0.6635856115608476, "context": 0.5581390237202868}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 1.003279209136963, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 17.76, "ce": 0.10201258618326392, "kl": 0.0651606731917127, "context": 0.1345929267408792}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.30607208609580994, "loss_scale": 4.0, "retries": 0, "seconds": 73.9, "max_memory_gb": 17.76, "ce": 0.013783033529762179, "kl": 0.02986034243076574, "context": 0.028713204839732498}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.862757921218872, "loss_scale": 4.0, "retries": 0, "seconds": 74.4, "max_memory_gb": 17.76, "ce": 0.43067175813484937, "kl": 0.09053373607457615, "context": 0.47593861317727715}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.188022643327713, "loss_scale": 4.0, "retries": 0, "seconds": 74.8, "max_memory_gb": 17.76, "ce": 0.02232055552303791, "kl": 0.004482645803363994, "context": 0.0245618789922446}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.1706693470478058, "loss_scale": 4.0, "retries": 0, "seconds": 75.2, "max_memory_gb": 17.76, "ce": 0.038073682924732566, "kl": 0.0021801362236146815, "context": 0.03916375065455213}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.279062032699585, "loss_scale": 4.0, "retries": 0, "seconds": 75.7, "max_memory_gb": 17.76, "ce": 0.0854973090463318, "kl": 1.215721234679222, "context": 0.6933579221367836}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.057717856019735336, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 17.76, "ce": 0.003225672146072611, "kl": 0.0004076594668731559, "context": 0.0034295019577257335}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.6785839796066284, "loss_scale": 4.0, "retries": 0, "seconds": 76.5, "max_memory_gb": 17.76, "ce": 0.03116613377642352, "kl": 0.24527402082458138, "context": 0.15380314632784575}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.185169517993927, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 17.76, "ce": 0.01815204188460484, "kl": 0.015405456826556474, "context": 0.02585477032698691}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.2321605384349823, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 17.76, "ce": 0.05330396210774779, "kl": 0.024868192806025036, "context": 0.06573806027881801}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.2701021730899811, "loss_scale": 4.0, "retries": 0, "seconds": 77.8, "max_memory_gb": 17.76, "ce": 0.040360778861213475, "kl": 0.02893872861704949, "context": 0.054830143417348154}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 1.0399155616760254, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 17.76, "ce": 0.096397599671036, "kl": 0.07648251252248883, "context": 0.13463885709643364}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.08719189465045929, "loss_scale": 4.0, "retries": 0, "seconds": 78.7, "max_memory_gb": 17.76, "ce": 0.010111571697052568, "kl": 0.0022357728848874103, "context": 0.011229458206798881}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.7653400897979736, "loss_scale": 4.0, "retries": 0, "seconds": 79.1, "max_memory_gb": 17.76, "ce": 0.03325559926452115, "kl": 0.5823945860538515, "context": 0.32445288950111717}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.7823708653450012, "loss_scale": 4.0, "retries": 0, "seconds": 79.6, "max_memory_gb": 17.76, "ce": 0.08243867487180978, "kl": 0.049154727603308856, "context": 0.1070160404779017}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.9520983695983887, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 17.76, "ce": 0.414304635807639, "kl": 0.014149077753245365, "context": 0.4213791764341295}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.4440711736679077, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 17.76, "ce": 0.054416080994997174, "kl": 0.03628408996883081, "context": 0.07255812527728267}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.21577617526054382, "loss_scale": 4.0, "retries": 0, "seconds": 80.9, "max_memory_gb": 17.76, "ce": 0.010128644935321063, "kl": 0.11076293955557048, "context": 0.06551011558622122}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.5942585468292236, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 17.76, "ce": 0.032653366622980684, "kl": 0.2576258731423877, "context": 0.16146630130242556}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.049579519778490067, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 17.76, "ce": 0.004110555681108963, "kl": 0.0008922657289076596, "context": 0.004556688567390665}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.11227907240390778, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 17.76, "ce": 0.005762052373029292, "kl": 0.009009262670588214, "context": 0.010266683471854776}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.20153909921646118, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 17.76, "ce": 0.015191715181572363, "kl": 0.07720465350575978, "context": 0.05379404002451338}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.695488452911377, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 17.76, "ce": 0.06836436028243043, "kl": 0.1499459714104887, "context": 0.1433373422914883}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.33492693305015564, "loss_scale": 4.0, "retries": 0, "seconds": 83.8, "max_memory_gb": 17.76, "ce": 0.07645231852075085, "kl": 0.47911122589721344, "context": 0.3160079236840829}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.1667629480361938, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 17.76, "ce": 0.20447683627571678, "kl": 0.013956251454146695, "context": 0.21145495827659033}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 2.363664150238037, "loss_scale": 4.0, "retries": 0, "seconds": 84.8, "max_memory_gb": 17.76, "ce": 0.3141199405072257, "kl": 0.27507177158258855, "context": 0.4516558228060603}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.1232505738735199, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 17.76, "ce": 0.005291048553772271, "kl": 0.03841685570660047, "context": 0.02449947636341676}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 2.4564568996429443, "loss_scale": 4.0, "retries": 0, "seconds": 85.7, "max_memory_gb": 17.76, "ce": 0.29739360918756574, "kl": 0.2412491838040296, "context": 0.4180182014242746}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 0.9528209567070007, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 17.76, "ce": 0.4400935329613276, "kl": 0.050634556195291225, "context": 0.465410809207242}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.07525520026683807, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 17.76, "ce": 0.005918960276176222, "kl": 0.0036723114644701127, "context": 0.007755116181215271}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.4375089406967163, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 17.76, "ce": 0.05792311945697293, "kl": 0.05917314013640862, "context": 0.08750969130778685}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.19115978479385376, "loss_scale": 4.0, "retries": 0, "seconds": 87.6, "max_memory_gb": 17.76, "ce": 0.018867230173782445, "kl": 0.0055556675397383515, "context": 0.021645064000040293}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.6303121447563171, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 17.76, "ce": 0.0023015912738628685, "kl": 0.43005563490442, "context": 0.2173294046369847}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.3285831212997437, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 17.76, "ce": 0.02578806277597323, "kl": 0.656653359666052, "context": 0.35411475636647083}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.3062499463558197, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 17.76, "ce": 0.07232000818476081, "kl": 0.008487863869959256, "context": 0.07656393792422023}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.6746798753738403, "loss_scale": 4.0, "retries": 0, "seconds": 89.5, "max_memory_gb": 17.76, "ce": 0.14806572685483843, "kl": 0.13468724189442582, "context": 0.21540934860240668}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.6194244027137756, "loss_scale": 4.0, "retries": 0, "seconds": 90.0, "max_memory_gb": 17.76, "ce": 0.25776460813358426, "kl": 0.03239419839519542, "context": 0.27396170247811824}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.6395638585090637, "loss_scale": 4.0, "retries": 0, "seconds": 90.5, "max_memory_gb": 17.76, "ce": 0.06946144855464809, "kl": 0.08588351105208858, "context": 0.11240320405340753}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.03859371319413185, "loss_scale": 4.0, "retries": 0, "seconds": 90.9, "max_memory_gb": 17.76, "ce": 0.002267458323331084, "kl": 0.003628897128692188, "context": 0.004081906903593335}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.3595825731754303, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 17.76, "ce": 0.01660524183535017, "kl": 0.01973398239351809, "context": 0.026472233119420707}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.24482221901416779, "loss_scale": 4.0, "retries": 0, "seconds": 91.8, "max_memory_gb": 17.76, "ce": 0.03999587646103464, "kl": 0.011560261191334575, "context": 0.0457760066492483}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.5030831694602966, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 17.76, "ce": 0.005080799892311916, "kl": 0.5101894842155161, "context": 0.2601755372597836}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.10396161675453186, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 17.76, "ce": 0.00861313434143085, "kl": 0.0033342337101203157, "context": 0.010280251197400503}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.21354961395263672, "loss_scale": 4.0, "retries": 0, "seconds": 93.3, "max_memory_gb": 17.76, "ce": 0.07337126712081954, "kl": 0.04029238864313811, "context": 0.09351746318861842}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.05763465538620949, "loss_scale": 4.0, "retries": 0, "seconds": 93.8, "max_memory_gb": 17.76, "ce": 0.00567964423680678, "kl": 0.0020556797717290465, "context": 0.006707484048092738}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 1.3940664529800415, "loss_scale": 4.0, "retries": 0, "seconds": 94.2, "max_memory_gb": 17.76, "ce": 0.213137477170676, "kl": 0.09504482227930566, "context": 0.26065989857306704}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 0.9328556060791016, "loss_scale": 4.0, "retries": 0, "seconds": 94.7, "max_memory_gb": 17.76, "ce": 0.47157431405503303, "kl": 0.15727548771246802, "context": 0.5502120542078046}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.5555849075317383, "loss_scale": 4.0, "retries": 0, "seconds": 95.2, "max_memory_gb": 17.76, "ce": 0.16292811569292098, "kl": 0.08606490204692818, "context": 0.20596056012436748}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 2.368835926055908, "loss_scale": 4.0, "retries": 0, "seconds": 95.7, "max_memory_gb": 17.76, "ce": 0.11255127791082487, "kl": 1.0514400774482056, "context": 0.6382713139755651}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.6445244550704956, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 17.76, "ce": 0.2245038754772395, "kl": 0.03650524090335239, "context": 0.24275649985065684}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 0.9562677145004272, "loss_scale": 4.0, "retries": 0, "seconds": 96.7, "max_memory_gb": 17.76, "ce": 0.41877738456241786, "kl": 0.1549378631170839, "context": 0.4962463018600829}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.497400164604187, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 17.76, "ce": 0.5395244858809747, "kl": 0.017734659800908048, "context": 0.5483918237005128}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.2367219775915146, "loss_scale": 4.0, "retries": 0, "seconds": 97.6, "max_memory_gb": 17.76, "ce": 0.017115890368586406, "kl": 0.09497011081839446, "context": 0.06460094579961151}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.05291162431240082, "loss_scale": 4.0, "retries": 0, "seconds": 98.1, "max_memory_gb": 17.76, "ce": 0.003122669630101882, "kl": 0.0032222686568275094, "context": 0.0047338038857560605}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.8192787766456604, "loss_scale": 4.0, "retries": 0, "seconds": 98.6, "max_memory_gb": 17.76, "ce": 0.18519493122585118, "kl": 0.29542140290141106, "context": 0.3329056315124035}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.6824762225151062, "loss_scale": 4.0, "retries": 0, "seconds": 99.1, "max_memory_gb": 17.76, "ce": 0.011830246963654645, "kl": 0.14615881032659672, "context": 0.08490965407690965}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.2487679421901703, "loss_scale": 4.0, "retries": 0, "seconds": 99.5, "max_memory_gb": 17.76, "ce": 0.06539929447171744, "kl": 0.0054385398270824226, "context": 0.0681185648572864}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.45599135756492615, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 17.76, "ce": 0.006582457775948569, "kl": 0.48474223726952914, "context": 0.24895357474451885}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.3493168354034424, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 17.76, "ce": 0.040285558643518016, "kl": 0.0729700718805475, "context": 0.07677059504203498}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 2.20438551902771, "loss_scale": 4.0, "retries": 0, "seconds": 101.0, "max_memory_gb": 17.76, "ce": 0.16373723090509884, "kl": 0.0106396822957322, "context": 0.1690570694336202}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 1.3995792865753174, "loss_scale": 4.0, "retries": 0, "seconds": 101.4, "max_memory_gb": 17.76, "ce": 0.08147621026728302, "kl": 0.20465619358037657, "context": 0.18380430416436866}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.3016379177570343, "loss_scale": 4.0, "retries": 0, "seconds": 101.9, "max_memory_gb": 17.76, "ce": 0.007447967771440744, "kl": 0.1809534824351431, "context": 0.09792470608954318}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.31985875964164734, "loss_scale": 4.0, "retries": 0, "seconds": 102.4, "max_memory_gb": 17.76, "ce": 0.02221910160733387, "kl": 0.02354936175106559, "context": 0.03399378265021369}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.22699925303459167, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 17.76, "ce": 0.01878916163695976, "kl": 0.007332117802434368, "context": 0.022455220110714436}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.45922765135765076, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 17.76, "ce": 0.00767939371871762, "kl": 0.25652008334873244, "context": 0.13593943428713828}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 3.201777935028076, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 17.76, "ce": 0.1228431032068329, "kl": 0.09660664762395754, "context": 0.17114642885280773}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.9272056818008423, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 17.76, "ce": 0.5027903365553357, "kl": 0.3017790992744267, "context": 0.6536798847373575}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.7826371788978577, "loss_scale": 4.0, "retries": 0, "seconds": 104.7, "max_memory_gb": 17.76, "ce": 0.05612515781831462, "kl": 0.15589997966890223, "context": 0.13407514625578187}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.23935401439666748, "loss_scale": 4.0, "retries": 0, "seconds": 105.2, "max_memory_gb": 17.76, "ce": 0.033443997643189505, "kl": 0.010089904797496274, "context": 0.03848895087139681}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.29737481474876404, "loss_scale": 4.0, "retries": 0, "seconds": 105.7, "max_memory_gb": 17.76, "ce": 0.012174335424788296, "kl": 0.08809996000491083, "context": 0.056224314495921135}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.30905747413635254, "loss_scale": 4.0, "retries": 0, "seconds": 106.2, "max_memory_gb": 17.76, "ce": 0.01578939234605059, "kl": 0.026997083812602796, "context": 0.02928793430328369}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.6592460870742798, "loss_scale": 4.0, "retries": 0, "seconds": 106.6, "max_memory_gb": 17.76, "ce": 0.389337262880872, "kl": 0.009750907549459953, "context": 0.39421270963794086}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.140055850148201, "loss_scale": 4.0, "retries": 0, "seconds": 107.1, "max_memory_gb": 17.76, "ce": 0.013254599703941494, "kl": 0.02081293225637637, "context": 0.023661065351916477}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.3994230628013611, "loss_scale": 4.0, "retries": 0, "seconds": 107.6, "max_memory_gb": 17.76, "ce": 0.07285418314859271, "kl": 0.01772310498392926, "context": 0.08171573482104577}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.316872775554657, "loss_scale": 4.0, "retries": 0, "seconds": 108.0, "max_memory_gb": 17.76, "ce": 0.03361470889649354, "kl": 0.006567332431586692, "context": 0.036898375779855996}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.4017370045185089, "loss_scale": 4.0, "retries": 0, "seconds": 108.5, "max_memory_gb": 17.76, "ce": 0.13112111610826105, "kl": 0.019825031180516817, "context": 0.14103362511377782}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.4132571518421173, "loss_scale": 4.0, "retries": 0, "seconds": 109.0, "max_memory_gb": 17.76, "ce": 0.11260443681385368, "kl": 0.03632304694474442, "context": 0.1307659560116008}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 1.3360940217971802, "loss_scale": 4.0, "retries": 0, "seconds": 109.5, "max_memory_gb": 17.76, "ce": 0.19482493663963396, "kl": 0.026850794376514386, "context": 0.20825032808352262}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.16322046518325806, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 17.76, "ce": 0.011096136586274952, "kl": 0.0530784435059104, "context": 0.037635357410181314}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.08828333020210266, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 17.76, "ce": 0.007170670200139284, "kl": 0.004954112198902294, "context": 0.009647726430557668}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 1.620833158493042, "loss_scale": 4.0, "retries": 0, "seconds": 110.9, "max_memory_gb": 17.76, "ce": 0.04731651008478366, "kl": 0.23164171431380964, "context": 0.16313736491429154}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 1.1102392673492432, "loss_scale": 4.0, "retries": 0, "seconds": 111.4, "max_memory_gb": 17.76, "ce": 0.40667667239904404, "kl": 0.06009319017175585, "context": 0.43672326672822237}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.03187653049826622, "loss_scale": 4.0, "retries": 0, "seconds": 111.9, "max_memory_gb": 17.76, "ce": 0.0021903177694184706, "kl": 0.0010327360305382172, "context": 0.0027066858165198937}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.4834247827529907, "loss_scale": 4.0, "retries": 0, "seconds": 112.3, "max_memory_gb": 17.76, "ce": 0.009634551432100125, "kl": 0.47410824719190714, "context": 0.2466886822367087}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.2816905677318573, "loss_scale": 4.0, "retries": 0, "seconds": 112.8, "max_memory_gb": 17.76, "ce": 0.036656665964983404, "kl": 0.009864516323432326, "context": 0.041588925290852785}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.37335219979286194, "loss_scale": 4.0, "retries": 0, "seconds": 113.3, "max_memory_gb": 17.76, "ce": 0.04693033150397241, "kl": 0.050812262401450425, "context": 0.07233646255917847}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.08638470619916916, "loss_scale": 4.0, "retries": 0, "seconds": 113.8, "max_memory_gb": 17.76, "ce": 0.008923978894017637, "kl": 0.004037753889861051, "context": 0.010942856024485081}
|
runs/single_internvl/context/L20_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L20_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.20.mlp.down_proj",
|
| 35 |
+
"model.language_model.layers.20.self_attn.o_proj",
|
| 36 |
+
"model.language_model.layers.20.mlp.gate_proj",
|
| 37 |
+
"model.language_model.layers.20.self_attn.q_proj",
|
| 38 |
+
"model.language_model.layers.20.self_attn.k_proj",
|
| 39 |
+
"model.language_model.layers.20.self_attn.v_proj",
|
| 40 |
+
"model.language_model.layers.20.mlp.up_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L20_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 20,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L20_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 1.4590295553207397, "loss_scale": 4.0, "retries": 0, "seconds": 1.2, "max_memory_gb": 17.41, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 1.78007173538208, "loss_scale": 4.0, "retries": 0, "seconds": 1.5, "max_memory_gb": 17.43, "ce": 0.6366294079707586, "kl": 0.4660195497110635, "context": 0.8696392080310034}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 5.842503070831299, "loss_scale": 4.0, "retries": 0, "seconds": 1.9, "max_memory_gb": 17.43, "ce": 1.0934037640690804, "kl": 0.06840741634368896, "context": 1.1276074573397636}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 3.0623605251312256, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 17.43, "ce": 0.4437175701750675, "kl": 0.4299729057972854, "context": 0.6587040300000808}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.03569492697715759, "loss_scale": 4.0, "retries": 0, "seconds": 2.8, "max_memory_gb": 17.43, "ce": 0.00918852480026544, "kl": 0.008196011491236277, "context": 0.013286530618643155}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 1.599719762802124, "loss_scale": 4.0, "retries": 0, "seconds": 3.2, "max_memory_gb": 17.43, "ce": 0.27817397758190054, "kl": 0.07494046157444245, "context": 0.315644205926219}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 5.15983247756958, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 17.43, "ce": 0.7515319935773732, "kl": 0.12480065323666167, "context": 0.813932306924471}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 5.100747108459473, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 17.43, "ce": 0.6904933545738459, "kl": 0.283041100949049, "context": 0.8320138938724995}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 3.38368558883667, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 17.43, "ce": 0.4274580981582403, "kl": 0.13683928083628416, "context": 0.4958777353167534}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.5007935762405396, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 17.43, "ce": 0.03491228540951852, "kl": 0.04464597976766527, "context": 0.05723527551162988}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 4.241944313049316, "loss_scale": 4.0, "retries": 0, "seconds": 5.5, "max_memory_gb": 17.43, "ce": 0.49639689180185087, "kl": 0.004847807378610014, "context": 0.4988207812712062}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 2.6355199813842773, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 17.43, "ce": 0.3132792016094754, "kl": 0.003863185357204202, "context": 0.31521079959202325}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 2.2507987022399902, "loss_scale": 4.0, "retries": 0, "seconds": 6.3, "max_memory_gb": 17.43, "ce": 0.26802006128491485, "kl": 0.014076146499064635, "context": 0.27505814053074573}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 3.808793067932129, "loss_scale": 4.0, "retries": 0, "seconds": 6.8, "max_memory_gb": 17.43, "ce": 0.3734566271305084, "kl": 0.9031106205657125, "context": 0.8250119350850582}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.638715982437134, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 17.43, "ce": 0.17491526532876378, "kl": 0.4327442004456543, "context": 0.39128736555881005}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.8238896727561951, "loss_scale": 4.0, "retries": 0, "seconds": 7.7, "max_memory_gb": 17.43, "ce": 0.059422593367344234, "kl": 0.004056053744307064, "context": 0.06145062118594069}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 1.0013560056686401, "loss_scale": 4.0, "retries": 0, "seconds": 8.1, "max_memory_gb": 17.43, "ce": 0.1371016499469988, "kl": 0.8286110732587986, "context": 0.5514071777288336}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 0.5703598856925964, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 17.43, "ce": 0.032141720250365324, "kl": 0.24249948128999677, "context": 0.15339146347832866}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 0.6333978772163391, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 17.43, "ce": 0.028871652204543352, "kl": 0.17019960378820542, "context": 0.1139714540913701}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.08504945039749146, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 17.43, "ce": 0.0023108667473934474, "kl": 0.019928770980186528, "context": 0.012275251910978113}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 0.08364845812320709, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 17.43, "ce": 0.0027906876184715657, "kl": 0.01677944063249015, "context": 0.011180408371728845}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.16146539151668549, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 17.43, "ce": 0.02810489048715681, "kl": 0.013898886591050541, "context": 0.03505433407553937}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.0699622631073, "loss_scale": 4.0, "retries": 0, "seconds": 10.8, "max_memory_gb": 17.43, "ce": 0.2059809923857756, "kl": 0.9915683351068765, "context": 0.7017651593569099}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 0.20677560567855835, "loss_scale": 4.0, "retries": 0, "seconds": 11.2, "max_memory_gb": 17.43, "ce": 0.007989924168214202, "kl": 0.0649815768119879, "context": 0.04048071219585836}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.0563938617706299, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 17.43, "ce": 0.3158848427283374, "kl": 0.11903471150435507, "context": 0.3754021916538477}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 0.2925988733768463, "loss_scale": 4.0, "retries": 0, "seconds": 12.1, "max_memory_gb": 17.43, "ce": 0.017313783027930185, "kl": 0.6509802414802834, "context": 0.34280390210915357}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 0.3114781677722931, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 17.43, "ce": 0.08420245898378198, "kl": 0.3152887004639524, "context": 0.24184680593316443}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 1.1801263093948364, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 17.43, "ce": 0.44300341978669167, "kl": 0.987000689143315, "context": 0.9365037470124662}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.4496458172798157, "loss_scale": 4.0, "retries": 0, "seconds": 13.4, "max_memory_gb": 17.43, "ce": 0.05082357977516949, "kl": 0.3629166174941929, "context": 0.23228189459769055}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.23124264180660248, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 17.43, "ce": 0.012387108705297578, "kl": 0.0008308256960845029, "context": 0.01280252151627792}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.7295592427253723, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 17.43, "ce": 0.176368516869843, "kl": 0.11039229782181792, "context": 0.2315646675415337}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.10880402475595474, "loss_scale": 4.0, "retries": 0, "seconds": 14.7, "max_memory_gb": 17.43, "ce": 0.006317058519925922, "kl": 0.005484528897795826, "context": 0.009059322823304683}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.5952839851379395, "loss_scale": 4.0, "retries": 0, "seconds": 15.2, "max_memory_gb": 17.43, "ce": 0.05794963100925088, "kl": 0.017276032827794552, "context": 0.06658764788880944}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.6218534111976624, "loss_scale": 4.0, "retries": 0, "seconds": 15.6, "max_memory_gb": 17.43, "ce": 0.003538993016263703, "kl": 0.769402709465794, "context": 0.3882403482857626}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.0334814190864563, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 17.43, "ce": 0.0021921198786003515, "kl": 0.0008407424511460704, "context": 0.0026124911819351837}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.3715822994709015, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 17.43, "ce": 0.0866326438554097, "kl": 0.05331891038804315, "context": 0.1132921010866994}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.6451432704925537, "loss_scale": 4.0, "retries": 0, "seconds": 16.9, "max_memory_gb": 17.43, "ce": 0.11707513089640997, "kl": 0.775784456927795, "context": 0.5049673655303195}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.3308546245098114, "loss_scale": 4.0, "retries": 0, "seconds": 17.4, "max_memory_gb": 17.43, "ce": 0.021094653522595763, "kl": 0.46581189129210543, "context": 0.25400060089305043}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.5844442844390869, "loss_scale": 4.0, "retries": 0, "seconds": 17.8, "max_memory_gb": 17.43, "ce": 0.10489695554133505, "kl": 0.009716941283841152, "context": 0.10975542708183639}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.08519139140844345, "loss_scale": 4.0, "retries": 0, "seconds": 18.2, "max_memory_gb": 17.43, "ce": 0.007418603723635897, "kl": 0.006961180624784902, "context": 0.010899193934164941}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.0804206132888794, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 17.43, "ce": 0.3684101404505782, "kl": 0.11210921686142683, "context": 0.4244647556915879}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.020837802439928055, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 17.43, "ce": 0.0031220944656524807, "kl": 0.000338321398885455, "context": 0.003291255154181272}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.03592628240585327, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 17.43, "ce": 0.002527493066736497, "kl": 0.002524338138755411, "context": 0.003789662121562287}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 5.256229400634766, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 17.43, "ce": 0.3888126241945429, "kl": 1.2045047816936858, "context": 0.9910650504025398}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.9830766916275024, "loss_scale": 4.0, "retries": 0, "seconds": 20.4, "max_memory_gb": 17.43, "ce": 0.7345142573467456, "kl": 0.3110611511656316, "context": 0.8900448159256484}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.6168208718299866, "loss_scale": 4.0, "retries": 0, "seconds": 20.8, "max_memory_gb": 17.43, "ce": 0.06050255085574463, "kl": 0.24570858286460862, "context": 0.1833568379515782}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.16110387444496155, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 17.43, "ce": 0.007095586137438659, "kl": 0.018647240998689085, "context": 0.01641920703696087}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.06652951240539551, "loss_scale": 4.0, "retries": 0, "seconds": 21.7, "max_memory_gb": 17.43, "ce": 0.002800227332045324, "kl": 0.0032193428924074396, "context": 0.004409898887388408}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.43266505002975464, "loss_scale": 4.0, "retries": 0, "seconds": 22.1, "max_memory_gb": 17.43, "ce": 0.012503619909693953, "kl": 0.3061961956577761, "context": 0.16560171075252583}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.44683152437210083, "loss_scale": 4.0, "retries": 0, "seconds": 22.6, "max_memory_gb": 17.43, "ce": 0.07518033167434623, "kl": 0.008170026543666609, "context": 0.07926534212310798}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 3.214966058731079, "loss_scale": 4.0, "retries": 0, "seconds": 23.0, "max_memory_gb": 17.43, "ce": 0.3396517083856452, "kl": 0.044177739826864126, "context": 0.36174058947653975}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.3487529754638672, "loss_scale": 4.0, "retries": 0, "seconds": 23.4, "max_memory_gb": 17.43, "ce": 0.022971620288444683, "kl": 0.143459607148543, "context": 0.09470142144709826}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.6475163698196411, "loss_scale": 4.0, "retries": 0, "seconds": 23.9, "max_memory_gb": 17.43, "ce": 0.09023297717794776, "kl": 0.46499473229050636, "context": 0.32273034006357193}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 1.9940190315246582, "loss_scale": 4.0, "retries": 0, "seconds": 24.3, "max_memory_gb": 17.43, "ce": 0.041798733931500465, "kl": 0.2107516900432529, "context": 0.14717457591905259}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.1597791612148285, "loss_scale": 4.0, "retries": 0, "seconds": 24.7, "max_memory_gb": 17.43, "ce": 0.020175372570520267, "kl": 0.014988486822403502, "context": 0.02766961621819064}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.30648520588874817, "loss_scale": 4.0, "retries": 0, "seconds": 25.2, "max_memory_gb": 17.43, "ce": 0.02265405119396746, "kl": 0.24025317071937025, "context": 0.14278063410893083}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.13155564665794373, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 17.43, "ce": 0.01005603585508652, "kl": 0.019770676117332187, "context": 0.019941373931942508}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.04355127736926079, "loss_scale": 4.0, "retries": 0, "seconds": 26.0, "max_memory_gb": 17.43, "ce": 0.0019352707095094956, "kl": 0.0010422195705359627, "context": 0.0024563805709476583}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.3471390902996063, "loss_scale": 4.0, "retries": 0, "seconds": 26.5, "max_memory_gb": 17.43, "ce": 0.04413880512584001, "kl": 0.06929239176679403, "context": 0.07878499914659187}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.35079559683799744, "loss_scale": 4.0, "retries": 0, "seconds": 26.9, "max_memory_gb": 17.43, "ce": 0.08115507813636214, "kl": 0.06525251944549382, "context": 0.1137813376262784}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.8055300712585449, "loss_scale": 4.0, "retries": 0, "seconds": 27.3, "max_memory_gb": 17.43, "ce": 0.03022584517020732, "kl": 0.7561344536443357, "context": 0.40829307385138236}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.14643117785453796, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 17.43, "ce": 0.0203113044553902, "kl": 0.009705642631161027, "context": 0.025164124788716435}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.3106098473072052, "loss_scale": 4.0, "retries": 0, "seconds": 28.2, "max_memory_gb": 17.43, "ce": 0.04272790369577706, "kl": 0.1165416735020699, "context": 0.10099874250590801}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.12850423157215118, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 17.43, "ce": 0.015987171849701554, "kl": 0.01221005863044411, "context": 0.022092201048508286}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.6521686911582947, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 17.43, "ce": 0.12016367982141674, "kl": 0.1478828187682666, "context": 0.19410509569570422}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.6648927330970764, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 17.43, "ce": 0.14172255479547857, "kl": 0.06009675246923507, "context": 0.17177093478449024}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.06657734513282776, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 17.43, "ce": 0.00612570348312147, "kl": 0.008521442159690196, "context": 0.010386424837633967}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.1079975962638855, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 17.43, "ce": 0.006359036844514776, "kl": 0.02101546252379194, "context": 0.016866768193722237}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 1.5151896476745605, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 17.43, "ce": 0.3813709410314914, "kl": 0.02287343882198911, "context": 0.3928076521260664}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.07775301486253738, "loss_scale": 4.0, "retries": 0, "seconds": 31.3, "max_memory_gb": 17.43, "ce": 0.004152380104642361, "kl": 0.0004707925008915481, "context": 0.004387776396470144}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.10408937931060791, "loss_scale": 4.0, "retries": 0, "seconds": 31.7, "max_memory_gb": 17.43, "ce": 0.009601881603884976, "kl": 0.01500730845873477, "context": 0.017105535604059696}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.33736470341682434, "loss_scale": 4.0, "retries": 0, "seconds": 32.2, "max_memory_gb": 17.43, "ce": 0.015269813331542537, "kl": 0.38211002841126174, "context": 0.20632483263034374}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.3399064540863037, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 17.43, "ce": 0.21682346364832483, "kl": 0.044994297555604135, "context": 0.2393206086999271}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.2504763603210449, "loss_scale": 4.0, "retries": 0, "seconds": 33.0, "max_memory_gb": 17.43, "ce": 0.0751807359229133, "kl": 0.08913200673123356, "context": 0.11974673937584157}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.0622806549072266, "loss_scale": 4.0, "retries": 0, "seconds": 33.5, "max_memory_gb": 17.43, "ce": 0.47386760004155803, "kl": 0.893038097769022, "context": 0.9203866412863135}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.2058604955673218, "loss_scale": 4.0, "retries": 0, "seconds": 33.9, "max_memory_gb": 17.43, "ce": 0.01687932331697084, "kl": 1.623361369012855, "context": 0.8285599981900305}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.13544629514217377, "loss_scale": 4.0, "retries": 0, "seconds": 34.3, "max_memory_gb": 17.43, "ce": 0.014501469791866839, "kl": 0.014645274786744267, "context": 0.021824107156135142}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.2065483182668686, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 17.43, "ce": 0.0023367464600596577, "kl": 0.08321637190238107, "context": 0.043944933742750436}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.21659167110919952, "loss_scale": 4.0, "retries": 0, "seconds": 35.2, "max_memory_gb": 17.43, "ce": 0.013792799261864275, "kl": 0.015937344956910238, "context": 0.02176147077989299}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.1891402304172516, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 17.43, "ce": 0.003346259831232601, "kl": 0.06757338315844663, "context": 0.037132950468730996}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.4086894989013672, "loss_scale": 4.0, "retries": 0, "seconds": 36.0, "max_memory_gb": 17.43, "ce": 0.08717048610560596, "kl": 0.028728247198159806, "context": 0.10153460968285799}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.07796826958656311, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 17.43, "ce": 0.008054286998230964, "kl": 0.0022424659109674394, "context": 0.009175520099233836}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.8515313267707825, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 17.43, "ce": 0.1511872747214511, "kl": 0.6323062885130639, "context": 0.4673404161585495}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.3333689272403717, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 17.43, "ce": 0.028020928022669978, "kl": 0.04483404051825346, "context": 0.0504379472185974}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 0.06441715359687805, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 17.43, "ce": 0.002285135213242029, "kl": 0.000849529958472317, "context": 0.0027099002199975075}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.0212243795394897, "loss_scale": 4.0, "retries": 0, "seconds": 38.2, "max_memory_gb": 17.43, "ce": 0.05146839760709554, "kl": 0.23303778923582286, "context": 0.167987288441509}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.919080913066864, "loss_scale": 4.0, "retries": 0, "seconds": 38.6, "max_memory_gb": 17.43, "ce": 0.008744585502427071, "kl": 0.6363969976082444, "context": 0.32694309717044234}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.2870903015136719, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 17.43, "ce": 0.0223031496097974, "kl": 0.013695430367576478, "context": 0.029150864778785035}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.3578581213951111, "loss_scale": 4.0, "retries": 0, "seconds": 39.5, "max_memory_gb": 17.43, "ce": 0.025923942286681267, "kl": 0.021779876267828513, "context": 0.03681387849064777}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.15714870393276215, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 17.43, "ce": 0.011107431433629245, "kl": 0.005350622421246953, "context": 0.013782742549665272}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 2.752161741256714, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 17.43, "ce": 0.5620770691311918, "kl": 0.22485357311961707, "context": 0.674503842194099}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.9731318950653076, "loss_scale": 4.0, "retries": 0, "seconds": 40.8, "max_memory_gb": 17.43, "ce": 0.06884440993599128, "kl": 0.2411640100481236, "context": 0.18942641423200257}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.1871824264526367, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 17.43, "ce": 0.050266940001165494, "kl": 0.1967766959569417, "context": 0.14865528437076136}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 0.48708540201187134, "loss_scale": 4.0, "retries": 0, "seconds": 41.7, "max_memory_gb": 17.43, "ce": 0.02731466552359052, "kl": 0.022778558763093315, "context": 0.038703944912413135}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.8533868789672852, "loss_scale": 4.0, "retries": 0, "seconds": 42.1, "max_memory_gb": 17.43, "ce": 0.32720155267452355, "kl": 0.06440478900913149, "context": 0.3594039441086352}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.7158684730529785, "loss_scale": 4.0, "retries": 0, "seconds": 42.6, "max_memory_gb": 17.43, "ce": 0.26567180597339757, "kl": 0.012080118562153075, "context": 0.2717118702130392}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 0.3808671534061432, "loss_scale": 4.0, "retries": 0, "seconds": 43.0, "max_memory_gb": 17.43, "ce": 0.016993138153338805, "kl": 0.0325813292292878, "context": 0.033283802680671215}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.16177786886692047, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 17.43, "ce": 0.004596342914737761, "kl": 0.08229222648151335, "context": 0.04574245730327675}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.10119668394327164, "loss_scale": 4.0, "retries": 0, "seconds": 43.9, "max_memory_gb": 17.43, "ce": 0.013608573670353508, "kl": 0.0037594231216644403, "context": 0.015488285003812052}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.7325934767723083, "loss_scale": 4.0, "retries": 0, "seconds": 44.3, "max_memory_gb": 17.43, "ce": 0.04031593873514794, "kl": 0.37348852603827254, "context": 0.22706019831821322}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.39551404118537903, "loss_scale": 4.0, "retries": 0, "seconds": 44.7, "max_memory_gb": 17.43, "ce": 0.005332737433491275, "kl": 0.09055385447572917, "context": 0.05060966464225203}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.30780673027038574, "loss_scale": 4.0, "retries": 0, "seconds": 45.2, "max_memory_gb": 17.43, "ce": 0.06024105520918965, "kl": 0.02938046772032976, "context": 0.07493129116483033}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.10728354007005692, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 17.43, "ce": 0.009640236996347085, "kl": 0.008846853670547716, "context": 0.014063664071727544}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.20399275422096252, "loss_scale": 4.0, "retries": 0, "seconds": 46.0, "max_memory_gb": 17.43, "ce": 0.011648617415630724, "kl": 0.005115276864671614, "context": 0.014206255524186417}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.46492430567741394, "loss_scale": 4.0, "retries": 0, "seconds": 46.4, "max_memory_gb": 17.43, "ce": 0.03758461441611871, "kl": 0.3149016574025154, "context": 0.19503544154576957}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.04618300497531891, "loss_scale": 4.0, "retries": 0, "seconds": 46.9, "max_memory_gb": 17.43, "ce": 0.004708804422989488, "kl": 0.002523512150219176, "context": 0.005970560596324503}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 0.4975419044494629, "loss_scale": 4.0, "retries": 0, "seconds": 47.3, "max_memory_gb": 17.43, "ce": 0.018177331949118525, "kl": 0.17779076874853672, "context": 0.10707271473074798}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 1.897992730140686, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 17.43, "ce": 0.3628122572990833, "kl": 0.013431661176355192, "context": 0.36952810172806494}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 1.8871122598648071, "loss_scale": 4.0, "retries": 0, "seconds": 48.2, "max_memory_gb": 17.43, "ce": 0.14720601204317063, "kl": 0.3518263138830662, "context": 0.32311917189508677}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 0.9294307231903076, "loss_scale": 4.0, "retries": 0, "seconds": 48.6, "max_memory_gb": 17.43, "ce": 0.03929390059784055, "kl": 0.036022286163643, "context": 0.05730504426173866}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.022702375426888466, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 17.43, "ce": 0.0018701931403484195, "kl": 0.0017409043575753458, "context": 0.0027406452572904527}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.09127989411354065, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 17.43, "ce": 0.00872225931379944, "kl": 0.0015916419815766858, "context": 0.009518080158159137}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.7182163000106812, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 17.43, "ce": 0.016893642488867044, "kl": 0.2631824993295595, "context": 0.14848489011637866}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.339119791984558, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 17.43, "ce": 0.016464120970340446, "kl": 0.8116697008663323, "context": 0.42229895264608786}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.7177527546882629, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 17.43, "ce": 0.017375864612404257, "kl": 0.05089364654122619, "context": 0.04282268678070977}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 0.5892314314842224, "loss_scale": 4.0, "retries": 0, "seconds": 51.2, "max_memory_gb": 17.43, "ce": 0.061770426109433174, "kl": 0.36993687456606494, "context": 0.24673887081735302}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 0.12474849075078964, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 17.44, "ce": 0.005949444457655773, "kl": 0.005534401621844154, "context": 0.008716645301319659}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.9120950102806091, "loss_scale": 4.0, "retries": 0, "seconds": 52.0, "max_memory_gb": 17.44, "ce": 0.24902319205284584, "kl": 0.004609224737578188, "context": 0.25132780810236}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.17609155178070068, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 17.44, "ce": 0.011862875500810333, "kl": 0.019088812479026274, "context": 0.021407281528809108}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.4973896741867065, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 17.44, "ce": 0.21465037914458662, "kl": 0.2005518880323507, "context": 0.3149263139348477}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.7288948893547058, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 17.44, "ce": 0.08601042210648302, "kl": 0.04481054850111832, "context": 0.10841569821059238}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 1.4408925771713257, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 17.44, "ce": 0.2027026844516513, "kl": 0.018982908452017, "context": 0.21219413296785206}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.3471967279911041, "loss_scale": 4.0, "retries": 0, "seconds": 54.2, "max_memory_gb": 17.44, "ce": 0.04475720995105803, "kl": 0.0299777791951783, "context": 0.059746101149357855}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.6607682704925537, "loss_scale": 4.0, "retries": 0, "seconds": 54.7, "max_memory_gb": 17.44, "ce": 0.6766616315289866, "kl": 0.3915491316729458, "context": 0.8724361815548036}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.5751215219497681, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 17.44, "ce": 0.16993627676856704, "kl": 0.05813453498558374, "context": 0.19900353875709698}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.9519726037979126, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 17.44, "ce": 0.7117552161216736, "kl": 0.1060451939702034, "context": 0.7647778242826462}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.1534527987241745, "loss_scale": 4.0, "retries": 0, "seconds": 56.0, "max_memory_gb": 17.44, "ce": 0.017438133945688605, "kl": 0.0008098506586975418, "context": 0.01784305917681195}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.4248194694519043, "loss_scale": 4.0, "retries": 0, "seconds": 56.4, "max_memory_gb": 17.44, "ce": 0.04205541124974843, "kl": 0.8243524878635071, "context": 0.4542316566221416}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.3368087410926819, "loss_scale": 4.0, "retries": 0, "seconds": 56.9, "max_memory_gb": 17.44, "ce": 0.0843405710766092, "kl": 0.028234510340553243, "context": 0.09845782537013292}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.25774237513542175, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 17.44, "ce": 0.04619685828220099, "kl": 0.2835838664250332, "context": 0.18798879568930715}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 3.063838481903076, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 17.44, "ce": 0.37667796231107786, "kl": 0.3032837637292687, "context": 0.5283198292599991}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 0.5085240602493286, "loss_scale": 4.0, "retries": 0, "seconds": 58.2, "max_memory_gb": 17.44, "ce": 0.1707285970041994, "kl": 0.03403631550463615, "context": 0.18774675985332578}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 1.2576524019241333, "loss_scale": 4.0, "retries": 0, "seconds": 58.6, "max_memory_gb": 17.44, "ce": 0.3815022477065213, "kl": 0.15178912552073598, "context": 0.4573968155309558}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.5528141260147095, "loss_scale": 4.0, "retries": 0, "seconds": 59.0, "max_memory_gb": 17.44, "ce": 0.13555632450152189, "kl": 0.0024999873628530622, "context": 0.1368063256086316}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.21274833381175995, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 17.44, "ce": 0.016190930851735175, "kl": 0.11563970544375479, "context": 0.07401078712427989}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 1.5158169269561768, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 17.44, "ce": 0.0353808244690299, "kl": 0.18906222764053382, "context": 0.12991193949710578}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.49042657017707825, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 17.44, "ce": 0.03656268924532924, "kl": 0.0768446993897669, "context": 0.07498503886745311}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 0.4863108992576599, "loss_scale": 4.0, "retries": 0, "seconds": 60.7, "max_memory_gb": 17.44, "ce": 0.06822736607864499, "kl": 0.09333021892234683, "context": 0.11489247251302004}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.9081736207008362, "loss_scale": 4.0, "retries": 0, "seconds": 61.2, "max_memory_gb": 17.44, "ce": 0.008133156254189089, "kl": 0.09661234832128685, "context": 0.05643933133978862}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.0850965678691864, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 17.44, "ce": 0.009195544669637457, "kl": 0.0038366234584827907, "context": 0.011113856104202569}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.1403789222240448, "loss_scale": 4.0, "retries": 0, "seconds": 62.1, "max_memory_gb": 17.44, "ce": 0.011927231680601835, "kl": 0.005820182210300118, "context": 0.014837322407402098}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.26932889223098755, "loss_scale": 4.0, "retries": 0, "seconds": 62.5, "max_memory_gb": 17.44, "ce": 0.03195401199627668, "kl": 0.012189406290417537, "context": 0.03804871469037607}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.8961385488510132, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 17.44, "ce": 0.3258642994769616, "kl": 0.023536186057754094, "context": 0.3376323847915046}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.7084206938743591, "loss_scale": 4.0, "retries": 0, "seconds": 63.4, "max_memory_gb": 17.44, "ce": 0.04122256813570857, "kl": 0.189561280538328, "context": 0.13600320648401976}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.042604122310876846, "loss_scale": 4.0, "retries": 0, "seconds": 63.8, "max_memory_gb": 17.44, "ce": 0.002653244649991393, "kl": 0.006617161083340761, "context": 0.005961825110716745}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.15715667605400085, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 17.44, "ce": 0.01736276736482978, "kl": 0.007787728405673988, "context": 0.02125663112383336}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.3081551790237427, "loss_scale": 4.0, "retries": 0, "seconds": 64.7, "max_memory_gb": 17.44, "ce": 0.43453878158470616, "kl": 0.4365038828109391, "context": 0.6527907423733268}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.14775905013084412, "loss_scale": 4.0, "retries": 0, "seconds": 65.1, "max_memory_gb": 17.44, "ce": 0.009060713062353898, "kl": 0.019024658380658366, "context": 0.01857304225268308}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.12054844945669174, "loss_scale": 4.0, "retries": 0, "seconds": 65.5, "max_memory_gb": 17.44, "ce": 0.014305992342997342, "kl": 0.0043464515038067475, "context": 0.016479217680171132}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.26041179895401, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 17.44, "ce": 0.05841323744971305, "kl": 0.02325980761088431, "context": 0.07004314288496971}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.36780160665512085, "loss_scale": 4.0, "retries": 0, "seconds": 66.4, "max_memory_gb": 17.44, "ce": 0.029550086765084416, "kl": 0.03179397237181547, "context": 0.04544707291643135}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.6610993146896362, "loss_scale": 4.0, "retries": 0, "seconds": 66.8, "max_memory_gb": 17.44, "ce": 0.009157744701951742, "kl": 0.4712741695384466, "context": 0.2447948312619701}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.49421021342277527, "loss_scale": 4.0, "retries": 0, "seconds": 67.3, "max_memory_gb": 17.44, "ce": 0.013039601792115718, "kl": 0.36418798775230243, "context": 0.19513359008124098}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 0.32275745272636414, "loss_scale": 4.0, "retries": 0, "seconds": 67.7, "max_memory_gb": 17.44, "ce": 0.05691598710836843, "kl": 0.06833901716163382, "context": 0.09108549583470449}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.47943341732025146, "loss_scale": 4.0, "retries": 0, "seconds": 68.1, "max_memory_gb": 17.44, "ce": 0.16189597420452628, "kl": 0.008772267028689384, "context": 0.16628210166527424}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.19050337374210358, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 17.44, "ce": 0.016567753606068436, "kl": 0.007365813882643124, "context": 0.020250660374586005}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.23434270918369293, "loss_scale": 4.0, "retries": 0, "seconds": 69.0, "max_memory_gb": 17.44, "ce": 0.02033697720617056, "kl": 0.029195878640166484, "context": 0.034934915718622506}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.8608682155609131, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 17.44, "ce": 0.21446965765790083, "kl": 0.11723991252074484, "context": 0.27308962476672605}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.5015860795974731, "loss_scale": 4.0, "retries": 0, "seconds": 69.9, "max_memory_gb": 17.44, "ce": 0.0070190281549002975, "kl": 0.0646919715218246, "context": 0.039365015807561576}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.4162678122520447, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 17.44, "ce": 0.05985602951841429, "kl": 0.2420002314975136, "context": 0.1808561374200508}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.6541949510574341, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 17.44, "ce": 0.11182200101029593, "kl": 0.06641335651556801, "context": 0.1450286792969564}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.4594573974609375, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 17.44, "ce": 0.0659897358273156, "kl": 0.0912122153386008, "context": 0.11159584351116791}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.075258493423462, "loss_scale": 4.0, "retries": 0, "seconds": 71.6, "max_memory_gb": 17.44, "ce": 0.21696927316952497, "kl": 0.5818815374514088, "context": 0.5079100439324975}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 0.7456574440002441, "loss_scale": 4.0, "retries": 0, "seconds": 72.0, "max_memory_gb": 17.44, "ce": 0.08760684843582567, "kl": 0.0620378679705027, "context": 0.11862577828287613}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.17111727595329285, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 17.44, "ce": 0.010323645605240017, "kl": 0.021768846330814995, "context": 0.02120806893799454}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.9159077405929565, "loss_scale": 4.0, "retries": 0, "seconds": 72.9, "max_memory_gb": 17.44, "ce": 0.41746362240519375, "kl": 0.09069868677761406, "context": 0.46281296154484153}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.19487303495407104, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 17.44, "ce": 0.02524505346082151, "kl": 0.006156618008390069, "context": 0.028323363047093153}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.180902361869812, "loss_scale": 4.0, "retries": 0, "seconds": 73.7, "max_memory_gb": 17.44, "ce": 0.03457109333248809, "kl": 0.004197972040856257, "context": 0.036670078116003424}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.248425006866455, "loss_scale": 4.0, "retries": 0, "seconds": 74.2, "max_memory_gb": 17.44, "ce": 0.06932701205369085, "kl": 1.1409137174487114, "context": 0.6397838816046715}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.042603496462106705, "loss_scale": 4.0, "retries": 0, "seconds": 74.6, "max_memory_gb": 17.44, "ce": 0.0030363374098669738, "kl": 0.0006252676330404938, "context": 0.0033489711640868336}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.7066062092781067, "loss_scale": 4.0, "retries": 0, "seconds": 75.0, "max_memory_gb": 17.44, "ce": 0.03169715098920278, "kl": 0.23517959972377867, "context": 0.14928695454727858}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.17722022533416748, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 17.44, "ce": 0.018491230788640678, "kl": 0.01706780877429992, "context": 0.02702513540862128}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.2589870095252991, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 17.44, "ce": 0.04960518382722512, "kl": 0.01926867212750949, "context": 0.05923952045850456}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.2125617116689682, "loss_scale": 4.0, "retries": 0, "seconds": 76.3, "max_memory_gb": 17.44, "ce": 0.033515953211463057, "kl": 0.020001492689516454, "context": 0.04351669974857941}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 1.4802006483078003, "loss_scale": 4.0, "retries": 0, "seconds": 76.8, "max_memory_gb": 17.44, "ce": 0.12013053707778454, "kl": 0.07493369374424219, "context": 0.15759738441556692}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.08370137959718704, "loss_scale": 4.0, "retries": 0, "seconds": 77.2, "max_memory_gb": 17.44, "ce": 0.010495648690266535, "kl": 0.00220549815276172, "context": 0.011598397832131013}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.7317657470703125, "loss_scale": 4.0, "retries": 0, "seconds": 77.6, "max_memory_gb": 17.44, "ce": 0.03860377741511911, "kl": 0.5573228736539022, "context": 0.317265212390339}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.527668833732605, "loss_scale": 4.0, "retries": 0, "seconds": 78.0, "max_memory_gb": 17.44, "ce": 0.06103824399178848, "kl": 0.04334524666774087, "context": 0.0827108679804951}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 1.0000944137573242, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 17.44, "ce": 0.41815048691933043, "kl": 0.0131310685792414, "context": 0.4247160154336598}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.5152769684791565, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 17.44, "ce": 0.052722403954248875, "kl": 0.03363932605134323, "context": 0.06954206674708985}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.28051501512527466, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 17.44, "ce": 0.00949212076375261, "kl": 0.12885896637453698, "context": 0.0739216028014198}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.5531642436981201, "loss_scale": 4.0, "retries": 0, "seconds": 79.7, "max_memory_gb": 17.44, "ce": 0.03319926781114191, "kl": 0.2634169900629786, "context": 0.1649077633046545}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.05661175027489662, "loss_scale": 4.0, "retries": 0, "seconds": 80.2, "max_memory_gb": 17.44, "ce": 0.005179335435968824, "kl": 0.0011161453257955145, "context": 0.005737408238928765}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.07351866364479065, "loss_scale": 4.0, "retries": 0, "seconds": 80.6, "max_memory_gb": 17.44, "ce": 0.004764364304719493, "kl": 0.005951668354100548, "context": 0.007740198343526572}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.18691422045230865, "loss_scale": 4.0, "retries": 0, "seconds": 81.0, "max_memory_gb": 17.44, "ce": 0.015058443532325327, "kl": 0.07531367841147585, "context": 0.052715283760335296}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.520707130432129, "loss_scale": 4.0, "retries": 0, "seconds": 81.5, "max_memory_gb": 17.44, "ce": 0.07788864079338964, "kl": 0.1503932776322472, "context": 0.15308527827437501}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.41876477003097534, "loss_scale": 4.0, "retries": 0, "seconds": 81.9, "max_memory_gb": 17.44, "ce": 0.06944575277157128, "kl": 0.4812805590336211, "context": 0.31008602562360466}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.203551173210144, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 17.44, "ce": 0.2245698753540637, "kl": 0.01599428779900336, "context": 0.23256702601793222}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.8283036947250366, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 17.44, "ce": 0.2637939705164172, "kl": 0.27883846129407175, "context": 0.4032132007414475}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.11457131057977676, "loss_scale": 4.0, "retries": 0, "seconds": 83.2, "max_memory_gb": 17.44, "ce": 0.005814751726575196, "kl": 0.027241703821346164, "context": 0.01943560387007892}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 2.0050671100616455, "loss_scale": 4.0, "retries": 0, "seconds": 83.6, "max_memory_gb": 17.44, "ce": 0.2972669439332094, "kl": 0.21998150852596154, "context": 0.4072576828766614}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 0.9663605093955994, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 17.44, "ce": 0.4491837488603778, "kl": 0.07109053510066587, "context": 0.48472900525666773}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.0800260603427887, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 17.44, "ce": 0.0074702304700622335, "kl": 0.0029333267848414835, "context": 0.008936894009821117}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.5466418266296387, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 17.44, "ce": 0.0650794118992053, "kl": 0.06081318186079443, "context": 0.0954860026249662}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.1674991399049759, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 17.44, "ce": 0.016751740564359352, "kl": 0.002962971462693531, "context": 0.01823322643758729}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 0.6412990093231201, "loss_scale": 4.0, "retries": 0, "seconds": 85.8, "max_memory_gb": 17.44, "ce": 0.0035091330792056397, "kl": 0.4053885666737642, "context": 0.20620341948233545}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.885278582572937, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 17.44, "ce": 0.05178657631040551, "kl": 0.6882797123917044, "context": 0.3959264423174318}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.2917458117008209, "loss_scale": 4.0, "retries": 0, "seconds": 86.7, "max_memory_gb": 17.44, "ce": 0.06871766920085065, "kl": 0.0058837608758040005, "context": 0.07165955239906907}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 0.5847971439361572, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 17.44, "ce": 0.1288332884432748, "kl": 0.09973152847669553, "context": 0.17869905184488744}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.7401729226112366, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 17.44, "ce": 0.25229497451800853, "kl": 0.0337247340939939, "context": 0.26915733818896115}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.7894347906112671, "loss_scale": 4.0, "retries": 0, "seconds": 88.0, "max_memory_gb": 17.44, "ce": 0.07442833195091225, "kl": 0.10747541533783078, "context": 0.1281660395907238}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.04209136217832565, "loss_scale": 4.0, "retries": 0, "seconds": 88.4, "max_memory_gb": 17.44, "ce": 0.003090486214205157, "kl": 0.00444782794738785, "context": 0.005314400375937112}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.2548060715198517, "loss_scale": 4.0, "retries": 0, "seconds": 88.8, "max_memory_gb": 17.44, "ce": 0.01634390295657795, "kl": 0.018208438530564308, "context": 0.025448121887166053}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.2685844302177429, "loss_scale": 4.0, "retries": 0, "seconds": 89.2, "max_memory_gb": 17.44, "ce": 0.044059862673748285, "kl": 0.010874023078940809, "context": 0.04949687194311991}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.5298177599906921, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 17.44, "ce": 0.005732922232709825, "kl": 0.48565543360018637, "context": 0.24856063607148826}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.14007796347141266, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 17.44, "ce": 0.012079191968950909, "kl": 0.003837137782738864, "context": 0.01399776036851108}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.27375704050064087, "loss_scale": 4.0, "retries": 0, "seconds": 90.5, "max_memory_gb": 17.44, "ce": 0.076754204579629, "kl": 0.03649671084713191, "context": 0.0950025599449873}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.060229431837797165, "loss_scale": 4.0, "retries": 0, "seconds": 91.0, "max_memory_gb": 17.44, "ce": 0.006746736151399091, "kl": 0.0025243611744372174, "context": 0.008008916571270674}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 1.1093343496322632, "loss_scale": 4.0, "retries": 0, "seconds": 91.4, "max_memory_gb": 17.44, "ce": 0.20636107865720987, "kl": 0.08581813272758154, "context": 0.24927014153217897}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 0.9411027431488037, "loss_scale": 4.0, "retries": 0, "seconds": 91.8, "max_memory_gb": 17.44, "ce": 0.4690406000299845, "kl": 0.13642690098640742, "context": 0.5372540621319786}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.5314802527427673, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 17.44, "ce": 0.173253926448524, "kl": 0.06779424662818201, "context": 0.20715105452109128}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 1.8546481132507324, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 17.44, "ce": 0.09362812910694629, "kl": 0.9933095221058466, "context": 0.5902828923426569}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.720806896686554, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 17.44, "ce": 0.2395758143684361, "kl": 0.043788866147224326, "context": 0.2614702539285645}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 1.0639702081680298, "loss_scale": 4.0, "retries": 0, "seconds": 93.6, "max_memory_gb": 17.44, "ce": 0.41170463801245205, "kl": 0.12135869672056288, "context": 0.4723839989164844}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.7505913972854614, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 17.44, "ce": 0.5811073939257767, "kl": 0.025278120606230914, "context": 0.5937464571325108}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.2282540649175644, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 17.44, "ce": 0.01245240870048292, "kl": 0.08342059279311798, "context": 0.054162703279871494}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.05403276905417442, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 17.44, "ce": 0.004129835491767153, "kl": 0.003164582871249877, "context": 0.005712127021979541}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.6711013317108154, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 17.44, "ce": 0.1758000128902495, "kl": 0.2621798515319824, "context": 0.30688993260264397}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.3883483111858368, "loss_scale": 4.0, "retries": 0, "seconds": 95.8, "max_memory_gb": 17.44, "ce": 0.013117191774654202, "kl": 0.11797631619265303, "context": 0.0721053495653905}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.27148377895355225, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 17.44, "ce": 0.05005647960933857, "kl": 0.0045545400062110275, "context": 0.0523337482009083}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.5294294953346252, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 17.44, "ce": 0.006631538533838466, "kl": 0.4163482530420879, "context": 0.21480565797537565}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.5469995141029358, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 17.44, "ce": 0.05788625404238701, "kl": 0.05812537183828681, "context": 0.08694894163636491}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 3.687875747680664, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 17.44, "ce": 0.24216359097044915, "kl": 0.024172105277102673, "context": 0.2542496373062022}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 1.4247239828109741, "loss_scale": 4.0, "retries": 0, "seconds": 97.9, "max_memory_gb": 17.44, "ce": 0.069984003494028, "kl": 0.16778120887101977, "context": 0.15387460181955248}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.36880454421043396, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 17.44, "ce": 0.006192745349835604, "kl": 0.14612657608449808, "context": 0.07925603384501301}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.6794047355651855, "loss_scale": 4.0, "retries": 0, "seconds": 98.8, "max_memory_gb": 17.44, "ce": 0.0391638966393657, "kl": 0.028843454289017245, "context": 0.053585624729748815}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.3253277540206909, "loss_scale": 4.0, "retries": 0, "seconds": 99.2, "max_memory_gb": 17.44, "ce": 0.032869173170183785, "kl": 0.013139800947783442, "context": 0.03943907179927919}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.5358420014381409, "loss_scale": 4.0, "retries": 0, "seconds": 99.6, "max_memory_gb": 17.44, "ce": 0.009110593920922838, "kl": 0.21209347067633644, "context": 0.11515733192209154}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 1.7915294170379639, "loss_scale": 4.0, "retries": 0, "seconds": 100.1, "max_memory_gb": 17.44, "ce": 0.10304591790190898, "kl": 0.06915788346350382, "context": 0.13762486216728576}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.824634075164795, "loss_scale": 4.0, "retries": 0, "seconds": 100.5, "max_memory_gb": 17.44, "ce": 0.5261201687389985, "kl": 0.28143306635320187, "context": 0.6668367111124098}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 1.1312415599822998, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 17.44, "ce": 0.0698705087415874, "kl": 0.1662152048374992, "context": 0.15297811396885663}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.2568143308162689, "loss_scale": 4.0, "retries": 0, "seconds": 101.3, "max_memory_gb": 17.44, "ce": 0.03950274342787452, "kl": 0.007589619708596729, "context": 0.04329755186336115}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.3629458248615265, "loss_scale": 4.0, "retries": 0, "seconds": 101.8, "max_memory_gb": 17.44, "ce": 0.01434144377708435, "kl": 0.09552034339867532, "context": 0.062101616989821196}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.22956934571266174, "loss_scale": 4.0, "retries": 0, "seconds": 102.2, "max_memory_gb": 17.44, "ce": 0.01293437328422442, "kl": 0.012563310730911326, "context": 0.01921602856600657}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.6706289052963257, "loss_scale": 4.0, "retries": 0, "seconds": 102.6, "max_memory_gb": 17.44, "ce": 0.3650902077788487, "kl": 0.021604967630992178, "context": 0.3758926876762416}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.13671788573265076, "loss_scale": 4.0, "retries": 0, "seconds": 103.1, "max_memory_gb": 17.44, "ce": 0.013609801622806117, "kl": 0.020432003191672266, "context": 0.02382580313133076}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.3887299597263336, "loss_scale": 4.0, "retries": 0, "seconds": 103.5, "max_memory_gb": 17.44, "ce": 0.08051875146338716, "kl": 0.024658569807286312, "context": 0.09284803568152711}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.22685550153255463, "loss_scale": 4.0, "retries": 0, "seconds": 103.9, "max_memory_gb": 17.44, "ce": 0.0278759922075551, "kl": 0.006583646181752556, "context": 0.03116781475546304}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.3455101251602173, "loss_scale": 4.0, "retries": 0, "seconds": 104.3, "max_memory_gb": 17.44, "ce": 0.123907262342982, "kl": 0.02159249763644766, "context": 0.1347035076469183}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.4530338943004608, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 17.44, "ce": 0.10762354370672256, "kl": 0.02973196018137969, "context": 0.12248952657682821}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 1.423787236213684, "loss_scale": 4.0, "retries": 0, "seconds": 105.2, "max_memory_gb": 17.44, "ce": 0.22502508186153136, "kl": 0.03015421591408085, "context": 0.24010218895273283}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.15139326453208923, "loss_scale": 4.0, "retries": 0, "seconds": 105.6, "max_memory_gb": 17.44, "ce": 0.009298542921897024, "kl": 0.04830929993204336, "context": 0.033453192620072514}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.1091562807559967, "loss_scale": 4.0, "retries": 0, "seconds": 106.1, "max_memory_gb": 17.44, "ce": 0.00955700408667326, "kl": 0.005369872262235731, "context": 0.012241940246894956}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.5413085222244263, "loss_scale": 4.0, "retries": 0, "seconds": 106.5, "max_memory_gb": 17.44, "ce": 0.060302945305011235, "kl": 0.14859933517072932, "context": 0.1346026128885569}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 1.0055218935012817, "loss_scale": 4.0, "retries": 0, "seconds": 106.9, "max_memory_gb": 17.44, "ce": 0.3816550327464938, "kl": 0.07939321978483349, "context": 0.4213516318704933}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.026745416224002838, "loss_scale": 4.0, "retries": 0, "seconds": 107.4, "max_memory_gb": 17.44, "ce": 0.0022504933731397614, "kl": 0.000935023035708582, "context": 0.002718004834605381}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.5240446925163269, "loss_scale": 4.0, "retries": 0, "seconds": 107.8, "max_memory_gb": 17.44, "ce": 0.009204930916894227, "kl": 0.46004432132758666, "context": 0.23922708712052554}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.2734729051589966, "loss_scale": 4.0, "retries": 0, "seconds": 108.2, "max_memory_gb": 17.44, "ce": 0.03674278501421213, "kl": 0.015257685910910368, "context": 0.04437162773683667}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.5165212154388428, "loss_scale": 4.0, "retries": 0, "seconds": 108.6, "max_memory_gb": 17.44, "ce": 0.061084243236109614, "kl": 0.059503668162506074, "context": 0.09083607792854309}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.09728644043207169, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 17.44, "ce": 0.010132987168617547, "kl": 0.004580398577672895, "context": 0.012423186388332397}
|
runs/single_internvl/context/L25_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L25_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.25.self_attn.k_proj",
|
| 35 |
+
"model.language_model.layers.25.self_attn.v_proj",
|
| 36 |
+
"model.language_model.layers.25.self_attn.o_proj",
|
| 37 |
+
"model.language_model.layers.25.self_attn.q_proj",
|
| 38 |
+
"model.language_model.layers.25.mlp.gate_proj",
|
| 39 |
+
"model.language_model.layers.25.mlp.up_proj",
|
| 40 |
+
"model.language_model.layers.25.mlp.down_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L25_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 25,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L25_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.5883803963661194, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 17.01, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.6018528938293457, "loss_scale": 4.0, "retries": 0, "seconds": 1.7, "max_memory_gb": 17.02, "ce": 0.6404333652244532, "kl": 0.4672744294241511, "context": 0.8740705911695841}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 2.07371187210083, "loss_scale": 4.0, "retries": 0, "seconds": 2.2, "max_memory_gb": 17.02, "ce": 1.0982282981276512, "kl": 0.06708946172147989, "context": 1.1317730098962784}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.2216459512710571, "loss_scale": 4.0, "retries": 0, "seconds": 2.6, "max_memory_gb": 17.02, "ce": 0.4485335658682743, "kl": 0.4298224066997136, "context": 0.6634447768738028}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.012583052739501, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 17.02, "ce": 0.0091884375524387, "kl": 0.008232076364947716, "context": 0.013304475676704897}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.5756418704986572, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 17.02, "ce": 0.28596588954678737, "kl": 0.07419521610427182, "context": 0.3230635010986589}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 1.8725969791412354, "loss_scale": 4.0, "retries": 0, "seconds": 4.0, "max_memory_gb": 17.02, "ce": 0.8147359094618878, "kl": 0.12323955487056537, "context": 0.8763556851517933}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 2.097053289413452, "loss_scale": 4.0, "retries": 0, "seconds": 4.4, "max_memory_gb": 17.03, "ce": 0.7779182773083448, "kl": 0.2881745472550392, "context": 0.9220055602490902}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 1.5860474109649658, "loss_scale": 4.0, "retries": 0, "seconds": 4.9, "max_memory_gb": 17.03, "ce": 0.5088652726262808, "kl": 0.13699499261565506, "context": 0.5773627609014511}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.27133557200431824, "loss_scale": 4.0, "retries": 0, "seconds": 5.3, "max_memory_gb": 17.03, "ce": 0.049309223009913694, "kl": 0.04537611495470628, "context": 0.07199727860279381}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 1.7590290307998657, "loss_scale": 4.0, "retries": 0, "seconds": 5.7, "max_memory_gb": 17.03, "ce": 0.6654312550235773, "kl": 0.0050657084905196825, "context": 0.6679641216105665}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.1741905212402344, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 17.03, "ce": 0.44441517303948785, "kl": 0.002710095401653234, "context": 0.44577022876546835}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.145578145980835, "loss_scale": 4.0, "retries": 0, "seconds": 6.5, "max_memory_gb": 17.03, "ce": 0.39925142271385994, "kl": 0.0187787204158667, "context": 0.4086407827453513}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 2.44775652885437, "loss_scale": 4.0, "retries": 0, "seconds": 6.9, "max_memory_gb": 17.03, "ce": 0.6799478977918625, "kl": 0.9486082820221782, "context": 1.1542520076036453}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 2.014955520629883, "loss_scale": 4.0, "retries": 0, "seconds": 7.3, "max_memory_gb": 17.03, "ce": 0.4455733961704027, "kl": 0.4663476336850181, "context": 0.6787472111502666}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.8571985363960266, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 17.03, "ce": 0.17199777761561563, "kl": 0.0008986846428342687, "context": 0.17244711599778384}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.9717519283294678, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 17.03, "ce": 0.2769453158834949, "kl": 0.825930904422421, "context": 0.6899107478384394}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 2.7226998805999756, "loss_scale": 4.0, "retries": 0, "seconds": 8.6, "max_memory_gb": 17.03, "ce": 0.3590149078518152, "kl": 0.21990333545545582, "context": 0.4689665929181501}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 4.263156414031982, "loss_scale": 4.0, "retries": 0, "seconds": 9.1, "max_memory_gb": 17.03, "ce": 0.4625355862081051, "kl": 0.17147006560117006, "context": 0.5482706166803837}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.03494376316666603, "loss_scale": 4.0, "retries": 0, "seconds": 9.5, "max_memory_gb": 17.03, "ce": 0.002319384506336064, "kl": 0.021213895099208457, "context": 0.012926331626658794}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 2.152287483215332, "loss_scale": 4.0, "retries": 0, "seconds": 9.9, "max_memory_gb": 17.03, "ce": 0.18082309379860817, "kl": 0.036486873455032764, "context": 0.1990665305256698}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 1.8844300508499146, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 17.03, "ce": 0.18615324969869107, "kl": 0.033913985134859104, "context": 0.20311024208785966}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.0362184047698975, "loss_scale": 4.0, "retries": 0, "seconds": 10.8, "max_memory_gb": 17.03, "ce": 0.35921427206176304, "kl": 0.970214748394028, "context": 0.8443216398700315}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 2.0983030796051025, "loss_scale": 4.0, "retries": 0, "seconds": 11.2, "max_memory_gb": 17.03, "ce": 0.09754919586703181, "kl": 0.22379334713332355, "context": 0.20944586396217346}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.9635854959487915, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 17.03, "ce": 0.42098192686898983, "kl": 0.16148384660482407, "context": 0.50172383710742}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 1.2059119939804077, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 17.03, "ce": 0.0757646500132978, "kl": 0.6657166795339435, "context": 0.4086229810491204}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 1.5041186809539795, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 17.03, "ce": 0.13551380793796852, "kl": 0.36740465975344705, "context": 0.31921613297163276}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.6702799797058105, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 17.03, "ce": 0.4857026836834848, "kl": 0.9502642345614731, "context": 0.9608347956091166}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 0.7582566738128662, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 17.03, "ce": 0.08213966991752386, "kl": 0.37779853062238544, "context": 0.27103893365710974}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 0.8398661613464355, "loss_scale": 4.0, "retries": 0, "seconds": 13.7, "max_memory_gb": 17.03, "ce": 0.06107084699215193, "kl": 0.002009998933317547, "context": 0.062075847299638554}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 0.9056372046470642, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 17.03, "ce": 0.19285876769572496, "kl": 0.11618714593350887, "context": 0.2509523360058665}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 0.5165380239486694, "loss_scale": 4.0, "retries": 0, "seconds": 14.6, "max_memory_gb": 17.03, "ce": 0.018824590952135623, "kl": 0.016624662865069695, "context": 0.027136922348290682}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.46021685004234314, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 17.03, "ce": 0.053450383711606264, "kl": 0.027688548434525728, "context": 0.0672946572303772}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 0.5405283570289612, "loss_scale": 4.0, "retries": 0, "seconds": 15.5, "max_memory_gb": 17.03, "ce": 0.0036093350317969453, "kl": 0.7894694572219123, "context": 0.39834406021509494}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 0.09846634417772293, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 17.03, "ce": 0.003590363048715517, "kl": 0.0029023185270489193, "context": 0.005041522250394337}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 0.4762556552886963, "loss_scale": 4.0, "retries": 0, "seconds": 16.4, "max_memory_gb": 17.03, "ce": 0.0904133084404748, "kl": 0.053309373288357165, "context": 0.1170679933129577}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 0.6668826341629028, "loss_scale": 4.0, "retries": 0, "seconds": 16.8, "max_memory_gb": 17.03, "ce": 0.12812985161144752, "kl": 0.7802973578218371, "context": 0.5182785344077274}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 0.3699702024459839, "loss_scale": 4.0, "retries": 0, "seconds": 17.3, "max_memory_gb": 17.03, "ce": 0.019534793682396412, "kl": 0.46295720801572315, "context": 0.25101339165121317}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.32537129521369934, "loss_scale": 4.0, "retries": 0, "seconds": 17.7, "max_memory_gb": 17.03, "ce": 0.11225691212166566, "kl": 0.009604482431313954, "context": 0.11705915149650536}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.059092652052640915, "loss_scale": 4.0, "retries": 0, "seconds": 18.2, "max_memory_gb": 17.03, "ce": 0.005978546178084798, "kl": 0.00878378565539606, "context": 0.010370439063990489}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 0.6495232582092285, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 17.03, "ce": 0.3648601232562214, "kl": 0.11942991754040122, "context": 0.42457508598454297}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.020746495574712753, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 17.03, "ce": 0.002792502440570388, "kl": 0.0005278027238091454, "context": 0.0030564037733711302}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.022785823792219162, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 17.03, "ce": 0.0019012618213309906, "kl": 0.0026951846230076626, "context": 0.0032488540455233306}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 1.8987876176834106, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 17.03, "ce": 0.5095766170343268, "kl": 1.2970087654393865, "context": 1.1580809855367988}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 0.9252567291259766, "loss_scale": 4.0, "retries": 0, "seconds": 20.4, "max_memory_gb": 17.03, "ce": 0.8199016902362928, "kl": 0.27302640629932284, "context": 0.9564148568315431}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.3026173710823059, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 17.03, "ce": 0.07065047482319642, "kl": 0.26014358559041284, "context": 0.20072227058699355}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.05316576734185219, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 17.03, "ce": 0.004817881068447605, "kl": 0.014133454591501504, "context": 0.011884608014952391}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.0166016835719347, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 17.03, "ce": 0.0009886067709885538, "kl": 0.002823416686624114, "context": 0.002400315148406662}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.5973345041275024, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 17.03, "ce": 0.012892144892248325, "kl": 0.3017775985586013, "context": 0.16378094900210272}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 0.6920309066772461, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 17.03, "ce": 0.09822942760001752, "kl": 0.00798040757945273, "context": 0.10221963142976165}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.3096792697906494, "loss_scale": 4.0, "retries": 0, "seconds": 23.2, "max_memory_gb": 17.03, "ce": 0.47753341624411405, "kl": 0.02411270804407195, "context": 0.48958976817084476}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.3982016444206238, "loss_scale": 4.0, "retries": 0, "seconds": 23.6, "max_memory_gb": 17.03, "ce": 0.022759321378543973, "kl": 0.1893149031093344, "context": 0.11741677485406399}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.7930741906166077, "loss_scale": 4.0, "retries": 0, "seconds": 24.0, "max_memory_gb": 17.03, "ce": 0.12042401020880789, "kl": 0.39120370615273714, "context": 0.31602586060762405}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5449915528297424, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 17.03, "ce": 0.04203539360605646, "kl": 0.3255150227123522, "context": 0.20479290471121203}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.19594796001911163, "loss_scale": 4.0, "retries": 0, "seconds": 25.0, "max_memory_gb": 17.03, "ce": 0.021414118236862123, "kl": 0.014417447110190551, "context": 0.02862284248112701}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.3875943720340729, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 17.03, "ce": 0.02237821533344686, "kl": 0.21401527954731137, "context": 0.12938585225492716}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.16711507737636566, "loss_scale": 4.0, "retries": 0, "seconds": 25.9, "max_memory_gb": 17.03, "ce": 0.008422325270657893, "kl": 0.01662750477771624, "context": 0.016736078578105662}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.028492286801338196, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 17.03, "ce": 0.0016132289383676834, "kl": 0.0008999560209304036, "context": 0.002063206884486135}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.2726878225803375, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 17.03, "ce": 0.03286314845900051, "kl": 0.0623879442573525, "context": 0.06405711872503161}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.5190157890319824, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 17.03, "ce": 0.07814182684523985, "kl": 0.0654287159559317, "context": 0.11085618479410186}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 0.9531986713409424, "loss_scale": 4.0, "retries": 0, "seconds": 27.7, "max_memory_gb": 17.03, "ce": 0.02158917547785677, "kl": 0.7772277932162979, "context": 0.41020306485006586}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.10969998687505722, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 17.03, "ce": 0.02016347204335034, "kl": 0.009075642519746907, "context": 0.024701293325051665}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.5174950361251831, "loss_scale": 4.0, "retries": 0, "seconds": 28.6, "max_memory_gb": 17.03, "ce": 0.05134840466780588, "kl": 0.11910521311801858, "context": 0.11090101022273302}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.18598923087120056, "loss_scale": 4.0, "retries": 0, "seconds": 29.0, "max_memory_gb": 17.03, "ce": 0.01995462510967627, "kl": 0.016171357448911294, "context": 0.028040303324814886}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.5916532874107361, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 17.03, "ce": 0.15141260949894786, "kl": 0.1517234374769032, "context": 0.22727432753890753}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.840674638748169, "loss_scale": 4.0, "retries": 0, "seconds": 30.0, "max_memory_gb": 17.03, "ce": 0.15794898173521688, "kl": 0.06900354827098454, "context": 0.1924507596032754}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.06331270933151245, "loss_scale": 4.0, "retries": 0, "seconds": 30.4, "max_memory_gb": 17.03, "ce": 0.0058982543559977785, "kl": 0.007234948658151552, "context": 0.009515728859696537}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.09075813740491867, "loss_scale": 4.0, "retries": 0, "seconds": 30.9, "max_memory_gb": 17.03, "ce": 0.005249471274510142, "kl": 0.03186450110842998, "context": 0.021181722499022726}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.6685600280761719, "loss_scale": 4.0, "retries": 0, "seconds": 31.4, "max_memory_gb": 17.03, "ce": 0.3244938444986474, "kl": 0.026773757024784572, "context": 0.33788071089657024}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.01318266335874796, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 17.03, "ce": 0.0017261885077459738, "kl": 0.00016616403718217043, "context": 0.00180927048495505}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.11517094820737839, "loss_scale": 4.0, "retries": 0, "seconds": 32.3, "max_memory_gb": 17.03, "ce": 0.010707456231102697, "kl": 0.010895552870351821, "context": 0.016155232908204198}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.42940908670425415, "loss_scale": 4.0, "retries": 0, "seconds": 32.7, "max_memory_gb": 17.03, "ce": 0.01719764742301777, "kl": 0.39265445206547156, "context": 0.21352487371768802}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 0.9861831068992615, "loss_scale": 4.0, "retries": 0, "seconds": 33.2, "max_memory_gb": 17.03, "ce": 0.2911085589657887, "kl": 0.0838282144650293, "context": 0.3330226680773194}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.17128324508666992, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 17.03, "ce": 0.05912718737090472, "kl": 0.11091211012899294, "context": 0.11458324615705351}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 0.7294386625289917, "loss_scale": 4.0, "retries": 0, "seconds": 34.1, "max_memory_gb": 17.03, "ce": 0.4628721833432792, "kl": 0.9788778331130743, "context": 0.9523110999725759}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.0860651731491089, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 17.03, "ce": 0.011887717846548185, "kl": 1.6225410388142336, "context": 0.8231582619482651}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.1326388269662857, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 17.03, "ce": 0.010985344764776528, "kl": 0.011062886529543903, "context": 0.016516787800355814}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.44485628604888916, "loss_scale": 4.0, "retries": 0, "seconds": 35.5, "max_memory_gb": 17.03, "ce": 0.00188755095587112, "kl": 0.11202920452342369, "context": 0.057902154338080436}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.1760922074317932, "loss_scale": 4.0, "retries": 0, "seconds": 35.9, "max_memory_gb": 17.03, "ce": 0.016876722447705106, "kl": 0.02667155336439464, "context": 0.030212498242690344}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.3171546161174774, "loss_scale": 4.0, "retries": 0, "seconds": 36.3, "max_memory_gb": 17.03, "ce": 0.004161332311923616, "kl": 0.06979448854917791, "context": 0.039058576001480105}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.492328941822052, "loss_scale": 4.0, "retries": 0, "seconds": 36.8, "max_memory_gb": 17.03, "ce": 0.08422820980194956, "kl": 0.0447734919434879, "context": 0.10661495721433312}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.04819710925221443, "loss_scale": 4.0, "retries": 0, "seconds": 37.2, "max_memory_gb": 17.03, "ce": 0.006598527077585459, "kl": 0.001668263808824122, "context": 0.007432659040205181}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 1.3910094499588013, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 17.03, "ce": 0.13659551524324343, "kl": 0.6275358727871208, "context": 0.45036343904212117}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.56479412317276, "loss_scale": 4.0, "retries": 0, "seconds": 38.2, "max_memory_gb": 17.03, "ce": 0.037734911948973604, "kl": 0.04911188094411045, "context": 0.06229085201721318}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 0.22084961831569672, "loss_scale": 4.0, "retries": 0, "seconds": 38.6, "max_memory_gb": 17.03, "ce": 0.007968294897636952, "kl": 0.00545471440161549, "context": 0.010695652100366715}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.4515538215637207, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 17.03, "ce": 0.11365919938543811, "kl": 0.3424575293902308, "context": 0.2848879718221724}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.9042577147483826, "loss_scale": 4.0, "retries": 0, "seconds": 39.5, "max_memory_gb": 17.03, "ce": 0.007699473819229752, "kl": 0.6548030693084002, "context": 0.33510101214051247}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.1283699870109558, "loss_scale": 4.0, "retries": 0, "seconds": 40.0, "max_memory_gb": 17.03, "ce": 0.010951541540634935, "kl": 0.014890067492530079, "context": 0.018396575766928436}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.08313681930303574, "loss_scale": 4.0, "retries": 0, "seconds": 40.4, "max_memory_gb": 17.03, "ce": 0.014786536752126267, "kl": 0.011726911579899024, "context": 0.02064999149115465}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.0570790097117424, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 17.03, "ce": 0.007018532312940806, "kl": 0.002639731450472027, "context": 0.00833839809638448}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 1.4531819820404053, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 17.03, "ce": 0.40020177182304906, "kl": 0.3458884698775364, "context": 0.5731460137467366}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.944313645362854, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 17.03, "ce": 0.21000844473746838, "kl": 0.38547047879501406, "context": 0.40274368419341045}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 0.9694845080375671, "loss_scale": 4.0, "retries": 0, "seconds": 42.2, "max_memory_gb": 17.03, "ce": 0.06704297458054498, "kl": 0.3743771935987752, "context": 0.2542315673199482}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.05013906955719, "loss_scale": 4.0, "retries": 0, "seconds": 42.7, "max_memory_gb": 17.03, "ce": 0.06471366251935251, "kl": 0.03785419254563749, "context": 0.08364076240104623}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.5430810451507568, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 17.03, "ce": 0.40936883155518444, "kl": 0.040847307071089745, "context": 0.4297924926504493}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.0396592617034912, "loss_scale": 4.0, "retries": 0, "seconds": 43.6, "max_memory_gb": 17.03, "ce": 0.37167296788538806, "kl": 0.006465722606662894, "context": 0.37490584058105014}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 0.9739329814910889, "loss_scale": 4.0, "retries": 0, "seconds": 44.1, "max_memory_gb": 17.03, "ce": 0.014180020494677592, "kl": 0.1091620335355401, "context": 0.06876103766262531}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.282056987285614, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 17.03, "ce": 0.0036496852844720706, "kl": 0.0873579473081918, "context": 0.047328658933111}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.07208988815546036, "loss_scale": 4.0, "retries": 0, "seconds": 45.0, "max_memory_gb": 17.03, "ce": 0.01028372865766869, "kl": 0.006813220421463484, "context": 0.013690338399101165}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 1.021178126335144, "loss_scale": 4.0, "retries": 0, "seconds": 45.4, "max_memory_gb": 17.03, "ce": 0.028298218559939414, "kl": 0.44205462044192245, "context": 0.24932552941027097}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.25038596987724304, "loss_scale": 4.0, "retries": 0, "seconds": 45.9, "max_memory_gb": 17.03, "ce": 0.003833352995570749, "kl": 0.08867719446425326, "context": 0.04817195073701441}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.45333877205848694, "loss_scale": 4.0, "retries": 0, "seconds": 46.4, "max_memory_gb": 17.03, "ce": 0.05208042968297377, "kl": 0.02731550403404981, "context": 0.06573818216565996}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.11849364638328552, "loss_scale": 4.0, "retries": 0, "seconds": 46.8, "max_memory_gb": 17.03, "ce": 0.010483227506483672, "kl": 0.013883957879443187, "context": 0.017425206387997605}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.06797442585229874, "loss_scale": 4.0, "retries": 0, "seconds": 47.2, "max_memory_gb": 17.03, "ce": 0.004509986229095375, "kl": 0.007347288940763974, "context": 0.00818363036523806}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.49751296639442444, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 17.03, "ce": 0.03095902723725885, "kl": 0.3512296425178647, "context": 0.20657384453807026}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.06552024930715561, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 17.03, "ce": 0.00574508604768198, "kl": 0.002659450219653081, "context": 0.007074811335769482}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 0.42443516850471497, "loss_scale": 4.0, "retries": 0, "seconds": 48.6, "max_memory_gb": 17.03, "ce": 0.03529062820598483, "kl": 0.18507945343844767, "context": 0.12783035865868442}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 2.0436384677886963, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 17.03, "ce": 0.24757855100688175, "kl": 0.03148094756750197, "context": 0.2633190097476472}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 1.1742371320724487, "loss_scale": 4.0, "retries": 0, "seconds": 49.5, "max_memory_gb": 17.03, "ce": 0.12017881235806271, "kl": 0.3695967085659504, "context": 0.3049771748483181}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.0979045629501343, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 17.03, "ce": 0.03651611343957484, "kl": 0.10795420443173498, "context": 0.09049321501515806}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.02202354371547699, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 17.03, "ce": 0.001631044497116818, "kl": 0.0015675078921049135, "context": 0.002414798436802812}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.09950650483369827, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 17.03, "ce": 0.009248501854017377, "kl": 0.0009517269354546443, "context": 0.009724364877911285}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.547526478767395, "loss_scale": 4.0, "retries": 0, "seconds": 51.3, "max_memory_gb": 17.03, "ce": 0.010414859454613179, "kl": 0.22404723684303463, "context": 0.12243848224170506}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.2829176187515259, "loss_scale": 4.0, "retries": 0, "seconds": 51.8, "max_memory_gb": 17.03, "ce": 0.010290073769283481, "kl": 0.8492824825807475, "context": 0.43493131629657}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.12803243100643158, "loss_scale": 4.0, "retries": 0, "seconds": 52.3, "max_memory_gb": 17.03, "ce": 0.006330150514259003, "kl": 0.028514359622931806, "context": 0.02058733043668326}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 0.7060248255729675, "loss_scale": 4.0, "retries": 0, "seconds": 52.7, "max_memory_gb": 17.03, "ce": 0.05228008536869311, "kl": 0.4870599882636384, "context": 0.29581007576416596}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 0.23245637118816376, "loss_scale": 4.0, "retries": 0, "seconds": 53.2, "max_memory_gb": 17.03, "ce": 0.007272361792274751, "kl": 0.011291879483906087, "context": 0.0129183019453194}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.7925398349761963, "loss_scale": 4.0, "retries": 0, "seconds": 53.6, "max_memory_gb": 17.03, "ce": 0.2854174791100377, "kl": 0.00544866241307318, "context": 0.2881418038268748}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.15989096462726593, "loss_scale": 4.0, "retries": 0, "seconds": 54.1, "max_memory_gb": 17.03, "ce": 0.01509233702745405, "kl": 0.023622276677201626, "context": 0.026903475361905294}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 1.8607845306396484, "loss_scale": 4.0, "retries": 0, "seconds": 54.6, "max_memory_gb": 17.03, "ce": 0.25362767750630155, "kl": 0.1670017147262115, "context": 0.3371285400353372}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.32582011818885803, "loss_scale": 4.0, "retries": 0, "seconds": 55.1, "max_memory_gb": 17.03, "ce": 0.07047239800886018, "kl": 0.040578877395091695, "context": 0.09076183903380297}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 1.1192219257354736, "loss_scale": 4.0, "retries": 0, "seconds": 55.5, "max_memory_gb": 17.03, "ce": 0.19138703450153116, "kl": 0.01662173381919274, "context": 0.1996978996648977}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.3873857259750366, "loss_scale": 4.0, "retries": 0, "seconds": 56.0, "max_memory_gb": 17.03, "ce": 0.039366116165183485, "kl": 0.026533628930337727, "context": 0.05263292888412252}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.5917096138000488, "loss_scale": 4.0, "retries": 0, "seconds": 56.4, "max_memory_gb": 17.03, "ce": 0.7518948805081891, "kl": 0.4963836315146182, "context": 1.0000866627524374}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.3557108938694, "loss_scale": 4.0, "retries": 0, "seconds": 56.9, "max_memory_gb": 17.03, "ce": 0.13643087587843183, "kl": 0.06694549463645671, "context": 0.16990363056538627}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.581441879272461, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 17.03, "ce": 0.757691165432334, "kl": 0.09041343908756971, "context": 0.802897859364748}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.10198111087083817, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 17.03, "ce": 0.009884103994409088, "kl": 0.0015159232170844916, "context": 0.010642065739375539}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 1.7708078622817993, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 17.03, "ce": 0.06705376431636978, "kl": 0.8782407676917501, "context": 0.50617413490545}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.7329121232032776, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 17.03, "ce": 0.11665922921383753, "kl": 0.02876431388722267, "context": 0.13104138663038611}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.39186763763427734, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 17.03, "ce": 0.05048683303175494, "kl": 0.28171021211892366, "context": 0.19134194235084578}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 2.1831634044647217, "loss_scale": 4.0, "retries": 0, "seconds": 59.6, "max_memory_gb": 17.03, "ce": 0.2979839983745478, "kl": 0.32291715347673744, "context": 0.4594425620161928}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 1.01661217212677, "loss_scale": 4.0, "retries": 0, "seconds": 60.1, "max_memory_gb": 17.03, "ce": 0.20938155775365885, "kl": 0.029919255834101932, "context": 0.22434117825469002}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 1.0701940059661865, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 17.03, "ce": 0.4348538512713276, "kl": 0.09802910219877958, "context": 0.4838683931156993}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.4816232919692993, "loss_scale": 4.0, "retries": 0, "seconds": 61.0, "max_memory_gb": 17.03, "ce": 0.11785231171597843, "kl": 0.0026169502093580377, "context": 0.11916078612557612}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.38143593072891235, "loss_scale": 4.0, "retries": 0, "seconds": 61.4, "max_memory_gb": 17.03, "ce": 0.017068182001821697, "kl": 0.13301423390112177, "context": 0.08357529900968075}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 0.3733215928077698, "loss_scale": 4.0, "retries": 0, "seconds": 61.9, "max_memory_gb": 17.03, "ce": 0.027781492477515712, "kl": 0.050107816539821215, "context": 0.052835402137134224}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.43028372526168823, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 17.03, "ce": 0.038209167192690074, "kl": 0.09418106934754178, "context": 0.08529969834489748}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 0.7217217087745667, "loss_scale": 4.0, "retries": 0, "seconds": 62.8, "max_memory_gb": 17.03, "ce": 0.07018623524345458, "kl": 0.10873509990051389, "context": 0.12455378472805023}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.7847253680229187, "loss_scale": 4.0, "retries": 0, "seconds": 63.3, "max_memory_gb": 17.03, "ce": 0.0034984926824108697, "kl": 0.10119933201758613, "context": 0.054098159900604514}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.11256065964698792, "loss_scale": 4.0, "retries": 0, "seconds": 63.7, "max_memory_gb": 17.03, "ce": 0.00865374477871228, "kl": 0.0037992392244632356, "context": 0.010553364336374216}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.06278890371322632, "loss_scale": 4.0, "retries": 0, "seconds": 64.2, "max_memory_gb": 17.03, "ce": 0.006333414858090691, "kl": 0.0051343380473554134, "context": 0.008900583910872228}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.1468425989151001, "loss_scale": 4.0, "retries": 0, "seconds": 64.7, "max_memory_gb": 17.03, "ce": 0.012729324902466033, "kl": 0.008390945615246892, "context": 0.01692479767370969}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.5532841682434082, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 17.03, "ce": 0.33014116615231615, "kl": 0.054901811660784006, "context": 0.35759206553484546}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.6458211541175842, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 17.03, "ce": 0.027421454549767077, "kl": 0.22410322073847055, "context": 0.13947306433692575}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.044459860771894455, "loss_scale": 4.0, "retries": 0, "seconds": 66.1, "max_memory_gb": 17.03, "ce": 0.002290274016559124, "kl": 0.006311983289378986, "context": 0.005446265582577325}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.08699104934930801, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 17.03, "ce": 0.01069367304444313, "kl": 0.006202428268807125, "context": 0.013794886996038258}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.3830406665802002, "loss_scale": 4.0, "retries": 0, "seconds": 67.0, "max_memory_gb": 17.03, "ce": 0.5247288077953272, "kl": 0.39880173426354304, "context": 0.7241296680876985}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.07576215267181396, "loss_scale": 4.0, "retries": 0, "seconds": 67.5, "max_memory_gb": 17.03, "ce": 0.004946048153215088, "kl": 0.010359929321566597, "context": 0.010126012726686895}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.06464395672082901, "loss_scale": 4.0, "retries": 0, "seconds": 67.9, "max_memory_gb": 17.03, "ce": 0.007749440526822582, "kl": 0.002852310872185626, "context": 0.009175596118438989}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.19881440699100494, "loss_scale": 4.0, "retries": 0, "seconds": 68.4, "max_memory_gb": 17.03, "ce": 0.04221794509794563, "kl": 0.030742385773919523, "context": 0.05758913769386709}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.32710912823677063, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 17.03, "ce": 0.0159028768830467, "kl": 0.019501950771882548, "context": 0.02565385226625949}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 0.9660107493400574, "loss_scale": 4.0, "retries": 0, "seconds": 69.3, "max_memory_gb": 17.03, "ce": 0.011791162985900883, "kl": 0.46118513574992903, "context": 0.24238372579566203}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.6168968081474304, "loss_scale": 4.0, "retries": 0, "seconds": 69.8, "max_memory_gb": 17.03, "ce": 0.010861288395972224, "kl": 0.3862015298309416, "context": 0.20396205785073107}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 0.393751323223114, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 17.03, "ce": 0.04448456986574456, "kl": 0.07473974306776654, "context": 0.08185443776892498}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 0.9528965950012207, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 17.03, "ce": 0.20300727489666315, "kl": 0.008841387428901726, "context": 0.20742796948252362}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.3426593542098999, "loss_scale": 4.0, "retries": 0, "seconds": 71.1, "max_memory_gb": 17.03, "ce": 0.015056069423735607, "kl": 0.008825134327139494, "context": 0.01946863617922645}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.10129007697105408, "loss_scale": 4.0, "retries": 0, "seconds": 71.5, "max_memory_gb": 17.03, "ce": 0.008060934487730265, "kl": 0.01840615033142967, "context": 0.017264009919017553}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.9789071679115295, "loss_scale": 4.0, "retries": 0, "seconds": 71.9, "max_memory_gb": 17.03, "ce": 0.24024197226390243, "kl": 0.09044754020214896, "context": 0.2854657488060184}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.47290173172950745, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 17.03, "ce": 0.005474655183206778, "kl": 0.07463376154191792, "context": 0.042791535961441696}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.4661954641342163, "loss_scale": 4.0, "retries": 0, "seconds": 72.8, "max_memory_gb": 17.03, "ce": 0.04864725150400773, "kl": 0.33110148830746766, "context": 0.21419799292925745}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.6289879679679871, "loss_scale": 4.0, "retries": 0, "seconds": 73.2, "max_memory_gb": 17.03, "ce": 0.09312230058276327, "kl": 0.08761238165743634, "context": 0.13692849327344447}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.0681747198104858, "loss_scale": 4.0, "retries": 0, "seconds": 73.6, "max_memory_gb": 17.03, "ce": 0.07321049377787858, "kl": 0.126152463562903, "context": 0.13628672555205412}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.490954637527466, "loss_scale": 4.0, "retries": 0, "seconds": 74.0, "max_memory_gb": 17.03, "ce": 0.2198253006208688, "kl": 0.6236569188185968, "context": 0.5316537577891722}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 0.9794778823852539, "loss_scale": 4.0, "retries": 0, "seconds": 74.5, "max_memory_gb": 17.03, "ce": 0.10525978803343605, "kl": 0.07853399952728068, "context": 0.14452678785892203}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.10609472543001175, "loss_scale": 4.0, "retries": 0, "seconds": 74.9, "max_memory_gb": 17.03, "ce": 0.005300105374772102, "kl": 0.014486485774796165, "context": 0.012543348435428925}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.5901341438293457, "loss_scale": 4.0, "retries": 0, "seconds": 75.3, "max_memory_gb": 17.03, "ce": 0.5688748210377526, "kl": 0.11415106684580678, "context": 0.6259503744659014}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.2725522816181183, "loss_scale": 4.0, "retries": 0, "seconds": 75.8, "max_memory_gb": 17.03, "ce": 0.024177291430532932, "kl": 0.006739090389601188, "context": 0.027546836179681122}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.23770394921302795, "loss_scale": 4.0, "retries": 0, "seconds": 76.2, "max_memory_gb": 17.03, "ce": 0.026351861393777654, "kl": 0.005469133306178264, "context": 0.029086427646689117}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.319625735282898, "loss_scale": 4.0, "retries": 0, "seconds": 76.6, "max_memory_gb": 17.03, "ce": 0.06356582866283134, "kl": 1.1812528371810913, "context": 0.6541922390460968}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.06064365431666374, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 17.03, "ce": 0.003419166663661599, "kl": 0.0005057900798419723, "context": 0.0036720616626553237}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.6917200088500977, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 17.03, "ce": 0.0356911159997253, "kl": 0.19726267957594246, "context": 0.13432245544390753}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.19312314689159393, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 17.03, "ce": 0.01872201805235818, "kl": 0.020736611360916868, "context": 0.029090323485434055}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.2933748960494995, "loss_scale": 4.0, "retries": 0, "seconds": 78.3, "max_memory_gb": 17.03, "ce": 0.031157539633568376, "kl": 0.017510380159365013, "context": 0.03991272987332195}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.33721041679382324, "loss_scale": 4.0, "retries": 0, "seconds": 78.7, "max_memory_gb": 17.03, "ce": 0.03263972975764773, "kl": 0.02438539386548655, "context": 0.044832428564404836}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 1.0812634229660034, "loss_scale": 4.0, "retries": 0, "seconds": 79.1, "max_memory_gb": 17.03, "ce": 0.1329997160937637, "kl": 0.0756831404287368, "context": 0.1708412878215313}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.11225103586912155, "loss_scale": 4.0, "retries": 0, "seconds": 79.5, "max_memory_gb": 17.03, "ce": 0.01110675124800764, "kl": 0.002949470261228271, "context": 0.012581486138515174}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.7594441175460815, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 17.03, "ce": 0.030622158985352144, "kl": 0.6516925141768297, "context": 0.35646841234120075}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.27180856466293335, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 17.03, "ce": 0.0382986111799255, "kl": 0.04731054467265494, "context": 0.06195388198830187}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.8271092176437378, "loss_scale": 4.0, "retries": 0, "seconds": 80.7, "max_memory_gb": 17.03, "ce": 0.42855861806310713, "kl": 0.016734159144107252, "context": 0.43692570440180134}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.36870071291923523, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 17.03, "ce": 0.04123671640991233, "kl": 0.024486517522746, "context": 0.05347997319768183}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.39366614818573, "loss_scale": 4.0, "retries": 0, "seconds": 81.6, "max_memory_gb": 17.03, "ce": 0.00895782490260899, "kl": 0.12747281495830975, "context": 0.0726942319306545}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.43555688858032227, "loss_scale": 4.0, "retries": 0, "seconds": 82.0, "max_memory_gb": 17.03, "ce": 0.01700750837335363, "kl": 0.35860173347646196, "context": 0.19630837347358465}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.06283027678728104, "loss_scale": 4.0, "retries": 0, "seconds": 82.4, "max_memory_gb": 17.03, "ce": 0.004917282309179427, "kl": 0.0008670056849950925, "context": 0.005350785228074528}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.05623235926032066, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 17.03, "ce": 0.0030248176408349536, "kl": 0.0039012392808217555, "context": 0.004975437303073704}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.2644278407096863, "loss_scale": 4.0, "retries": 0, "seconds": 83.2, "max_memory_gb": 17.03, "ce": 0.018095926148816943, "kl": 0.10234395134466467, "context": 0.06926790380384773}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.4592841863632202, "loss_scale": 4.0, "retries": 0, "seconds": 83.7, "max_memory_gb": 17.03, "ce": 0.119017002536566, "kl": 0.23699821200352744, "context": 0.2375161081145052}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.5812423825263977, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 17.03, "ce": 0.05227982407086529, "kl": 0.4761710772436345, "context": 0.29036535893101245}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.0277632474899292, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 17.03, "ce": 0.23303340493839642, "kl": 0.017826313420982842, "context": 0.24194655861901992}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.1544181108474731, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 17.03, "ce": 0.2938855163520202, "kl": 0.36039742588764057, "context": 0.4740842153551057}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.11300656199455261, "loss_scale": 4.0, "retries": 0, "seconds": 85.3, "max_memory_gb": 17.03, "ce": 0.004160032054642215, "kl": 0.01837022410472855, "context": 0.013345143757760525}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 1.307254672050476, "loss_scale": 4.0, "retries": 0, "seconds": 85.7, "max_memory_gb": 17.03, "ce": 0.35139350923418533, "kl": 0.22044350085889164, "context": 0.46161525140632875}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.0238513946533203, "loss_scale": 4.0, "retries": 0, "seconds": 86.1, "max_memory_gb": 17.03, "ce": 0.5667436913354322, "kl": 0.10011550260242075, "context": 0.6168014630675316}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.051361437886953354, "loss_scale": 4.0, "retries": 0, "seconds": 86.6, "max_memory_gb": 17.03, "ce": 0.004702752106823027, "kl": 0.0023796807172402623, "context": 0.005892592558666365}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.33124712109565735, "loss_scale": 4.0, "retries": 0, "seconds": 87.0, "max_memory_gb": 17.03, "ce": 0.0437122427392751, "kl": 0.10149333086883416, "context": 0.09445891188806854}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.08974657207727432, "loss_scale": 4.0, "retries": 0, "seconds": 87.4, "max_memory_gb": 17.03, "ce": 0.01138039561919868, "kl": 0.0042718109471024945, "context": 0.0135163011145778}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 1.0476049184799194, "loss_scale": 4.0, "retries": 0, "seconds": 87.8, "max_memory_gb": 17.03, "ce": 0.001800829795683967, "kl": 0.4329603006092668, "context": 0.21828098478363245}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 2.3913276195526123, "loss_scale": 4.0, "retries": 0, "seconds": 88.2, "max_memory_gb": 17.03, "ce": 0.10487320221727714, "kl": 0.7637474794382797, "context": 0.48674695240333676}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.19971363246440887, "loss_scale": 4.0, "retries": 0, "seconds": 88.7, "max_memory_gb": 17.03, "ce": 0.06271432648645714, "kl": 0.010338675603634329, "context": 0.06788366244290955}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 1.0951231718063354, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 17.03, "ce": 0.1811512049753219, "kl": 0.07933440090255317, "context": 0.2208183978800662}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.4814234972000122, "loss_scale": 4.0, "retries": 0, "seconds": 89.5, "max_memory_gb": 17.03, "ce": 0.2591826236748602, "kl": 0.043484976566105615, "context": 0.28092511917930096}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.4331781566143036, "loss_scale": 4.0, "retries": 0, "seconds": 89.9, "max_memory_gb": 17.03, "ce": 0.0689631151908543, "kl": 0.1178324810207414, "context": 0.1278793631354347}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.030282583087682724, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 17.03, "ce": 0.0018578129893285222, "kl": 0.002896241433063551, "context": 0.003305933681986062}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.40700089931488037, "loss_scale": 4.0, "retries": 0, "seconds": 90.7, "max_memory_gb": 17.03, "ce": 0.012786234830855392, "kl": 0.02110551981604658, "context": 0.023338994767982513}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.33653727173805237, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 17.03, "ce": 0.044432983675505966, "kl": 0.010142287355847657, "context": 0.04950412758626044}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.7399264574050903, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 17.03, "ce": 0.0023914349294500425, "kl": 0.42553503411181737, "context": 0.21515894995536655}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.1276760697364807, "loss_scale": 4.0, "retries": 0, "seconds": 92.0, "max_memory_gb": 17.03, "ce": 0.009654847923229681, "kl": 0.003691989180993005, "context": 0.011500842687382828}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.15303780138492584, "loss_scale": 4.0, "retries": 0, "seconds": 92.4, "max_memory_gb": 17.03, "ce": 0.0644440292235231, "kl": 0.04007509260554798, "context": 0.08448157738894224}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.07133025676012039, "loss_scale": 4.0, "retries": 0, "seconds": 92.8, "max_memory_gb": 17.03, "ce": 0.005688674500561319, "kl": 0.0023967228462424828, "context": 0.006887035706313327}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.7325005531311035, "loss_scale": 4.0, "retries": 0, "seconds": 93.2, "max_memory_gb": 17.03, "ce": 0.17692685109796003, "kl": 0.09298358572414145, "context": 0.22341864299960434}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 0.8628749251365662, "loss_scale": 4.0, "retries": 0, "seconds": 93.7, "max_memory_gb": 17.03, "ce": 0.5233511367041501, "kl": 0.13913243213573878, "context": 0.5929173672229808}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.30474504828453064, "loss_scale": 4.0, "retries": 0, "seconds": 94.1, "max_memory_gb": 17.03, "ce": 0.11878074309788644, "kl": 0.12145576096372679, "context": 0.1795086159836501}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 1.28254234790802, "loss_scale": 4.0, "retries": 0, "seconds": 94.5, "max_memory_gb": 17.03, "ce": 0.0645886771089863, "kl": 0.9029845556542568, "context": 0.5160809705266729}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.6484095454216003, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 17.03, "ce": 0.26835493374528596, "kl": 0.04778774274745956, "context": 0.2922488045951468}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 1.629351258277893, "loss_scale": 4.0, "retries": 0, "seconds": 95.4, "max_memory_gb": 17.03, "ce": 0.4034676400551689, "kl": 0.1292928293987643, "context": 0.4681140653556213}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.2515625953674316, "loss_scale": 4.0, "retries": 0, "seconds": 95.8, "max_memory_gb": 17.03, "ce": 0.6144464304015855, "kl": 0.03008011630083729, "context": 0.629486465237278}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.32815074920654297, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 17.03, "ce": 0.006636434191023, "kl": 0.06980620152171468, "context": 0.041539535915944725}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.056760117411613464, "loss_scale": 4.0, "retries": 0, "seconds": 96.6, "max_memory_gb": 17.03, "ce": 0.002676919859368354, "kl": 0.004019138428702718, "context": 0.004686489060986787}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.6022874712944031, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 17.03, "ce": 0.16158835915848613, "kl": 0.24128401651978493, "context": 0.28223036974668503}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.32800039649009705, "loss_scale": 4.0, "retries": 0, "seconds": 97.5, "max_memory_gb": 17.03, "ce": 0.008013668200874235, "kl": 0.12004988442640752, "context": 0.0680386119929608}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.3862667679786682, "loss_scale": 4.0, "retries": 0, "seconds": 97.9, "max_memory_gb": 17.03, "ce": 0.05245736882352503, "kl": 0.007584188860164431, "context": 0.05624946278112475}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.685825765132904, "loss_scale": 4.0, "retries": 0, "seconds": 98.3, "max_memory_gb": 17.03, "ce": 0.004895298188785091, "kl": 0.35147889505606145, "context": 0.18063473975053057}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.7714501023292542, "loss_scale": 4.0, "retries": 0, "seconds": 98.7, "max_memory_gb": 17.03, "ce": 0.050158682686742395, "kl": 0.06593817769135057, "context": 0.08312777307583019}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 2.066169261932373, "loss_scale": 4.0, "retries": 0, "seconds": 99.2, "max_memory_gb": 17.03, "ce": 0.14881653164411546, "kl": 0.012514362124420586, "context": 0.15507371576313744}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 1.075538992881775, "loss_scale": 4.0, "retries": 0, "seconds": 99.6, "max_memory_gb": 17.03, "ce": 0.0503439761523623, "kl": 0.12940231251377554, "context": 0.1150451367138885}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.4573914408683777, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 17.03, "ce": 0.0034533535363152623, "kl": 0.16257655280060135, "context": 0.08474162861239165}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.5417963266372681, "loss_scale": 4.0, "retries": 0, "seconds": 100.4, "max_memory_gb": 17.03, "ce": 0.029309203804587014, "kl": 0.02707476970681455, "context": 0.04284658876713365}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.24052166938781738, "loss_scale": 4.0, "retries": 0, "seconds": 100.8, "max_memory_gb": 17.03, "ce": 0.02697902946238173, "kl": 0.012231526166715412, "context": 0.03309479253948666}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 0.6167959570884705, "loss_scale": 4.0, "retries": 0, "seconds": 101.2, "max_memory_gb": 17.03, "ce": 0.003786434550420381, "kl": 0.12579308165004477, "context": 0.06668297463329509}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.5709551572799683, "loss_scale": 4.0, "retries": 0, "seconds": 101.7, "max_memory_gb": 17.03, "ce": 0.066492917285359, "kl": 0.040216479904984226, "context": 0.08660115630482323}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.7019537687301636, "loss_scale": 4.0, "retries": 0, "seconds": 102.1, "max_memory_gb": 17.03, "ce": 0.6184547471930273, "kl": 0.35084958863444626, "context": 0.7938795284717344}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.49168601632118225, "loss_scale": 4.0, "retries": 0, "seconds": 102.5, "max_memory_gb": 17.03, "ce": 0.038199090224225074, "kl": 0.13003486705565592, "context": 0.10321652445418295}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.30870383977890015, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 17.03, "ce": 0.028533693781355396, "kl": 0.008247519610449672, "context": 0.03265745419776067}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.42206403613090515, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 17.03, "ce": 0.01063293102197349, "kl": 0.10418614011723548, "context": 0.06272600218653679}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.22887767851352692, "loss_scale": 4.0, "retries": 0, "seconds": 103.7, "max_memory_gb": 17.03, "ce": 0.008867055323207751, "kl": 0.021643368614604697, "context": 0.019688739368575625}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.38666832447052, "loss_scale": 4.0, "retries": 0, "seconds": 104.1, "max_memory_gb": 17.03, "ce": 0.49075317502138205, "kl": 0.01695840112370206, "context": 0.499232382979244}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.11291448771953583, "loss_scale": 4.0, "retries": 0, "seconds": 104.6, "max_memory_gb": 17.03, "ce": 0.009931797343597282, "kl": 0.013937149218691047, "context": 0.01690037218213547}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.23813332617282867, "loss_scale": 4.0, "retries": 0, "seconds": 105.0, "max_memory_gb": 17.03, "ce": 0.06159646951709874, "kl": 0.029016946004503552, "context": 0.07610493966785725}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.20570287108421326, "loss_scale": 4.0, "retries": 0, "seconds": 105.4, "max_memory_gb": 17.03, "ce": 0.021645031025400385, "kl": 0.0060105731718067545, "context": 0.024650317413033918}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.266798734664917, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 17.03, "ce": 0.10413038835395128, "kl": 0.026976854613167234, "context": 0.11761881277197972}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 0.6823749542236328, "loss_scale": 4.0, "retries": 0, "seconds": 106.2, "max_memory_gb": 17.03, "ce": 0.09551352561538806, "kl": 0.03468780497496482, "context": 0.11285742369364016}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 1.0921671390533447, "loss_scale": 4.0, "retries": 0, "seconds": 106.6, "max_memory_gb": 17.03, "ce": 0.2379776086017955, "kl": 0.03334883469506167, "context": 0.2546520227115252}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.16821356117725372, "loss_scale": 4.0, "retries": 0, "seconds": 107.0, "max_memory_gb": 17.03, "ce": 0.004588107512972783, "kl": 0.03732858348962509, "context": 0.02325239867786877}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.08887011557817459, "loss_scale": 4.0, "retries": 0, "seconds": 107.5, "max_memory_gb": 17.03, "ce": 0.006061189633328468, "kl": 0.005029521475080401, "context": 0.008575950341764838}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.35974183678627014, "loss_scale": 4.0, "retries": 0, "seconds": 107.9, "max_memory_gb": 17.03, "ce": 0.028948739913175814, "kl": 0.1389826595041086, "context": 0.09844007140782196}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 0.9534862041473389, "loss_scale": 4.0, "retries": 0, "seconds": 108.3, "max_memory_gb": 17.03, "ce": 0.39290873729623854, "kl": 0.09546407300513238, "context": 0.4406407750211656}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.021206054836511612, "loss_scale": 4.0, "retries": 0, "seconds": 108.7, "max_memory_gb": 17.03, "ce": 0.001529951608972624, "kl": 0.0007448694477716344, "context": 0.0019023862987523898}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.6905167698860168, "loss_scale": 4.0, "retries": 0, "seconds": 109.1, "max_memory_gb": 17.03, "ce": 0.002328258167835884, "kl": 0.44038506851211423, "context": 0.2225207869487349}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.2342296689748764, "loss_scale": 4.0, "retries": 0, "seconds": 109.6, "max_memory_gb": 17.03, "ce": 0.02894393193128053, "kl": 0.008247015066444874, "context": 0.03306744003202766}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 0.6701333522796631, "loss_scale": 4.0, "retries": 0, "seconds": 110.0, "max_memory_gb": 17.03, "ce": 0.046999166370369494, "kl": 0.04473653715103865, "context": 0.06936743739061058}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.09018982946872711, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 17.03, "ce": 0.009507904745987616, "kl": 0.005324674028088339, "context": 0.012170241825515404}
|
runs/single_internvl/context/L29_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L29_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.29.self_attn.q_proj",
|
| 35 |
+
"model.language_model.layers.29.self_attn.v_proj",
|
| 36 |
+
"model.language_model.layers.29.mlp.gate_proj",
|
| 37 |
+
"model.language_model.layers.29.self_attn.o_proj",
|
| 38 |
+
"model.language_model.layers.29.mlp.down_proj",
|
| 39 |
+
"model.language_model.layers.29.self_attn.k_proj",
|
| 40 |
+
"model.language_model.layers.29.mlp.up_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L29_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 29,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L29_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.2368045300245285, "loss_scale": 4.0, "retries": 0, "seconds": 1.4, "max_memory_gb": 16.69, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.327043354511261, "loss_scale": 4.0, "retries": 0, "seconds": 1.8, "max_memory_gb": 16.7, "ce": 0.6404375040219747, "kl": 0.46727469330153326, "context": 0.874074881321576}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 0.769045889377594, "loss_scale": 4.0, "retries": 0, "seconds": 2.2, "max_memory_gb": 16.7, "ce": 1.0932224094867706, "kl": 0.06784487329423428, "context": 1.1271448582410812}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 0.3907756507396698, "loss_scale": 4.0, "retries": 0, "seconds": 2.7, "max_memory_gb": 16.7, "ce": 0.451626935711829, "kl": 0.43012601514419657, "context": 0.6666899229676346}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.008808257058262825, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 16.7, "ce": 0.009193224415867007, "kl": 0.008233361233578762, "context": 0.01330990514907171}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.25807544589042664, "loss_scale": 4.0, "retries": 0, "seconds": 3.6, "max_memory_gb": 16.7, "ce": 0.29110770285478793, "kl": 0.0742342584162543, "context": 0.3282248232717393}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 0.6202875971794128, "loss_scale": 4.0, "retries": 0, "seconds": 4.0, "max_memory_gb": 16.7, "ce": 0.8280366486087587, "kl": 0.1227430608563509, "context": 0.8894081706548604}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 0.6656560301780701, "loss_scale": 4.0, "retries": 0, "seconds": 4.5, "max_memory_gb": 16.71, "ce": 0.795151861384511, "kl": 0.2897674401756376, "context": 0.9400355815887451}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 0.48557519912719727, "loss_scale": 4.0, "retries": 0, "seconds": 4.9, "max_memory_gb": 16.71, "ce": 0.531519902870059, "kl": 0.13788561860565096, "context": 0.6004627235233784}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.11990529298782349, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 16.71, "ce": 0.05340433833771385, "kl": 0.045304756436962634, "context": 0.0760567169636488}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 0.6163947582244873, "loss_scale": 4.0, "retries": 0, "seconds": 5.9, "max_memory_gb": 16.71, "ce": 0.7089882403961383, "kl": 0.004855293829677976, "context": 0.7114158784897882}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 0.4634014666080475, "loss_scale": 4.0, "retries": 0, "seconds": 6.3, "max_memory_gb": 16.71, "ce": 0.4747669590760779, "kl": 0.002514318157409434, "context": 0.47602413246568176}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 0.4194768965244293, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 16.71, "ce": 0.42903357234899886, "kl": 0.019577729583033943, "context": 0.4388224463800725}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 0.8380067944526672, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 16.71, "ce": 0.7802784293889999, "kl": 0.9456784292124212, "context": 1.2531176656484604}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 0.7394029498100281, "loss_scale": 4.0, "retries": 0, "seconds": 7.6, "max_memory_gb": 16.71, "ce": 0.5328527746364671, "kl": 0.4689808783755325, "context": 0.7673432250001042}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.32328182458877563, "loss_scale": 4.0, "retries": 0, "seconds": 8.1, "max_memory_gb": 16.71, "ce": 0.20833582297200337, "kl": 0.0009543582834794506, "context": 0.2088130015254137}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.39328280091285706, "loss_scale": 4.0, "retries": 0, "seconds": 8.5, "max_memory_gb": 16.71, "ce": 0.33080662053544074, "kl": 0.8246882430976257, "context": 0.7431507483124733}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 0.8903608918190002, "loss_scale": 4.0, "retries": 0, "seconds": 8.9, "max_memory_gb": 16.71, "ce": 0.5575144251051825, "kl": 0.20974945105990628, "context": 0.6623891585477395}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 1.8549394607543945, "loss_scale": 4.0, "retries": 0, "seconds": 9.4, "max_memory_gb": 16.71, "ce": 0.8175516128540039, "kl": 0.1628367763478309, "context": 0.8989700078964233}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.019542271271348, "loss_scale": 4.0, "retries": 0, "seconds": 9.8, "max_memory_gb": 16.71, "ce": 0.0023010416589386296, "kl": 0.02123097529693041, "context": 0.012916529067297233}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.0111737251281738, "loss_scale": 4.0, "retries": 0, "seconds": 10.3, "max_memory_gb": 16.71, "ce": 0.38248012181884405, "kl": 0.048789063383082976, "context": 0.4068746656321309}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 0.9598909020423889, "loss_scale": 4.0, "retries": 0, "seconds": 10.7, "max_memory_gb": 16.71, "ce": 0.423254440058372, "kl": 0.026696948742028326, "context": 0.43660291978449095}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 1.3687055110931396, "loss_scale": 4.0, "retries": 0, "seconds": 11.2, "max_memory_gb": 16.71, "ce": 0.5924570706556551, "kl": 0.9890876810237614, "context": 1.0870009209465934}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 1.0363179445266724, "loss_scale": 4.0, "retries": 0, "seconds": 11.6, "max_memory_gb": 16.71, "ce": 0.3033435009419918, "kl": 0.3332179016433656, "context": 0.46995246410369873}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 1.4582756757736206, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 16.71, "ce": 0.7006395137796062, "kl": 0.21715917019173503, "context": 0.8092190958559513}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 1.3045154809951782, "loss_scale": 4.0, "retries": 0, "seconds": 12.5, "max_memory_gb": 16.71, "ce": 0.35067669302225113, "kl": 0.7001139237545431, "context": 0.7007336467504501}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 1.391527771949768, "loss_scale": 4.0, "retries": 0, "seconds": 12.9, "max_memory_gb": 16.71, "ce": 0.4600100731040584, "kl": 0.4283303649935988, "context": 0.6741752581001492}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 0.8539785742759705, "loss_scale": 4.0, "retries": 0, "seconds": 13.4, "max_memory_gb": 16.71, "ce": 0.6140342019498348, "kl": 0.9417435270734131, "context": 1.0849059522151947}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 1.3099579811096191, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 16.71, "ce": 0.3630714048631489, "kl": 0.3595145691651851, "context": 0.5428286790847778}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 1.3632243871688843, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 16.71, "ce": 0.3477202722351649, "kl": 0.007463099724645872, "context": 0.35145183234271826}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 1.9161518812179565, "loss_scale": 4.0, "retries": 0, "seconds": 14.7, "max_memory_gb": 16.71, "ce": 0.6607154347002506, "kl": 0.1619139164686203, "context": 0.7416723966598511}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 2.2006258964538574, "loss_scale": 4.0, "retries": 0, "seconds": 15.1, "max_memory_gb": 16.71, "ce": 0.3962809219956398, "kl": 0.015312236821046099, "context": 0.40393703803420067}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 0.9485607147216797, "loss_scale": 4.0, "retries": 0, "seconds": 15.6, "max_memory_gb": 16.71, "ce": 0.21800466580316424, "kl": 0.07039269781671464, "context": 0.25320101249963045}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 1.100024700164795, "loss_scale": 4.0, "retries": 0, "seconds": 16.0, "max_memory_gb": 16.71, "ce": 0.10886622611360508, "kl": 0.7500241809002546, "context": 0.48387830905630835}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 1.0855989456176758, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 16.71, "ce": 0.1670604961545905, "kl": 0.004122340652429557, "context": 0.1691216647886904}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 1.0965862274169922, "loss_scale": 4.0, "retries": 0, "seconds": 17.0, "max_memory_gb": 16.71, "ce": 0.25313852942053927, "kl": 0.05021007238246966, "context": 0.27824357282952406}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 1.0173462629318237, "loss_scale": 4.0, "retries": 0, "seconds": 17.4, "max_memory_gb": 16.71, "ce": 0.2645794189738808, "kl": 0.7572606479516253, "context": 0.643209726200439}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 1.0166475772857666, "loss_scale": 4.0, "retries": 0, "seconds": 17.8, "max_memory_gb": 16.71, "ce": 0.14396110828965902, "kl": 0.5090394093422219, "context": 0.3984808037057519}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.3153465688228607, "loss_scale": 4.0, "retries": 0, "seconds": 18.2, "max_memory_gb": 16.71, "ce": 0.13841690152912633, "kl": 0.009311500383773819, "context": 0.1430726507678628}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.4071977734565735, "loss_scale": 4.0, "retries": 0, "seconds": 18.7, "max_memory_gb": 16.71, "ce": 0.0360455556656234, "kl": 0.01092245674226433, "context": 0.04150678447331302}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 0.5732959508895874, "loss_scale": 4.0, "retries": 0, "seconds": 19.1, "max_memory_gb": 16.71, "ce": 0.3898785043274984, "kl": 0.12497198674827814, "context": 0.4523644857108593}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 0.7198561429977417, "loss_scale": 4.0, "retries": 0, "seconds": 19.5, "max_memory_gb": 16.71, "ce": 0.0629174248824711, "kl": 0.006988826884480659, "context": 0.06641183742613066}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.7262969613075256, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 16.71, "ce": 0.05828394429408945, "kl": 0.09589021779538598, "context": 0.1062290531990584}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 1.2888867855072021, "loss_scale": 4.0, "retries": 0, "seconds": 20.4, "max_memory_gb": 16.71, "ce": 0.5736977792184916, "kl": 1.343023726662068, "context": 1.2452096462802729}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 0.8340447545051575, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 16.71, "ce": 0.8888028300134465, "kl": 0.24021743292541942, "context": 1.0089115366863552}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.27493610978126526, "loss_scale": 4.0, "retries": 0, "seconds": 21.3, "max_memory_gb": 16.71, "ce": 0.08701110724359751, "kl": 0.27264758537057787, "context": 0.22333490825258195}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.37824442982673645, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 16.71, "ce": 0.01760654376266757, "kl": 0.04345504107186571, "context": 0.03933406452415511}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.5810438394546509, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 16.71, "ce": 0.03545440763991792, "kl": 0.00795467289572116, "context": 0.03943174460437149}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 0.747589111328125, "loss_scale": 4.0, "retries": 0, "seconds": 22.7, "max_memory_gb": 16.71, "ce": 0.0445215639701928, "kl": 0.31313182958592733, "context": 0.2010874750376388}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 1.2603031396865845, "loss_scale": 4.0, "retries": 0, "seconds": 23.1, "max_memory_gb": 16.71, "ce": 0.13851953029734432, "kl": 0.011725601812941022, "context": 0.14438233120017685}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 0.8463196754455566, "loss_scale": 4.0, "retries": 0, "seconds": 23.6, "max_memory_gb": 16.71, "ce": 0.5454443816597632, "kl": 0.020497909824030103, "context": 0.5556933388979814}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.5768530964851379, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 16.71, "ce": 0.0444487122585997, "kl": 0.21368631371296942, "context": 0.15129186701960862}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 0.9573751091957092, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 16.71, "ce": 0.15706212702207267, "kl": 0.36911876406520605, "context": 0.34162150509655476}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.5050751566886902, "loss_scale": 4.0, "retries": 0, "seconds": 25.0, "max_memory_gb": 16.71, "ce": 0.03886442378279753, "kl": 0.33992804048466496, "context": 0.20882845192681998}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.1312277764081955, "loss_scale": 4.0, "retries": 0, "seconds": 25.4, "max_memory_gb": 16.71, "ce": 0.015492717400775291, "kl": 0.011091876755017438, "context": 0.02103865600656718}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.5078900456428528, "loss_scale": 4.0, "retries": 0, "seconds": 25.8, "max_memory_gb": 16.71, "ce": 0.025040312204509974, "kl": 0.27726728352718055, "context": 0.16367395734414458}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 0.29348644614219666, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 16.71, "ce": 0.014825327656581067, "kl": 0.018851741904654773, "context": 0.024251197661214974}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.10412448644638062, "loss_scale": 4.0, "retries": 0, "seconds": 26.7, "max_memory_gb": 16.71, "ce": 0.007554385181720136, "kl": 0.0011070561163109005, "context": 0.008107913443382131}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 0.2867010831832886, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 16.71, "ce": 0.02975150541169569, "kl": 0.05942369518743362, "context": 0.05946335208136588}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 0.4637559950351715, "loss_scale": 4.0, "retries": 0, "seconds": 27.6, "max_memory_gb": 16.71, "ce": 0.05299242938053794, "kl": 0.0498827209812589, "context": 0.07793378975475207}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 1.3525638580322266, "loss_scale": 4.0, "retries": 0, "seconds": 28.1, "max_memory_gb": 16.71, "ce": 0.0235990472137928, "kl": 0.862341269152239, "context": 0.45476967887952924}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.12171317636966705, "loss_scale": 4.0, "retries": 0, "seconds": 28.5, "max_memory_gb": 16.71, "ce": 0.028137307715951465, "kl": 0.008328624655405292, "context": 0.03230162005638704}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 0.5662246942520142, "loss_scale": 4.0, "retries": 0, "seconds": 29.0, "max_memory_gb": 16.71, "ce": 0.04043019510572776, "kl": 0.12189843076339457, "context": 0.1013794126920402}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.16281436383724213, "loss_scale": 4.0, "retries": 0, "seconds": 29.4, "max_memory_gb": 16.71, "ce": 0.016869472339749336, "kl": 0.016335611027898267, "context": 0.025037277373485267}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 0.6464684009552002, "loss_scale": 4.0, "retries": 0, "seconds": 29.9, "max_memory_gb": 16.71, "ce": 0.20452961651608348, "kl": 0.15183181827887893, "context": 0.2804455179721117}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 0.9418254494667053, "loss_scale": 4.0, "retries": 0, "seconds": 30.3, "max_memory_gb": 16.71, "ce": 0.1586373824961811, "kl": 0.06898525123745003, "context": 0.19313000070428643}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.09022855758666992, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 16.71, "ce": 0.005944273114437237, "kl": 0.008874883002135903, "context": 0.010381714528193697}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.1474025696516037, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 16.71, "ce": 0.01047287659093854, "kl": 0.043806473400763934, "context": 0.03237611320219003}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.5302525162696838, "loss_scale": 4.0, "retries": 0, "seconds": 31.7, "max_memory_gb": 16.71, "ce": 0.2958965701982379, "kl": 0.028403541538864374, "context": 0.31009834236465394}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.0688260868191719, "loss_scale": 4.0, "retries": 0, "seconds": 32.1, "max_memory_gb": 16.71, "ce": 0.003104300209088251, "kl": 0.0008234810875364929, "context": 0.0035160408151568845}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.12043743580579758, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 16.71, "ce": 0.00810248024208704, "kl": 0.010576749264146201, "context": 0.013390854757744819}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.47359955310821533, "loss_scale": 4.0, "retries": 0, "seconds": 33.0, "max_memory_gb": 16.71, "ce": 0.018363282666541636, "kl": 0.40850812022108585, "context": 0.22261733701452613}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 0.753589391708374, "loss_scale": 4.0, "retries": 0, "seconds": 33.4, "max_memory_gb": 16.71, "ce": 0.3673154679345316, "kl": 0.10377357929974096, "context": 0.4192022445749899}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.15636923909187317, "loss_scale": 4.0, "retries": 0, "seconds": 33.9, "max_memory_gb": 16.71, "ce": 0.061261469214514364, "kl": 0.12201028881463571, "context": 0.12226661739805422}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 0.5778779983520508, "loss_scale": 4.0, "retries": 0, "seconds": 34.3, "max_memory_gb": 16.71, "ce": 0.45779934019083157, "kl": 1.000898611266166, "context": 0.9582486720755696}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.1487363576889038, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 16.71, "ce": 0.011480047221994027, "kl": 1.7031282427487895, "context": 0.8630441925488412}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.10541275143623352, "loss_scale": 4.0, "retries": 0, "seconds": 35.2, "max_memory_gb": 16.71, "ce": 0.006289156423008535, "kl": 0.010978091879223939, "context": 0.011778202417190187}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.39423996210098267, "loss_scale": 4.0, "retries": 0, "seconds": 35.6, "max_memory_gb": 16.71, "ce": 0.0013913031871197745, "kl": 0.07212100806646049, "context": 0.03745180694386363}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.15400977432727814, "loss_scale": 4.0, "retries": 0, "seconds": 36.1, "max_memory_gb": 16.71, "ce": 0.018754970208647137, "kl": 0.03447577588440254, "context": 0.03599285818017961}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 0.4575210213661194, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 16.71, "ce": 0.003171216400915, "kl": 0.08595484515484486, "context": 0.04614864039740496}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.514958918094635, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 16.71, "ce": 0.0736928636324592, "kl": 0.06157279672333971, "context": 0.10447926499182358}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.07922427356243134, "loss_scale": 4.0, "retries": 0, "seconds": 37.4, "max_memory_gb": 16.71, "ce": 0.00797699237591587, "kl": 0.002039927025180077, "context": 0.008996955846669152}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 1.4993987083435059, "loss_scale": 4.0, "retries": 0, "seconds": 37.8, "max_memory_gb": 16.71, "ce": 0.1598056461662054, "kl": 0.7504231357888784, "context": 0.5350172028993256}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 0.4829157292842865, "loss_scale": 4.0, "retries": 0, "seconds": 38.3, "max_memory_gb": 16.71, "ce": 0.02589954252289317, "kl": 0.0441833840104664, "context": 0.04799123632164992}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 0.7096356153488159, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 16.71, "ce": 0.03429349738962628, "kl": 0.02974953286744153, "context": 0.04916826568705801}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.2602087259292603, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 16.71, "ce": 0.217115031555295, "kl": 0.45480364421382546, "context": 0.4445168530801311}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.7413747310638428, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 16.71, "ce": 0.006965069842408411, "kl": 0.7334089117357507, "context": 0.3736695144325495}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.17610470950603485, "loss_scale": 4.0, "retries": 0, "seconds": 40.1, "max_memory_gb": 16.71, "ce": 0.008511748852470191, "kl": 0.020101023950588548, "context": 0.01856226056588639}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.0433182418346405, "loss_scale": 4.0, "retries": 0, "seconds": 40.5, "max_memory_gb": 16.71, "ce": 0.012362602326447814, "kl": 0.010057802206574706, "context": 0.017391503433827893}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.056681931018829346, "loss_scale": 4.0, "retries": 0, "seconds": 40.9, "max_memory_gb": 16.71, "ce": 0.006081424071453512, "kl": 0.004243359493557364, "context": 0.00820310364360921}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 0.6189141273498535, "loss_scale": 4.0, "retries": 0, "seconds": 41.4, "max_memory_gb": 16.71, "ce": 0.23181540091900388, "kl": 0.5263204759685323, "context": 0.4949756344794878}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.274174690246582, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 16.71, "ce": 0.3520932191313477, "kl": 0.48381724672526616, "context": 0.5940018416185922}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 0.7131308913230896, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 16.71, "ce": 0.07104845946014393, "kl": 0.48102535076759523, "context": 0.3115611294924747}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.2233120203018188, "loss_scale": 4.0, "retries": 0, "seconds": 42.7, "max_memory_gb": 16.71, "ce": 0.18158194301940966, "kl": 0.10346697343629785, "context": 0.23331542624509893}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.4160939455032349, "loss_scale": 4.0, "retries": 0, "seconds": 43.2, "max_memory_gb": 16.71, "ce": 0.500761864852393, "kl": 0.030716669163666666, "context": 0.5161202121526003}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 0.7726480960845947, "loss_scale": 4.0, "retries": 0, "seconds": 43.6, "max_memory_gb": 16.71, "ce": 0.46687257062876597, "kl": 0.006340614743749029, "context": 0.47004288247262593}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 1.1073535680770874, "loss_scale": 4.0, "retries": 0, "seconds": 44.1, "max_memory_gb": 16.71, "ce": 0.015857250669796485, "kl": 0.395851100445725, "context": 0.21378279710188508}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.3725271224975586, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 16.71, "ce": 0.0031024569761939347, "kl": 0.08320402721801656, "context": 0.04470447197672911}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.11662140488624573, "loss_scale": 4.0, "retries": 0, "seconds": 45.0, "max_memory_gb": 16.71, "ce": 0.014580606004528818, "kl": 0.009454738430576981, "context": 0.019307975802803412}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 0.7715573310852051, "loss_scale": 4.0, "retries": 0, "seconds": 45.4, "max_memory_gb": 16.71, "ce": 0.02947282303648535, "kl": 0.6977854511351325, "context": 0.378365551281604}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.17554689943790436, "loss_scale": 4.0, "retries": 0, "seconds": 45.9, "max_memory_gb": 16.71, "ce": 0.004645833309041336, "kl": 0.09139214095193893, "context": 0.05034190486185253}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.5939478874206543, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 16.71, "ce": 0.04481267894152552, "kl": 0.027034853002987802, "context": 0.05833010550122708}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.1252141147851944, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 16.71, "ce": 0.010280071532179136, "kl": 0.020277320763852913, "context": 0.02041873225243762}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.0947798565030098, "loss_scale": 4.0, "retries": 0, "seconds": 47.2, "max_memory_gb": 16.71, "ce": 0.0047238194674719125, "kl": 0.008344574938291771, "context": 0.00889610669673857}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 0.634733259677887, "loss_scale": 4.0, "retries": 0, "seconds": 47.6, "max_memory_gb": 16.71, "ce": 0.05101200641365722, "kl": 0.40641220181714743, "context": 0.2542181061580777}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.03221597149968147, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 16.71, "ce": 0.002580111045972444, "kl": 0.0014188248405844206, "context": 0.003289523534476757}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.7471413612365723, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 16.71, "ce": 0.34437734522362007, "kl": 0.14398403624477396, "context": 0.41636937403382035}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.7013013362884521, "loss_scale": 4.0, "retries": 0, "seconds": 49.0, "max_memory_gb": 16.71, "ce": 0.04860791179089574, "kl": 0.013260778294224451, "context": 0.05523830228776205}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 0.9138124585151672, "loss_scale": 4.0, "retries": 0, "seconds": 49.4, "max_memory_gb": 16.71, "ce": 0.05922757138614543, "kl": 0.29234934551641345, "context": 0.20540223829448223}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.3075295686721802, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 16.71, "ce": 0.08226949721574783, "kl": 0.27057445084210485, "context": 0.21755672874860466}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.015218937769532204, "loss_scale": 4.0, "retries": 0, "seconds": 50.3, "max_memory_gb": 16.71, "ce": 0.0011360972375769052, "kl": 0.0010239974653813988, "context": 0.0016480959166074172}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.05767499655485153, "loss_scale": 4.0, "retries": 0, "seconds": 50.7, "max_memory_gb": 16.71, "ce": 0.004154552392719779, "kl": 0.0006125797162894742, "context": 0.004460842224943917}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 0.9578631520271301, "loss_scale": 4.0, "retries": 0, "seconds": 51.2, "max_memory_gb": 16.71, "ce": 0.006171071727294475, "kl": 0.2911490532569587, "context": 0.1517456010915339}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.383124589920044, "loss_scale": 4.0, "retries": 0, "seconds": 51.6, "max_memory_gb": 16.71, "ce": 0.010361607099184766, "kl": 0.9193729919497855, "context": 0.4700480934116058}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.18402741849422455, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 16.71, "ce": 0.005406147425674135, "kl": 0.032201576887018746, "context": 0.021506935721845366}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 1.069904088973999, "loss_scale": 4.0, "retries": 0, "seconds": 52.5, "max_memory_gb": 16.71, "ce": 0.10144173370099452, "kl": 0.5243120627246185, "context": 0.36359776506469643}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 1.5753618478775024, "loss_scale": 4.0, "retries": 0, "seconds": 52.9, "max_memory_gb": 16.71, "ce": 0.05465306757832877, "kl": 0.10171535922836483, "context": 0.10551074721297482}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.46985381841659546, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 16.71, "ce": 0.21198910542443627, "kl": 0.008683587510176949, "context": 0.21633089733950328}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.13115957379341125, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 16.71, "ce": 0.01580141131853452, "kl": 0.021349271857843632, "context": 0.026476046739844605}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 2.2988944053649902, "loss_scale": 4.0, "retries": 0, "seconds": 54.3, "max_memory_gb": 16.71, "ce": 0.42899152600148227, "kl": 0.16052397682506125, "context": 0.5092534966242965}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.19642528891563416, "loss_scale": 4.0, "retries": 0, "seconds": 54.8, "max_memory_gb": 16.71, "ce": 0.060893563448189525, "kl": 0.038475211476907134, "context": 0.08013116963775246}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.5918744206428528, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 16.71, "ce": 0.09097812068648636, "kl": 0.019989633190562017, "context": 0.10097293716535205}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.3699270188808441, "loss_scale": 4.0, "retries": 0, "seconds": 55.6, "max_memory_gb": 16.71, "ce": 0.030098459334112704, "kl": 0.019091302761808038, "context": 0.03964411071501672}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.3693232536315918, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 16.71, "ce": 0.734637500892859, "kl": 0.6113216503581498, "context": 1.040298349398654}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.15954823791980743, "loss_scale": 4.0, "retries": 0, "seconds": 56.5, "max_memory_gb": 16.71, "ce": 0.11157838342478499, "kl": 0.07891165509499842, "context": 0.15103420364903286}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.4899884462356567, "loss_scale": 4.0, "retries": 0, "seconds": 57.0, "max_memory_gb": 16.71, "ce": 0.6607798151671886, "kl": 0.12692392198368907, "context": 0.7242417596280575}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.05602286756038666, "loss_scale": 4.0, "retries": 0, "seconds": 57.4, "max_memory_gb": 16.71, "ce": 0.004240061403834261, "kl": 0.0006616969058086397, "context": 0.00457090996496845}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 2.9427216053009033, "loss_scale": 4.0, "retries": 0, "seconds": 57.8, "max_memory_gb": 16.71, "ce": 0.16898957910598256, "kl": 1.0649576223513577, "context": 0.7014683860470541}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 0.7505726218223572, "loss_scale": 4.0, "retries": 0, "seconds": 58.3, "max_memory_gb": 16.71, "ce": 0.07602168811718002, "kl": 0.03306893883200246, "context": 0.09255616122391075}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.4982530474662781, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 16.71, "ce": 0.04235470632556826, "kl": 0.2709699947772606, "context": 0.17783969570882618}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 1.537609577178955, "loss_scale": 4.0, "retries": 0, "seconds": 59.2, "max_memory_gb": 16.71, "ce": 0.21041193984274287, "kl": 0.30502206430537626, "context": 0.36292297317413613}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 1.28470778465271, "loss_scale": 4.0, "retries": 0, "seconds": 59.6, "max_memory_gb": 16.71, "ce": 0.19866602565161884, "kl": 0.02989489381434396, "context": 0.21361347768106498}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.8809653520584106, "loss_scale": 4.0, "retries": 0, "seconds": 60.0, "max_memory_gb": 16.71, "ce": 0.38239821937168017, "kl": 0.05607438343577087, "context": 0.41043541952967644}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.5466724634170532, "loss_scale": 4.0, "retries": 0, "seconds": 60.5, "max_memory_gb": 16.71, "ce": 0.16299361077471985, "kl": 0.002304727406198026, "context": 0.16414596912363777}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.553953230381012, "loss_scale": 4.0, "retries": 0, "seconds": 60.9, "max_memory_gb": 16.71, "ce": 0.009703548159450293, "kl": 0.10276824159518583, "context": 0.06108766995021142}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 1.234946846961975, "loss_scale": 4.0, "retries": 0, "seconds": 61.3, "max_memory_gb": 16.71, "ce": 0.05667556058324408, "kl": 0.07236247430773801, "context": 0.0928568017843645}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.3837984800338745, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 16.71, "ce": 0.014168832887662575, "kl": 0.05299561898573302, "context": 0.04066664236597717}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 0.8844560980796814, "loss_scale": 4.0, "retries": 0, "seconds": 62.3, "max_memory_gb": 16.71, "ce": 0.06040463445242494, "kl": 0.11135801766067743, "context": 0.11608364339917898}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 1.0091288089752197, "loss_scale": 4.0, "retries": 0, "seconds": 62.7, "max_memory_gb": 16.71, "ce": 0.003378009765583556, "kl": 0.22437468711268593, "context": 0.11556535107956734}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.08126889169216156, "loss_scale": 4.0, "retries": 0, "seconds": 63.2, "max_memory_gb": 16.71, "ce": 0.0047739853253006, "kl": 0.0028628174532059347, "context": 0.006205394180142321}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.04643251374363899, "loss_scale": 4.0, "retries": 0, "seconds": 63.6, "max_memory_gb": 16.71, "ce": 0.003603604760428425, "kl": 0.0034481337534089107, "context": 0.005327671664417721}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.21796993911266327, "loss_scale": 4.0, "retries": 0, "seconds": 64.1, "max_memory_gb": 16.71, "ce": 0.014615579864766914, "kl": 0.008864728617481887, "context": 0.019047943991608918}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.8241688013076782, "loss_scale": 4.0, "retries": 0, "seconds": 64.5, "max_memory_gb": 16.71, "ce": 0.339815295432345, "kl": 0.08667577264941428, "context": 0.38315318897366524}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.7050882577896118, "loss_scale": 4.0, "retries": 0, "seconds": 65.0, "max_memory_gb": 16.71, "ce": 0.035232391936006024, "kl": 0.2145392156962771, "context": 0.14250200189417228}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.07749767601490021, "loss_scale": 4.0, "retries": 0, "seconds": 65.4, "max_memory_gb": 16.71, "ce": 0.001700727385468781, "kl": 0.008759223919696524, "context": 0.0060803393280366436}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.18957076966762543, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 16.71, "ce": 0.008986407774500549, "kl": 0.013926351761256228, "context": 0.015949583961628377}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.4763610363006592, "loss_scale": 4.0, "retries": 0, "seconds": 66.3, "max_memory_gb": 16.71, "ce": 0.5092421722511062, "kl": 0.484220547194127, "context": 0.7513524332462111}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.08148222416639328, "loss_scale": 4.0, "retries": 0, "seconds": 66.7, "max_memory_gb": 16.71, "ce": 0.004460379013835336, "kl": 0.011859442092827521, "context": 0.010390099756477866}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.07100266963243484, "loss_scale": 4.0, "retries": 0, "seconds": 67.2, "max_memory_gb": 16.71, "ce": 0.0078302816618816, "kl": 0.0020040814706590027, "context": 0.008832322404487059}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.20121361315250397, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 16.71, "ce": 0.03698073371197097, "kl": 0.025577153894118965, "context": 0.04976930859265849}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.31113266944885254, "loss_scale": 4.0, "retries": 0, "seconds": 68.1, "max_memory_gb": 16.71, "ce": 0.011391480704332935, "kl": 0.016186903050765977, "context": 0.01948493222880643}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 1.2242207527160645, "loss_scale": 4.0, "retries": 0, "seconds": 68.5, "max_memory_gb": 16.71, "ce": 0.010872962331632152, "kl": 0.47305833307200373, "context": 0.24740212771575898}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 0.7952424883842468, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 16.71, "ce": 0.0072291401847905945, "kl": 0.4160452163305308, "context": 0.2152517509966856}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 0.8161841034889221, "loss_scale": 4.0, "retries": 0, "seconds": 69.4, "max_memory_gb": 16.71, "ce": 0.06713232496986166, "kl": 0.07820802432252094, "context": 0.10623634164221585}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 1.5322774648666382, "loss_scale": 4.0, "retries": 0, "seconds": 69.8, "max_memory_gb": 16.71, "ce": 0.22141521272351383, "kl": 0.011130610574582533, "context": 0.22698051802581176}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.18527641892433167, "loss_scale": 4.0, "retries": 0, "seconds": 70.3, "max_memory_gb": 16.71, "ce": 0.010600751425954513, "kl": 0.005980090337061483, "context": 0.013590796377684455}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.24941034615039825, "loss_scale": 4.0, "retries": 0, "seconds": 70.7, "max_memory_gb": 16.71, "ce": 0.011947727238293737, "kl": 0.026220194915367756, "context": 0.02505782520165667}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.8233588337898254, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 16.71, "ce": 0.18439486376882996, "kl": 0.0673708093672758, "context": 0.2180802694929298}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.3812306523323059, "loss_scale": 4.0, "retries": 0, "seconds": 71.6, "max_memory_gb": 16.71, "ce": 0.005530665173864691, "kl": 0.08463068748824298, "context": 0.04784600998391397}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 0.850827157497406, "loss_scale": 4.0, "retries": 0, "seconds": 72.1, "max_memory_gb": 16.71, "ce": 0.03811356643564068, "kl": 0.3835885763692204, "context": 0.22990785795263946}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 0.6899948120117188, "loss_scale": 4.0, "retries": 0, "seconds": 72.5, "max_memory_gb": 16.71, "ce": 0.0813293045212049, "kl": 0.08082721672963089, "context": 0.12174291288101813}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.1962462663650513, "loss_scale": 4.0, "retries": 0, "seconds": 73.0, "max_memory_gb": 16.71, "ce": 0.0927574787492631, "kl": 0.1785185935877962, "context": 0.18201677556498908}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 2.5697832107543945, "loss_scale": 4.0, "retries": 0, "seconds": 73.3, "max_memory_gb": 16.71, "ce": 0.2330172373331152, "kl": 0.6724347355775535, "context": 0.5692346016294323}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 0.7112966179847717, "loss_scale": 4.0, "retries": 0, "seconds": 73.7, "max_memory_gb": 16.71, "ce": 0.08604453012230806, "kl": 0.07414408546173945, "context": 0.12311657308600843}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.10644535720348358, "loss_scale": 4.0, "retries": 0, "seconds": 74.1, "max_memory_gb": 16.71, "ce": 0.0051848420553142205, "kl": 0.012808190885152726, "context": 0.011588937064516358}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.696427583694458, "loss_scale": 4.0, "retries": 0, "seconds": 74.5, "max_memory_gb": 16.71, "ce": 0.4945782715512905, "kl": 0.10635858309979085, "context": 0.5477575621625874}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.4064151346683502, "loss_scale": 4.0, "retries": 0, "seconds": 74.9, "max_memory_gb": 16.71, "ce": 0.02520511334296316, "kl": 0.00878405045659747, "context": 0.029597138112876564}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.41470953822135925, "loss_scale": 4.0, "retries": 0, "seconds": 75.4, "max_memory_gb": 16.71, "ce": 0.028323719416221138, "kl": 0.004052875010529533, "context": 0.030350156150234398}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.5999475717544556, "loss_scale": 4.0, "retries": 0, "seconds": 75.7, "max_memory_gb": 16.71, "ce": 0.06426927947904915, "kl": 1.153262972831726, "context": 0.6409007497131824}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.02439856342971325, "loss_scale": 4.0, "retries": 0, "seconds": 76.1, "max_memory_gb": 16.71, "ce": 0.0013683197248610668, "kl": 0.0002985957644341397, "context": 0.0015176176020759158}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 0.9757713079452515, "loss_scale": 4.0, "retries": 0, "seconds": 76.5, "max_memory_gb": 16.71, "ce": 0.02459436873323284, "kl": 0.21867510967422277, "context": 0.1339319198159501}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.2097717970609665, "loss_scale": 4.0, "retries": 0, "seconds": 76.9, "max_memory_gb": 16.71, "ce": 0.012605406431248412, "kl": 0.01918991914135404, "context": 0.02220036496873945}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.4217269718647003, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 16.71, "ce": 0.02773283055284992, "kl": 0.01804577963775955, "context": 0.036755720793735236}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.4552392363548279, "loss_scale": 4.0, "retries": 0, "seconds": 77.7, "max_memory_gb": 16.71, "ce": 0.03040705765306484, "kl": 0.02159059582595546, "context": 0.04120235451409826}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 1.860949993133545, "loss_scale": 4.0, "retries": 0, "seconds": 78.2, "max_memory_gb": 16.71, "ce": 0.19531967234797776, "kl": 0.06897702277638018, "context": 0.22980818897485733}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.1081150472164154, "loss_scale": 4.0, "retries": 0, "seconds": 78.6, "max_memory_gb": 16.71, "ce": 0.0071619321970501915, "kl": 0.003324082614199142, "context": 0.008823973592370749}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 0.8179239630699158, "loss_scale": 4.0, "retries": 0, "seconds": 79.1, "max_memory_gb": 16.71, "ce": 0.02780293249816168, "kl": 0.7513666799204657, "context": 0.4034862500993768}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.7400793433189392, "loss_scale": 4.0, "retries": 0, "seconds": 79.5, "max_memory_gb": 16.71, "ce": 0.044825607270468026, "kl": 0.08244425214070361, "context": 0.0860477332607843}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.8545299172401428, "loss_scale": 4.0, "retries": 0, "seconds": 79.9, "max_memory_gb": 16.71, "ce": 0.4289486824563937, "kl": 0.01345342428976437, "context": 0.43567540602816734}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.47611570358276367, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 16.71, "ce": 0.03903971931140404, "kl": 0.023571549403641257, "context": 0.05082549389771884}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.5193780064582825, "loss_scale": 4.0, "retries": 0, "seconds": 80.8, "max_memory_gb": 16.71, "ce": 0.00861426406481769, "kl": 0.1349772744870279, "context": 0.07610290152661037}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.4446464776992798, "loss_scale": 4.0, "retries": 0, "seconds": 81.3, "max_memory_gb": 16.71, "ce": 0.01883307858952321, "kl": 0.2979753613435605, "context": 0.16782076298841275}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.044058188796043396, "loss_scale": 4.0, "retries": 0, "seconds": 81.8, "max_memory_gb": 16.71, "ce": 0.003362491022926406, "kl": 0.0007088639176799916, "context": 0.0037169229181017727}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.04901159927248955, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 16.71, "ce": 0.0021696796393371187, "kl": 0.0027329839367666864, "context": 0.0035361715854378417}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.47745996713638306, "loss_scale": 4.0, "retries": 0, "seconds": 82.8, "max_memory_gb": 16.71, "ce": 0.009616556228138506, "kl": 0.12086866940080654, "context": 0.07005089096492156}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.643538236618042, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 16.71, "ce": 0.12910731192096137, "kl": 0.26206237154838163, "context": 0.2601385050002136}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 0.7913607954978943, "loss_scale": 4.0, "retries": 0, "seconds": 83.7, "max_memory_gb": 16.71, "ce": 0.049782474408857524, "kl": 0.47572728982777335, "context": 0.28764611075166613}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.180463433265686, "loss_scale": 4.0, "retries": 0, "seconds": 84.2, "max_memory_gb": 16.71, "ce": 0.25717698784865206, "kl": 0.018517542281642818, "context": 0.2664357487410598}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.0000441074371338, "loss_scale": 4.0, "retries": 0, "seconds": 84.6, "max_memory_gb": 16.71, "ce": 0.2927595311193727, "kl": 0.2647052349639125, "context": 0.42511215931153856}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.09843477606773376, "loss_scale": 4.0, "retries": 0, "seconds": 85.0, "max_memory_gb": 16.71, "ce": 0.0018258185591548681, "kl": 0.010077834594994783, "context": 0.006864735827548429}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 1.4671674966812134, "loss_scale": 4.0, "retries": 0, "seconds": 85.5, "max_memory_gb": 16.71, "ce": 0.3119582537910901, "kl": 0.1817436372384691, "context": 0.4028300869103987}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.269222617149353, "loss_scale": 4.0, "retries": 0, "seconds": 85.9, "max_memory_gb": 16.71, "ce": 0.5636607748747338, "kl": 0.09788094550458482, "context": 0.6126012234017253}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.07623818516731262, "loss_scale": 4.0, "retries": 0, "seconds": 86.3, "max_memory_gb": 16.71, "ce": 0.0049179950437974185, "kl": 0.0031530402739008423, "context": 0.006494515320810024}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.4359196126461029, "loss_scale": 4.0, "retries": 0, "seconds": 86.8, "max_memory_gb": 16.71, "ce": 0.049083784222602844, "kl": 0.07084876646695193, "context": 0.08450817118864506}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.09845951944589615, "loss_scale": 4.0, "retries": 0, "seconds": 87.2, "max_memory_gb": 16.71, "ce": 0.008838967332849279, "kl": 0.0037996953178662807, "context": 0.01073881532647647}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 1.4749562740325928, "loss_scale": 4.0, "retries": 0, "seconds": 87.7, "max_memory_gb": 16.71, "ce": 0.001243407568836119, "kl": 0.4389821417932467, "context": 0.22073447624279652}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.3682529926300049, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 16.71, "ce": 0.060160838984302245, "kl": 0.7639652977741207, "context": 0.44214347773231566}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.26359426975250244, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 16.71, "ce": 0.07070977652620059, "kl": 0.009663644486863632, "context": 0.07554159691790119}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 1.761228084564209, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 16.71, "ce": 0.1747294524757308, "kl": 0.04913011817552615, "context": 0.19929451344069093}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.7693033218383789, "loss_scale": 4.0, "retries": 0, "seconds": 89.5, "max_memory_gb": 16.71, "ce": 0.296380577608943, "kl": 0.03837903249950614, "context": 0.315570087346714}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.38673102855682373, "loss_scale": 4.0, "retries": 0, "seconds": 89.9, "max_memory_gb": 16.71, "ce": 0.07155851217976306, "kl": 0.11776564522369881, "context": 0.13044133846415207}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.05055180937051773, "loss_scale": 4.0, "retries": 0, "seconds": 90.3, "max_memory_gb": 16.71, "ce": 0.0019501004753692541, "kl": 0.0033988879138178163, "context": 0.003649544552899897}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.416126549243927, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 16.71, "ce": 0.011354575064615346, "kl": 0.017855685873655602, "context": 0.020282418001443148}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.4761333465576172, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 16.71, "ce": 0.03969378164038062, "kl": 0.009160347166471183, "context": 0.044273955281823874}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 1.0375559329986572, "loss_scale": 4.0, "retries": 0, "seconds": 91.7, "max_memory_gb": 16.71, "ce": 0.002855285791156348, "kl": 0.42509326541767223, "context": 0.21540192337124608}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.08228770643472672, "loss_scale": 4.0, "retries": 0, "seconds": 92.2, "max_memory_gb": 16.71, "ce": 0.004621352389222011, "kl": 0.002827683219948085, "context": 0.0060351940264808945}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.16019822657108307, "loss_scale": 4.0, "retries": 0, "seconds": 92.6, "max_memory_gb": 16.71, "ce": 0.06152741461119149, "kl": 0.04397523097577505, "context": 0.08351503382436931}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.07645486295223236, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 16.71, "ce": 0.004391890070110094, "kl": 0.0022215715689526405, "context": 0.005502675703610294}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.7476220726966858, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 16.71, "ce": 0.178247246745741, "kl": 0.07938163804647047, "context": 0.2179380698944442}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 1.0427380800247192, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 16.71, "ce": 0.5398981235193787, "kl": 0.1377046208435786, "context": 0.608750427069026}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.43511027097702026, "loss_scale": 4.0, "retries": 0, "seconds": 94.4, "max_memory_gb": 16.71, "ce": 0.10635353380348533, "kl": 0.13491400322527625, "context": 0.17381052812561393}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 1.3100106716156006, "loss_scale": 4.0, "retries": 0, "seconds": 94.9, "max_memory_gb": 16.71, "ce": 0.05497748948982917, "kl": 0.9837258050720266, "context": 0.546840381633956}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.751321017742157, "loss_scale": 4.0, "retries": 0, "seconds": 95.3, "max_memory_gb": 16.71, "ce": 0.25982198168640025, "kl": 0.04257186244285549, "context": 0.28110791107610567}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 2.266218662261963, "loss_scale": 4.0, "retries": 0, "seconds": 95.8, "max_memory_gb": 16.71, "ce": 0.40552738566475455, "kl": 0.14177078567445278, "context": 0.4764127812231891}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.3241349458694458, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 16.71, "ce": 0.6486403101444012, "kl": 0.029734225485526622, "context": 0.6635074503865326}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.7546284198760986, "loss_scale": 4.0, "retries": 0, "seconds": 96.7, "max_memory_gb": 16.71, "ce": 0.008609694014012348, "kl": 0.14874649154080544, "context": 0.08298294222913682}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.03741403669118881, "loss_scale": 4.0, "retries": 0, "seconds": 97.1, "max_memory_gb": 16.71, "ce": 0.001924399723066017, "kl": 0.0017690397426122217, "context": 0.002808919583912939}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.7428407073020935, "loss_scale": 4.0, "retries": 0, "seconds": 97.6, "max_memory_gb": 16.71, "ce": 0.15308568067848682, "kl": 0.23455457389354706, "context": 0.2703629657626152}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.296287477016449, "loss_scale": 4.0, "retries": 0, "seconds": 98.0, "max_memory_gb": 16.71, "ce": 0.008535549881344195, "kl": 0.11415630863120896, "context": 0.06561370413692202}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.4003655016422272, "loss_scale": 4.0, "retries": 0, "seconds": 98.5, "max_memory_gb": 16.71, "ce": 0.03625480303890072, "kl": 0.005281375062622828, "context": 0.03889549011364579}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 1.352281093597412, "loss_scale": 4.0, "retries": 0, "seconds": 98.9, "max_memory_gb": 16.71, "ce": 0.0049794182705227286, "kl": 0.40758624994850834, "context": 0.20877254553488456}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 0.8316126465797424, "loss_scale": 4.0, "retries": 0, "seconds": 99.4, "max_memory_gb": 16.71, "ce": 0.03155179225723259, "kl": 0.0663077851113485, "context": 0.06470568527583964}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 1.6795313358306885, "loss_scale": 4.0, "retries": 0, "seconds": 99.8, "max_memory_gb": 16.71, "ce": 0.09228127076858073, "kl": 0.011393118536943803, "context": 0.09797783142857952}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.922089159488678, "loss_scale": 4.0, "retries": 0, "seconds": 100.3, "max_memory_gb": 16.71, "ce": 0.04135878733359277, "kl": 0.12818968301326095, "context": 0.10545363259734586}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 0.7243639826774597, "loss_scale": 4.0, "retries": 0, "seconds": 100.7, "max_memory_gb": 16.71, "ce": 0.002670426416443661, "kl": 0.1342218959944148, "context": 0.06978137584519573}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.7595767974853516, "loss_scale": 4.0, "retries": 0, "seconds": 101.2, "max_memory_gb": 16.71, "ce": 0.039773064461769536, "kl": 0.031141035186010413, "context": 0.05534358086879365}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.2559490203857422, "loss_scale": 4.0, "retries": 0, "seconds": 101.6, "max_memory_gb": 16.71, "ce": 0.03021840141445864, "kl": 0.013403967200702027, "context": 0.03692038501321804}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 1.0401334762573242, "loss_scale": 4.0, "retries": 0, "seconds": 102.0, "max_memory_gb": 16.71, "ce": 0.0034088095162587706, "kl": 0.16077539324760437, "context": 0.08379650418646634}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.40622901916503906, "loss_scale": 4.0, "retries": 0, "seconds": 102.5, "max_memory_gb": 16.71, "ce": 0.0639987960603321, "kl": 0.025303671811116146, "context": 0.07665063382592052}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.8644514083862305, "loss_scale": 4.0, "retries": 0, "seconds": 102.9, "max_memory_gb": 16.71, "ce": 0.5671948347880971, "kl": 0.3379666559631005, "context": 0.7361781593644992}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.554814875125885, "loss_scale": 4.0, "retries": 0, "seconds": 103.3, "max_memory_gb": 16.71, "ce": 0.03927257026953157, "kl": 0.15143142243323382, "context": 0.1149882786849048}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.4297504425048828, "loss_scale": 4.0, "retries": 0, "seconds": 103.7, "max_memory_gb": 16.71, "ce": 0.023474101879401132, "kl": 0.01196859165065689, "context": 0.029458398115821183}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.5879163146018982, "loss_scale": 4.0, "retries": 0, "seconds": 104.2, "max_memory_gb": 16.71, "ce": 0.006988727313000709, "kl": 0.11971918802009895, "context": 0.06684832088649273}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.3307815492153168, "loss_scale": 4.0, "retries": 0, "seconds": 104.6, "max_memory_gb": 16.71, "ce": 0.00797227508155629, "kl": 0.03647392056154786, "context": 0.026209234725683928}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.5439237356185913, "loss_scale": 4.0, "retries": 0, "seconds": 105.1, "max_memory_gb": 16.71, "ce": 0.46142754128959496, "kl": 0.02313649763163994, "context": 0.47299579583341256}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.17799057066440582, "loss_scale": 4.0, "retries": 0, "seconds": 105.5, "max_memory_gb": 16.71, "ce": 0.009183914822642691, "kl": 0.01849630217839149, "context": 0.01843206601915881}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.2599031329154968, "loss_scale": 4.0, "retries": 0, "seconds": 106.0, "max_memory_gb": 16.71, "ce": 0.06663757175556384, "kl": 0.029207479503781997, "context": 0.08124130958458409}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.25375667214393616, "loss_scale": 4.0, "retries": 0, "seconds": 106.4, "max_memory_gb": 16.71, "ce": 0.021733979774580803, "kl": 0.008173513910151087, "context": 0.025820736918831244}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.2965341806411743, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 16.71, "ce": 0.097240423085168, "kl": 0.02391043478564825, "context": 0.10919563670177013}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 1.0109524726867676, "loss_scale": 4.0, "retries": 0, "seconds": 107.2, "max_memory_gb": 16.71, "ce": 0.1069060717272805, "kl": 0.046794740854238626, "context": 0.13030344530125149}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 1.1630357503890991, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 16.71, "ce": 0.2056777434045216, "kl": 0.028419492031389382, "context": 0.21988749648153316}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.19477276504039764, "loss_scale": 4.0, "retries": 0, "seconds": 108.1, "max_memory_gb": 16.71, "ce": 0.003290535882115364, "kl": 0.026884311171102127, "context": 0.016732691816287115}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.06715390086174011, "loss_scale": 4.0, "retries": 0, "seconds": 108.6, "max_memory_gb": 16.71, "ce": 0.003166153939673677, "kl": 0.0041159137326758355, "context": 0.005224110733252019}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 0.5538956522941589, "loss_scale": 4.0, "retries": 0, "seconds": 109.0, "max_memory_gb": 16.71, "ce": 0.03460235166858183, "kl": 0.13033297130459687, "context": 0.09976883883064147}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 0.9753162860870361, "loss_scale": 4.0, "retries": 0, "seconds": 109.5, "max_memory_gb": 16.71, "ce": 0.3950878292089328, "kl": 0.062057274975813925, "context": 0.4261164660565555}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.016602231189608574, "loss_scale": 4.0, "retries": 0, "seconds": 109.9, "max_memory_gb": 16.71, "ce": 0.001144015957834199, "kl": 0.0004790911129930464, "context": 0.0013835614954587072}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 0.8274972438812256, "loss_scale": 4.0, "retries": 0, "seconds": 110.4, "max_memory_gb": 16.71, "ce": 0.003601079515647143, "kl": 0.46074639431390096, "context": 0.233974274204229}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.2989153563976288, "loss_scale": 4.0, "retries": 0, "seconds": 110.8, "max_memory_gb": 16.71, "ce": 0.0247425195411779, "kl": 0.008599508670158684, "context": 0.029042274341918528}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 1.4432365894317627, "loss_scale": 4.0, "retries": 0, "seconds": 111.3, "max_memory_gb": 16.71, "ce": 0.05559809447731823, "kl": 0.0712228994234465, "context": 0.09120954782702029}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.0649186298251152, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 16.71, "ce": 0.0064979665039572865, "kl": 0.0031367175397463143, "context": 0.0080663250409998}
|
runs/single_internvl/context/L34_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/context/L34_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.34.mlp.down_proj",
|
| 35 |
+
"model.language_model.layers.34.self_attn.q_proj",
|
| 36 |
+
"model.language_model.layers.34.mlp.gate_proj",
|
| 37 |
+
"model.language_model.layers.34.self_attn.o_proj",
|
| 38 |
+
"model.language_model.layers.34.mlp.up_proj",
|
| 39 |
+
"model.language_model.layers.34.self_attn.k_proj",
|
| 40 |
+
"model.language_model.layers.34.self_attn.v_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/context/L34_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "context",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/context.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 34,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/context/L34_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.37761035561561584, "loss_scale": 4.0, "retries": 0, "seconds": 1.3, "max_memory_gb": 16.47, "ce": 0.5227766726748087, "kl": 0.08654405153356493, "context": 0.5660487064160407}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.5186718106269836, "loss_scale": 4.0, "retries": 0, "seconds": 1.7, "max_memory_gb": 16.47, "ce": 0.6404395870849839, "kl": 0.4683591094252506, "context": 0.874619144036842}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 1.0828431844711304, "loss_scale": 4.0, "retries": 0, "seconds": 2.0, "max_memory_gb": 16.47, "ce": 1.0941549465060234, "kl": 0.06738869473338127, "context": 1.127849280834198}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 0.6129411458969116, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 16.47, "ce": 0.45207827698322944, "kl": 0.42941547930308843, "context": 0.666786012268858}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.013608083128929138, "loss_scale": 4.0, "retries": 0, "seconds": 2.8, "max_memory_gb": 16.47, "ce": 0.009195431375701446, "kl": 0.00819971573582734, "context": 0.013295289332745597}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.3128042221069336, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 16.47, "ce": 0.29097037747851573, "kl": 0.0746842822409235, "context": 0.3283125229063444}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 0.8752612471580505, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 16.47, "ce": 0.8299291558323603, "kl": 0.12237092932673477, "context": 0.8911146149075648}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 1.103765606880188, "loss_scale": 4.0, "retries": 0, "seconds": 3.9, "max_memory_gb": 16.47, "ce": 0.7970096506178379, "kl": 0.2904033618979156, "context": 0.9422113485634327}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 0.7307001948356628, "loss_scale": 4.0, "retries": 0, "seconds": 4.3, "max_memory_gb": 16.47, "ce": 0.529319217428565, "kl": 0.13603398314444348, "context": 0.5973362177610397}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.14946189522743225, "loss_scale": 4.0, "retries": 0, "seconds": 4.6, "max_memory_gb": 16.47, "ce": 0.05406907516589854, "kl": 0.04538091190624982, "context": 0.07675953186117113}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 0.9589400291442871, "loss_scale": 4.0, "retries": 0, "seconds": 5.0, "max_memory_gb": 16.47, "ce": 0.707253715780098, "kl": 0.004943941062265367, "context": 0.7097257009372697}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 0.6590546369552612, "loss_scale": 4.0, "retries": 0, "seconds": 5.4, "max_memory_gb": 16.47, "ce": 0.47689084079775057, "kl": 0.002505653024854837, "context": 0.47814365851445473}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 0.6652080416679382, "loss_scale": 4.0, "retries": 0, "seconds": 5.7, "max_memory_gb": 16.47, "ce": 0.4309131596091902, "kl": 0.01905656784583698, "context": 0.44044144966755994}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 1.219215989112854, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 16.47, "ce": 0.7773014008998871, "kl": 0.9423743607476354, "context": 1.2484886199235916}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 1.0410902500152588, "loss_scale": 4.0, "retries": 0, "seconds": 6.4, "max_memory_gb": 16.47, "ce": 0.5317166905338127, "kl": 0.46841133894531595, "context": 0.7659223339294385}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 0.5452935099601746, "loss_scale": 4.0, "retries": 0, "seconds": 6.8, "max_memory_gb": 16.47, "ce": 0.20521015085250838, "kl": 0.0009570344000167097, "context": 0.20568866158282617}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 0.7989943623542786, "loss_scale": 4.0, "retries": 0, "seconds": 7.2, "max_memory_gb": 16.47, "ce": 0.3286481923132669, "kl": 0.8176099721167702, "context": 0.7374531725654379}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.458905577659607, "loss_scale": 4.0, "retries": 0, "seconds": 7.5, "max_memory_gb": 16.47, "ce": 0.5464778242021566, "kl": 0.2091637388075469, "context": 0.6510596894077025}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.792428970336914, "loss_scale": 4.0, "retries": 0, "seconds": 7.9, "max_memory_gb": 16.47, "ce": 0.8029928058385849, "kl": 0.16543459240347147, "context": 0.8857101127505302}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.021210303530097008, "loss_scale": 4.0, "retries": 0, "seconds": 8.2, "max_memory_gb": 16.47, "ce": 0.0023261074329639086, "kl": 0.021195027917201514, "context": 0.012923620975925587}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.5256315469741821, "loss_scale": 4.0, "retries": 0, "seconds": 8.6, "max_memory_gb": 16.47, "ce": 0.3759386041592734, "kl": 0.05062838021240168, "context": 0.40125279519634205}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 1.6551647186279297, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 16.47, "ce": 0.4098642658937024, "kl": 0.027115583594422787, "context": 0.4234220684011234}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.134916305541992, "loss_scale": 4.0, "retries": 0, "seconds": 9.3, "max_memory_gb": 16.47, "ce": 0.5732468810547289, "kl": 0.980443716443574, "context": 1.0634687118026704}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 1.6743794679641724, "loss_scale": 4.0, "retries": 0, "seconds": 9.6, "max_memory_gb": 16.47, "ce": 0.2790686786174774, "kl": 0.3385501215234399, "context": 0.4483437482267618}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 2.6279423236846924, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 16.47, "ce": 0.6429021182066208, "kl": 0.21956343296915293, "context": 0.7526838295161724}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 2.2995517253875732, "loss_scale": 4.0, "retries": 0, "seconds": 10.4, "max_memory_gb": 16.47, "ce": 0.31581367179751396, "kl": 0.6854472709819674, "context": 0.6585372984409332}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 2.2470264434814453, "loss_scale": 4.0, "retries": 0, "seconds": 10.7, "max_memory_gb": 16.47, "ce": 0.42172479729924817, "kl": 0.41442567374542705, "context": 0.6289376515196636}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 1.3578767776489258, "loss_scale": 4.0, "retries": 0, "seconds": 11.1, "max_memory_gb": 16.47, "ce": 0.586861427873373, "kl": 0.9242429820587859, "context": 1.0489829033613205}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 1.9864777326583862, "loss_scale": 4.0, "retries": 0, "seconds": 11.5, "max_memory_gb": 16.47, "ce": 0.35618973802775145, "kl": 0.3606715979985893, "context": 0.5365255456417799}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 2.278364896774292, "loss_scale": 4.0, "retries": 0, "seconds": 11.8, "max_memory_gb": 16.47, "ce": 0.30045791371412633, "kl": 0.009465701120461745, "context": 0.30519075582742516}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 3.0495870113372803, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 16.47, "ce": 0.6068007834255695, "kl": 0.15922486782073975, "context": 0.6864132359623909}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 3.303603172302246, "loss_scale": 4.0, "retries": 0, "seconds": 12.6, "max_memory_gb": 16.47, "ce": 0.32879478484392166, "kl": 0.01751886901911348, "context": 0.3375542163848877}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 1.644532322883606, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 16.47, "ce": 0.1716457987204194, "kl": 0.06106753833591938, "context": 0.2021795753389597}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 1.4247065782546997, "loss_scale": 4.0, "retries": 0, "seconds": 13.4, "max_memory_gb": 16.47, "ce": 0.09061342109998805, "kl": 0.7398322010240008, "context": 0.4605295068286068}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 1.5644445419311523, "loss_scale": 4.0, "retries": 0, "seconds": 13.8, "max_memory_gb": 16.47, "ce": 0.11415828792087268, "kl": 0.004089853685400158, "context": 0.11620321893133223}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 1.680168628692627, "loss_scale": 4.0, "retries": 0, "seconds": 14.2, "max_memory_gb": 16.47, "ce": 0.24066647718427703, "kl": 0.04151513797114603, "context": 0.26142404349229764}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 1.2988855838775635, "loss_scale": 4.0, "retries": 0, "seconds": 14.6, "max_memory_gb": 16.47, "ce": 0.20160232647322118, "kl": 0.7391269218642265, "context": 0.5711657957872376}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 1.5227571725845337, "loss_scale": 4.0, "retries": 0, "seconds": 15.0, "max_memory_gb": 16.47, "ce": 0.13600982073694468, "kl": 0.5118064093985595, "context": 0.39191302191466093}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 0.27973541617393494, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 16.47, "ce": 0.1229011340692523, "kl": 0.009806103291339241, "context": 0.1278041836048942}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 0.38713106513023376, "loss_scale": 4.0, "retries": 0, "seconds": 15.8, "max_memory_gb": 16.47, "ce": 0.02480542630655691, "kl": 0.007932591193821281, "context": 0.028771722689270973}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 0.6594567894935608, "loss_scale": 4.0, "retries": 0, "seconds": 16.1, "max_memory_gb": 16.47, "ce": 0.3783110890071839, "kl": 0.12381945899687707, "context": 0.44022082770243287}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 1.0558871030807495, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 16.47, "ce": 0.06884965513745556, "kl": 0.010017899410740938, "context": 0.07385860378417419}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 0.9899312257766724, "loss_scale": 4.0, "retries": 0, "seconds": 16.9, "max_memory_gb": 16.47, "ce": 0.03732533962465823, "kl": 0.04601217378512956, "context": 0.06033142654632684}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 1.5511966943740845, "loss_scale": 4.0, "retries": 0, "seconds": 17.3, "max_memory_gb": 16.47, "ce": 0.5580377080113976, "kl": 1.3266719271414331, "context": 1.2213736734556733}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 1.4160007238388062, "loss_scale": 4.0, "retries": 0, "seconds": 17.7, "max_memory_gb": 16.47, "ce": 0.8767894050979521, "kl": 0.21264849188446533, "context": 0.983113668160513}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 0.3288963735103607, "loss_scale": 4.0, "retries": 0, "seconds": 18.0, "max_memory_gb": 16.47, "ce": 0.08710252097807825, "kl": 0.25664880813565105, "context": 0.21542692324146628}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 0.3190377354621887, "loss_scale": 4.0, "retries": 0, "seconds": 18.4, "max_memory_gb": 16.47, "ce": 0.012811800683266483, "kl": 0.026383135351352394, "context": 0.0260033673257567}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 0.7356830835342407, "loss_scale": 4.0, "retries": 0, "seconds": 18.8, "max_memory_gb": 16.47, "ce": 0.027977079793345183, "kl": 0.008884651942935307, "context": 0.03241940477164462}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 1.2481225728988647, "loss_scale": 4.0, "retries": 0, "seconds": 19.2, "max_memory_gb": 16.47, "ce": 0.03372259564639535, "kl": 0.25314849342180423, "context": 0.16029684049499338}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 1.3849934339523315, "loss_scale": 4.0, "retries": 0, "seconds": 19.6, "max_memory_gb": 16.47, "ce": 0.09054279038537061, "kl": 0.01462170206650626, "context": 0.09785364205890801}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 1.1050041913986206, "loss_scale": 4.0, "retries": 0, "seconds": 20.0, "max_memory_gb": 16.47, "ce": 0.5346348782659334, "kl": 0.021364169223176077, "context": 0.5453169430802518}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 0.6214925050735474, "loss_scale": 4.0, "retries": 0, "seconds": 20.3, "max_memory_gb": 16.47, "ce": 0.024725373863475397, "kl": 0.22056901163887233, "context": 0.1350098834373057}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 1.3009779453277588, "loss_scale": 4.0, "retries": 0, "seconds": 20.7, "max_memory_gb": 16.47, "ce": 0.18340773973613977, "kl": 0.32028638990595937, "context": 0.3435509353876114}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 0.7596674561500549, "loss_scale": 4.0, "retries": 0, "seconds": 21.1, "max_memory_gb": 16.47, "ce": 0.04197191438288428, "kl": 0.3070386714389315, "context": 0.1954912464134395}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 0.2958359122276306, "loss_scale": 4.0, "retries": 0, "seconds": 21.4, "max_memory_gb": 16.47, "ce": 0.01915532586281188, "kl": 0.013050557648057293, "context": 0.025680605875095353}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.6505310535430908, "loss_scale": 4.0, "retries": 0, "seconds": 21.8, "max_memory_gb": 16.47, "ce": 0.026160964975133538, "kl": 0.22079393197782338, "context": 0.13655793317593634}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 1.0424697399139404, "loss_scale": 4.0, "retries": 0, "seconds": 22.2, "max_memory_gb": 16.47, "ce": 0.03314849469461478, "kl": 0.03183406507378095, "context": 0.049065525410696864}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 0.05739349126815796, "loss_scale": 4.0, "retries": 0, "seconds": 22.6, "max_memory_gb": 16.47, "ce": 0.003399612905923277, "kl": 0.0016180769915834503, "context": 0.004208651473163627}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 1.015641689300537, "loss_scale": 4.0, "retries": 0, "seconds": 23.0, "max_memory_gb": 16.47, "ce": 0.03994853119365871, "kl": 0.05865753337275237, "context": 0.0692772978800349}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 1.6636821031570435, "loss_scale": 4.0, "retries": 0, "seconds": 23.3, "max_memory_gb": 16.47, "ce": 0.06407979904906824, "kl": 0.05871231440687552, "context": 0.09343595802783966}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 1.6113951206207275, "loss_scale": 4.0, "retries": 0, "seconds": 23.7, "max_memory_gb": 16.47, "ce": 0.023819448659196496, "kl": 0.8361918111331761, "context": 0.4419153684284538}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 0.18082670867443085, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 16.47, "ce": 0.03501736468751915, "kl": 0.009694977670733351, "context": 0.039864853548351675}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 1.0709245204925537, "loss_scale": 4.0, "retries": 0, "seconds": 24.5, "max_memory_gb": 16.47, "ce": 0.04108122136676684, "kl": 0.07051847489492502, "context": 0.0763404592871666}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.2672809362411499, "loss_scale": 4.0, "retries": 0, "seconds": 24.8, "max_memory_gb": 16.47, "ce": 0.016198195808101445, "kl": 0.012789112370228395, "context": 0.022592751774936914}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 1.5865362882614136, "loss_scale": 4.0, "retries": 0, "seconds": 25.2, "max_memory_gb": 16.47, "ce": 0.2636011531576514, "kl": 0.11128308088518679, "context": 0.3192427009344101}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 2.257746696472168, "loss_scale": 4.0, "retries": 0, "seconds": 25.6, "max_memory_gb": 16.47, "ce": 0.16411478790632827, "kl": 0.07415785459784274, "context": 0.20119370774000345}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.16051939129829407, "loss_scale": 4.0, "retries": 0, "seconds": 26.0, "max_memory_gb": 16.47, "ce": 0.00829287149826996, "kl": 0.010847485915292054, "context": 0.013716614805161953}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 0.1115679070353508, "loss_scale": 4.0, "retries": 0, "seconds": 26.4, "max_memory_gb": 16.47, "ce": 0.008297624244733015, "kl": 0.04493565726079396, "context": 0.030765453542699106}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 0.7930982708930969, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 16.47, "ce": 0.25218116329051554, "kl": 0.029988321155542508, "context": 0.26717532600741833}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 0.055742401629686356, "loss_scale": 4.0, "retries": 0, "seconds": 27.2, "max_memory_gb": 16.47, "ce": 0.00292969866131898, "kl": 0.0006226273399079219, "context": 0.003241012367652729}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 0.42156440019607544, "loss_scale": 4.0, "retries": 0, "seconds": 27.5, "max_memory_gb": 16.47, "ce": 0.02006226755111129, "kl": 0.012907995289424434, "context": 0.026516265017562546}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 0.8080861568450928, "loss_scale": 4.0, "retries": 0, "seconds": 27.9, "max_memory_gb": 16.47, "ce": 0.019341501174494624, "kl": 0.4133800126146525, "context": 0.22603151015937328}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 0.9471169710159302, "loss_scale": 4.0, "retries": 0, "seconds": 28.3, "max_memory_gb": 16.47, "ce": 0.3684332705233828, "kl": 0.11612568562850356, "context": 0.4264961263979785}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 0.270221471786499, "loss_scale": 4.0, "retries": 0, "seconds": 28.7, "max_memory_gb": 16.47, "ce": 0.07602135361776163, "kl": 0.10834721320134122, "context": 0.13019496022025123}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.0820882320404053, "loss_scale": 4.0, "retries": 0, "seconds": 29.1, "max_memory_gb": 16.47, "ce": 0.4637548547470942, "kl": 0.9706895519047976, "context": 0.9490996268577874}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.7432626485824585, "loss_scale": 4.0, "retries": 0, "seconds": 29.4, "max_memory_gb": 16.47, "ce": 0.005576135183218867, "kl": 1.531206208979711, "context": 0.7711792411282659}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 0.3452589809894562, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 16.47, "ce": 0.01312896772287786, "kl": 0.014786228392040357, "context": 0.020522082384559326}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 0.9154940843582153, "loss_scale": 4.0, "retries": 0, "seconds": 30.2, "max_memory_gb": 16.47, "ce": 0.0019845255301333964, "kl": 0.08114306336210575, "context": 0.04255605611251667}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 0.26931676268577576, "loss_scale": 4.0, "retries": 0, "seconds": 30.5, "max_memory_gb": 16.47, "ce": 0.025921085327354376, "kl": 0.04202663330488576, "context": 0.046934401980252005}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 1.0699714422225952, "loss_scale": 4.0, "retries": 0, "seconds": 30.9, "max_memory_gb": 16.47, "ce": 0.005854386288774549, "kl": 0.11774213060067495, "context": 0.06472545211363467}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.8156082630157471, "loss_scale": 4.0, "retries": 0, "seconds": 31.2, "max_memory_gb": 16.47, "ce": 0.08651831519091502, "kl": 0.06888969092688058, "context": 0.12096315552480519}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 0.11501165479421616, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 16.47, "ce": 0.011510388692840934, "kl": 0.003772030917389202, "context": 0.013396404567174613}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 2.092132091522217, "loss_scale": 4.0, "retries": 0, "seconds": 32.0, "max_memory_gb": 16.47, "ce": 0.12405738295638002, "kl": 0.634978319700167, "context": 0.44154655071906745}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 1.3233458995819092, "loss_scale": 4.0, "retries": 0, "seconds": 32.4, "max_memory_gb": 16.47, "ce": 0.049353984278241114, "kl": 0.04187541298870201, "context": 0.07029169155066484}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 0.8251069784164429, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 16.47, "ce": 0.07316299463764153, "kl": 0.046257248103245274, "context": 0.09629161868906522}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 1.5796597003936768, "loss_scale": 4.0, "retries": 0, "seconds": 33.1, "max_memory_gb": 16.47, "ce": 0.22184788656886667, "kl": 0.40838580249692313, "context": 0.42604078841395676}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 0.9872198700904846, "loss_scale": 4.0, "retries": 0, "seconds": 33.5, "max_memory_gb": 16.47, "ce": 0.009165915922494605, "kl": 0.7055542243178934, "context": 0.3619430218823254}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 0.25431588292121887, "loss_scale": 4.0, "retries": 0, "seconds": 33.9, "max_memory_gb": 16.47, "ce": 0.012707001244052663, "kl": 0.01959943644510531, "context": 0.022506719917146256}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 0.07689652591943741, "loss_scale": 4.0, "retries": 0, "seconds": 34.2, "max_memory_gb": 16.47, "ce": 0.014672392970169312, "kl": 0.011188653214048827, "context": 0.020266718522179872}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 0.1032879427075386, "loss_scale": 4.0, "retries": 0, "seconds": 34.6, "max_memory_gb": 16.47, "ce": 0.010993881296599284, "kl": 0.008115267126413528, "context": 0.015051514375954866}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 0.6612429618835449, "loss_scale": 4.0, "retries": 0, "seconds": 35.0, "max_memory_gb": 16.47, "ce": 0.1581683697659173, "kl": 0.5502810619436787, "context": 0.4333089053980075}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 1.5353977680206299, "loss_scale": 4.0, "retries": 0, "seconds": 35.4, "max_memory_gb": 16.47, "ce": 0.37572451653977623, "kl": 0.42926444474142045, "context": 0.5903567541463417}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 0.8907526731491089, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 16.47, "ce": 0.07501881229109131, "kl": 0.45712568430099054, "context": 0.30358166160294786}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 1.1790088415145874, "loss_scale": 4.0, "retries": 0, "seconds": 36.2, "max_memory_gb": 16.47, "ce": 0.2124678185209632, "kl": 0.09898994903778657, "context": 0.2619627930398565}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 1.687898874282837, "loss_scale": 4.0, "retries": 0, "seconds": 36.5, "max_memory_gb": 16.47, "ce": 0.5213842554658186, "kl": 0.026118438225239515, "context": 0.5344434762373567}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.1689647436141968, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 16.47, "ce": 0.41177474657888524, "kl": 0.004433104957570322, "context": 0.4139912980608642}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 0.9632822275161743, "loss_scale": 4.0, "retries": 0, "seconds": 37.3, "max_memory_gb": 16.47, "ce": 0.04259736306266859, "kl": 0.4333916368195787, "context": 0.2592931818217039}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 0.6146625876426697, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 16.47, "ce": 0.00371652607282158, "kl": 0.08380205888897763, "context": 0.04561755574832205}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 0.39156803488731384, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 16.47, "ce": 0.04217839246848598, "kl": 0.013318693287146743, "context": 0.04883774027985055}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 1.103805422782898, "loss_scale": 4.0, "retries": 0, "seconds": 38.4, "max_memory_gb": 16.47, "ce": 0.058493990916758776, "kl": 0.5362830762969679, "context": 0.3266355280065909}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 0.28458812832832336, "loss_scale": 4.0, "retries": 0, "seconds": 38.8, "max_memory_gb": 16.47, "ce": 0.009554541087709367, "kl": 0.09514158536330797, "context": 0.05712533416226506}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 0.7398524284362793, "loss_scale": 4.0, "retries": 0, "seconds": 39.2, "max_memory_gb": 16.47, "ce": 0.03467882494442165, "kl": 0.015646414132788777, "context": 0.04250203096307814}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.3240455389022827, "loss_scale": 4.0, "retries": 0, "seconds": 39.6, "max_memory_gb": 16.47, "ce": 0.020816633943468332, "kl": 0.03899145418108674, "context": 0.04031236085575074}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 0.23146109282970428, "loss_scale": 4.0, "retries": 0, "seconds": 39.9, "max_memory_gb": 16.47, "ce": 0.008133341238135472, "kl": 0.0204123499988782, "context": 0.018339516020205338}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 1.2316596508026123, "loss_scale": 4.0, "retries": 0, "seconds": 40.3, "max_memory_gb": 16.47, "ce": 0.08627842932764906, "kl": 0.28396609362971503, "context": 0.22826147021260113}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 0.07407756894826889, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 16.47, "ce": 0.004106615291675553, "kl": 0.0011316654217807809, "context": 0.004672448048950173}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.3115715980529785, "loss_scale": 4.0, "retries": 0, "seconds": 41.1, "max_memory_gb": 16.47, "ce": 0.46876133265323006, "kl": 0.08135742007084446, "context": 0.5094400303642033}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 0.3444644808769226, "loss_scale": 4.0, "retries": 0, "seconds": 41.4, "max_memory_gb": 16.47, "ce": 0.03545179047068814, "kl": 0.002393407071394904, "context": 0.03664849404594861}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 2.1307523250579834, "loss_scale": 4.0, "retries": 0, "seconds": 41.8, "max_memory_gb": 16.47, "ce": 0.09097798162838444, "kl": 0.26249488443136215, "context": 0.2222254192456603}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 1.41319739818573, "loss_scale": 4.0, "retries": 0, "seconds": 42.2, "max_memory_gb": 16.47, "ce": 0.17683868203312159, "kl": 0.2137802231591195, "context": 0.2837287988513708}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 0.026664255186915398, "loss_scale": 4.0, "retries": 0, "seconds": 42.5, "max_memory_gb": 16.47, "ce": 0.0015646039028069936, "kl": 0.001495811087806942, "context": 0.0023125095176510513}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 0.2606724500656128, "loss_scale": 4.0, "retries": 0, "seconds": 42.9, "max_memory_gb": 16.47, "ce": 0.013972120432299562, "kl": 0.0007974577747518197, "context": 0.014370849370607175}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 1.003149390220642, "loss_scale": 4.0, "retries": 0, "seconds": 43.3, "max_memory_gb": 16.47, "ce": 0.01589862850960344, "kl": 0.21297428314574063, "context": 0.12238577241078019}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 2.2295455932617188, "loss_scale": 4.0, "retries": 0, "seconds": 43.6, "max_memory_gb": 16.47, "ce": 0.020181935295113362, "kl": 0.7746704818273429, "context": 0.40751716651720926}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 0.3809361755847931, "loss_scale": 4.0, "retries": 0, "seconds": 44.0, "max_memory_gb": 16.47, "ce": 0.018725941015873104, "kl": 0.05096399039030075, "context": 0.044207935337908566}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 1.2033956050872803, "loss_scale": 4.0, "retries": 0, "seconds": 44.4, "max_memory_gb": 16.47, "ce": 0.14461008729995228, "kl": 0.3844216663172233, "context": 0.33682090560614597}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 1.3950989246368408, "loss_scale": 4.0, "retries": 0, "seconds": 44.8, "max_memory_gb": 16.47, "ce": 0.10672100364172366, "kl": 0.10078440966026392, "context": 0.15711320110131055}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 0.4858960509300232, "loss_scale": 4.0, "retries": 0, "seconds": 45.1, "max_memory_gb": 16.47, "ce": 0.1820290530740749, "kl": 0.008557275924431451, "context": 0.18630768642469775}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 0.21822938323020935, "loss_scale": 4.0, "retries": 0, "seconds": 45.5, "max_memory_gb": 16.47, "ce": 0.033931256337382365, "kl": 0.01667159140026797, "context": 0.042267051481758244}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 2.6122398376464844, "loss_scale": 4.0, "retries": 0, "seconds": 45.9, "max_memory_gb": 16.47, "ce": 0.426505598181393, "kl": 0.15100115039967932, "context": 0.5020061662653461}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 0.3272062838077545, "loss_scale": 4.0, "retries": 0, "seconds": 46.3, "max_memory_gb": 16.47, "ce": 0.07083234805759275, "kl": 0.044951880950975465, "context": 0.09330829084501602}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 0.6482672095298767, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 16.47, "ce": 0.10292622481210856, "kl": 0.016833016234159004, "context": 0.11134273106290493}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 0.9160653948783875, "loss_scale": 4.0, "retries": 0, "seconds": 47.0, "max_memory_gb": 16.47, "ce": 0.07066653715446591, "kl": 0.029950531956274062, "context": 0.08564180415123701}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.7597171068191528, "loss_scale": 4.0, "retries": 0, "seconds": 47.4, "max_memory_gb": 16.47, "ce": 0.6099894177459646, "kl": 0.4873403614219569, "context": 0.8536596052581444}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 0.44371581077575684, "loss_scale": 4.0, "retries": 0, "seconds": 47.8, "max_memory_gb": 16.47, "ce": 0.1635266748489812, "kl": 0.07616794607019983, "context": 0.20161064231069759}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 1.5422512292861938, "loss_scale": 4.0, "retries": 0, "seconds": 48.1, "max_memory_gb": 16.47, "ce": 0.5987408012151718, "kl": 0.11619376996532083, "context": 0.6568376813083887}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 0.06374993175268173, "loss_scale": 4.0, "retries": 0, "seconds": 48.5, "max_memory_gb": 16.47, "ce": 0.0037477003497770056, "kl": 0.0011063418023695704, "context": 0.0043008712091250345}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 2.90609073638916, "loss_scale": 4.0, "retries": 0, "seconds": 48.9, "max_memory_gb": 16.47, "ce": 0.23128023685421795, "kl": 0.7317641217669006, "context": 0.5971623013028875}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 1.3704829216003418, "loss_scale": 4.0, "retries": 0, "seconds": 49.3, "max_memory_gb": 16.47, "ce": 0.0839061014703475, "kl": 0.03113722954003606, "context": 0.09947471623308957}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 0.7709740996360779, "loss_scale": 4.0, "retries": 0, "seconds": 49.6, "max_memory_gb": 16.47, "ce": 0.05090293777175248, "kl": 0.31602115221176064, "context": 0.208913512295112}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 1.1484224796295166, "loss_scale": 4.0, "retries": 0, "seconds": 50.0, "max_memory_gb": 16.47, "ce": 0.13161687966203317, "kl": 0.17231077142059803, "context": 0.21777226449921727}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 2.306229829788208, "loss_scale": 4.0, "retries": 0, "seconds": 50.4, "max_memory_gb": 16.47, "ce": 0.23444913735147566, "kl": 0.03915017869803705, "context": 0.25402422205661424}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 0.9727643132209778, "loss_scale": 4.0, "retries": 0, "seconds": 50.8, "max_memory_gb": 16.47, "ce": 0.30812741311092395, "kl": 0.03786984644830227, "context": 0.32706233114004135}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 0.6157997846603394, "loss_scale": 4.0, "retries": 0, "seconds": 51.1, "max_memory_gb": 16.47, "ce": 0.17459103028522804, "kl": 0.002068694026434059, "context": 0.17562537307094317}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 0.42000654339790344, "loss_scale": 4.0, "retries": 0, "seconds": 51.5, "max_memory_gb": 16.47, "ce": 0.0187924865167588, "kl": 0.03616737360425759, "context": 0.03687617293326184}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 1.5617010593414307, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 16.47, "ce": 0.16347522870637476, "kl": 0.17225511044671293, "context": 0.24960278428625315}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 0.23609937727451324, "loss_scale": 4.0, "retries": 0, "seconds": 52.2, "max_memory_gb": 16.47, "ce": 0.018513087939936668, "kl": 0.016684538015397266, "context": 0.026855356874875724}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 2.640674352645874, "loss_scale": 4.0, "retries": 0, "seconds": 52.6, "max_memory_gb": 16.47, "ce": 0.1533635735977441, "kl": 0.19895915407687426, "context": 0.2528431573882699}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 0.5753344893455505, "loss_scale": 4.0, "retries": 0, "seconds": 53.0, "max_memory_gb": 16.47, "ce": 0.01417516719084233, "kl": 0.17562367348728003, "context": 0.101987000904046}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 0.1508786380290985, "loss_scale": 4.0, "retries": 0, "seconds": 53.4, "max_memory_gb": 16.47, "ce": 0.009626108178053983, "kl": 0.004571168523398228, "context": 0.01191169218509458}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.09940017759799957, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 16.47, "ce": 0.007711589008977171, "kl": 0.002533894316002261, "context": 0.008978535974165425}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 0.4540635645389557, "loss_scale": 4.0, "retries": 0, "seconds": 54.2, "max_memory_gb": 16.47, "ce": 0.05035934518673457, "kl": 0.022265802705078386, "context": 0.061492245411500335}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 1.64764404296875, "loss_scale": 4.0, "retries": 0, "seconds": 54.5, "max_memory_gb": 16.47, "ce": 0.3434288035496138, "kl": 0.037122779143828666, "context": 0.36199019983178005}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 0.8765448927879333, "loss_scale": 4.0, "retries": 0, "seconds": 54.9, "max_memory_gb": 16.47, "ce": 0.08237503044074401, "kl": 0.15535728106624447, "context": 0.16005367378238589}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.10428954660892487, "loss_scale": 4.0, "retries": 0, "seconds": 55.3, "max_memory_gb": 16.47, "ce": 0.003535341063980013, "kl": 0.007715902484051185, "context": 0.0073932922969106585}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 0.35915717482566833, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 16.47, "ce": 0.03469140542438254, "kl": 0.007582697919133352, "context": 0.038482753559947014}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 1.6535667181015015, "loss_scale": 4.0, "retries": 0, "seconds": 56.1, "max_memory_gb": 16.47, "ce": 0.373264098278014, "kl": 0.4573690151446499, "context": 0.6019486117875203}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 0.2816430926322937, "loss_scale": 4.0, "retries": 0, "seconds": 56.5, "max_memory_gb": 16.47, "ce": 0.013541343956603669, "kl": 0.03600516918231733, "context": 0.03154392752912827}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 0.32513654232025146, "loss_scale": 4.0, "retries": 0, "seconds": 56.8, "max_memory_gb": 16.47, "ce": 0.0397790651768446, "kl": 0.003909030961949611, "context": 0.04173358163097873}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 0.423282265663147, "loss_scale": 4.0, "retries": 0, "seconds": 57.2, "max_memory_gb": 16.47, "ce": 0.061545148069853894, "kl": 0.01929120472050272, "context": 0.0711907485674601}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 0.3939381539821625, "loss_scale": 4.0, "retries": 0, "seconds": 57.6, "max_memory_gb": 16.47, "ce": 0.012647942290641367, "kl": 0.01364965246830252, "context": 0.019472768006380647}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 1.7087198495864868, "loss_scale": 4.0, "retries": 0, "seconds": 58.0, "max_memory_gb": 16.47, "ce": 0.014376570441527292, "kl": 0.5900594571476176, "context": 0.3094062876771204}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 1.2607839107513428, "loss_scale": 4.0, "retries": 0, "seconds": 58.4, "max_memory_gb": 16.47, "ce": 0.014673112069431227, "kl": 0.41883540638445993, "context": 0.22409081817022525}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 1.3185656070709229, "loss_scale": 4.0, "retries": 0, "seconds": 58.8, "max_memory_gb": 16.47, "ce": 0.1410541163641028, "kl": 0.08757490083371522, "context": 0.18484156604972668}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 3.389451265335083, "loss_scale": 4.0, "retries": 0, "seconds": 59.1, "max_memory_gb": 16.47, "ce": 0.25146054945798824, "kl": 0.009440482960599184, "context": 0.2561807904203306}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 0.35409489274024963, "loss_scale": 4.0, "retries": 0, "seconds": 59.5, "max_memory_gb": 16.47, "ce": 0.011382283992134035, "kl": 0.04879733971574751, "context": 0.035780953359790146}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 0.7731274962425232, "loss_scale": 4.0, "retries": 0, "seconds": 59.9, "max_memory_gb": 16.47, "ce": 0.08323618758004159, "kl": 0.0508612533158157, "context": 0.10866681789048016}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 0.9155433773994446, "loss_scale": 4.0, "retries": 0, "seconds": 60.3, "max_memory_gb": 16.47, "ce": 0.16574142553145066, "kl": 0.07312176068080589, "context": 0.20230230828747153}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 0.6482465267181396, "loss_scale": 4.0, "retries": 0, "seconds": 60.6, "max_memory_gb": 16.47, "ce": 0.017321132792858407, "kl": 0.12422748806420714, "context": 0.07943487667944282}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 1.3648407459259033, "loss_scale": 4.0, "retries": 0, "seconds": 61.0, "max_memory_gb": 16.47, "ce": 0.05285628873389214, "kl": 0.28020469000330195, "context": 0.1929586308542639}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 1.6535708904266357, "loss_scale": 4.0, "retries": 0, "seconds": 61.4, "max_memory_gb": 16.47, "ce": 0.14873241024906747, "kl": 0.10015133182787395, "context": 0.19880807242589071}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.228124976158142, "loss_scale": 4.0, "retries": 0, "seconds": 61.8, "max_memory_gb": 16.47, "ce": 0.14658278910792433, "kl": 0.20012966719514225, "context": 0.24664762275642715}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 3.0756523609161377, "loss_scale": 4.0, "retries": 0, "seconds": 62.1, "max_memory_gb": 16.47, "ce": 0.2736607240221929, "kl": 0.5433752180251759, "context": 0.5453483503661118}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 0.6024351119995117, "loss_scale": 4.0, "retries": 0, "seconds": 62.5, "max_memory_gb": 16.47, "ce": 0.09910765997483395, "kl": 0.07888251366603072, "context": 0.13854891833034344}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 0.23399659991264343, "loss_scale": 4.0, "retries": 0, "seconds": 62.9, "max_memory_gb": 16.47, "ce": 0.013955814509245101, "kl": 0.01891580155643169, "context": 0.023413715534843504}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 1.534469485282898, "loss_scale": 4.0, "retries": 0, "seconds": 63.3, "max_memory_gb": 16.47, "ce": 0.25932274910155684, "kl": 0.14050840091658756, "context": 0.3295769449323416}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 0.7162711024284363, "loss_scale": 4.0, "retries": 0, "seconds": 63.7, "max_memory_gb": 16.47, "ce": 0.06563666253350675, "kl": 0.010640355059877038, "context": 0.07095683924853802}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.23449170589447021, "loss_scale": 4.0, "retries": 0, "seconds": 64.1, "max_memory_gb": 16.47, "ce": 0.015161020601226483, "kl": 0.003211345392628573, "context": 0.016766693588579074}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.7217001914978027, "loss_scale": 4.0, "retries": 0, "seconds": 64.4, "max_memory_gb": 16.47, "ce": 0.08413251704769209, "kl": 0.952951405197382, "context": 0.5606082323938608}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 0.04280557483434677, "loss_scale": 4.0, "retries": 0, "seconds": 64.8, "max_memory_gb": 16.47, "ce": 0.002306578680872917, "kl": 0.00031897708504402544, "context": 0.0024660671770107}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 1.6938555240631104, "loss_scale": 4.0, "retries": 0, "seconds": 65.2, "max_memory_gb": 16.47, "ce": 0.04754379038058687, "kl": 0.17426306777633727, "context": 0.13467532116919756}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.5032947063446045, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 16.47, "ce": 0.018851318513043225, "kl": 0.025597029700293206, "context": 0.03164983360329643}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.33910253643989563, "loss_scale": 4.0, "retries": 0, "seconds": 65.9, "max_memory_gb": 16.47, "ce": 0.018326543096918613, "kl": 0.007650433326489292, "context": 0.0221517599420622}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 0.9617010354995728, "loss_scale": 4.0, "retries": 0, "seconds": 66.3, "max_memory_gb": 16.47, "ce": 0.04340904932178091, "kl": 0.042734970625588176, "context": 0.06477653277397621}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 1.3058862686157227, "loss_scale": 4.0, "retries": 0, "seconds": 66.7, "max_memory_gb": 16.47, "ce": 0.26696532836649567, "kl": 0.0374913098057732, "context": 0.28571098297834396}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 0.1587214469909668, "loss_scale": 4.0, "retries": 0, "seconds": 67.1, "max_memory_gb": 16.47, "ce": 0.007426342985127121, "kl": 0.0027485323807923123, "context": 0.008800609211903065}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 1.0197340250015259, "loss_scale": 4.0, "retries": 0, "seconds": 67.4, "max_memory_gb": 16.47, "ce": 0.02566056548675988, "kl": 0.6996946984472743, "context": 0.3755079165712232}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 0.6800634264945984, "loss_scale": 4.0, "retries": 0, "seconds": 67.8, "max_memory_gb": 16.47, "ce": 0.15252121147932485, "kl": 0.03588447594665922, "context": 0.17046344478148967}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 0.8461829423904419, "loss_scale": 4.0, "retries": 0, "seconds": 68.2, "max_memory_gb": 16.47, "ce": 0.41354010306531563, "kl": 0.009129666072112741, "context": 0.4181049325852655}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 1.2973538637161255, "loss_scale": 4.0, "retries": 0, "seconds": 68.5, "max_memory_gb": 16.47, "ce": 0.055293579905992374, "kl": 0.03419249832586502, "context": 0.07238983023853507}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 0.6040036678314209, "loss_scale": 4.0, "retries": 0, "seconds": 68.9, "max_memory_gb": 16.47, "ce": 0.01714630408969242, "kl": 0.15360285257338546, "context": 0.09394773273379542}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 0.7336249351501465, "loss_scale": 4.0, "retries": 0, "seconds": 69.3, "max_memory_gb": 16.47, "ce": 0.11308077222201973, "kl": 0.11812185827147914, "context": 0.17214170203078538}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.06218017637729645, "loss_scale": 4.0, "retries": 0, "seconds": 69.7, "max_memory_gb": 16.47, "ce": 0.0037850177905056626, "kl": 0.0006555576401297003, "context": 0.004112796625122428}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 0.07674992084503174, "loss_scale": 4.0, "retries": 0, "seconds": 70.1, "max_memory_gb": 16.47, "ce": 0.0031217253126669675, "kl": 0.002835927480191458, "context": 0.004539689020020887}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 0.8030687570571899, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 16.47, "ce": 0.017595787707250565, "kl": 0.1121953700640006, "context": 0.07369347184430808}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 1.466934323310852, "loss_scale": 4.0, "retries": 0, "seconds": 70.9, "max_memory_gb": 16.47, "ce": 0.17902536226029042, "kl": 0.2646285150767653, "context": 0.31133960490114987}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 1.2443082332611084, "loss_scale": 4.0, "retries": 0, "seconds": 71.2, "max_memory_gb": 16.47, "ce": 0.03670887931366451, "kl": 0.43585354535025544, "context": 0.2546356492675841}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 1.0394351482391357, "loss_scale": 4.0, "retries": 0, "seconds": 71.6, "max_memory_gb": 16.47, "ce": 0.24790627509355545, "kl": 0.019005087700861623, "context": 0.2574088138353545}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 1.3132929801940918, "loss_scale": 4.0, "retries": 0, "seconds": 72.0, "max_memory_gb": 16.47, "ce": 0.3354244108486455, "kl": 0.092861259370693, "context": 0.38185504809371196}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 0.1094052717089653, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 16.47, "ce": 0.0019801703965640627, "kl": 0.010180479439441115, "context": 0.007070409948937595}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 0.9418196082115173, "loss_scale": 4.0, "retries": 0, "seconds": 72.8, "max_memory_gb": 16.47, "ce": 0.1574037903919816, "kl": 0.08402810191182652, "context": 0.19941784534603357}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.490408182144165, "loss_scale": 4.0, "retries": 0, "seconds": 73.1, "max_memory_gb": 16.47, "ce": 0.485149459913373, "kl": 0.14138677113078302, "context": 0.5558428450021893}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 0.10789324343204498, "loss_scale": 4.0, "retries": 0, "seconds": 73.6, "max_memory_gb": 16.47, "ce": 0.005615483969449997, "kl": 0.0023205966881505447, "context": 0.00677578235627152}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 0.5004081130027771, "loss_scale": 4.0, "retries": 0, "seconds": 74.0, "max_memory_gb": 16.47, "ce": 0.0645717044826597, "kl": 0.02822284633293748, "context": 0.0786831303848885}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 0.18108510971069336, "loss_scale": 4.0, "retries": 0, "seconds": 74.4, "max_memory_gb": 16.47, "ce": 0.01270779746118933, "kl": 0.004684885861934163, "context": 0.015050240617711097}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 2.215398073196411, "loss_scale": 4.0, "retries": 0, "seconds": 74.7, "max_memory_gb": 16.47, "ce": 0.0014248650841182098, "kl": 0.33015964144988175, "context": 0.1665046868729405}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 0.7424083948135376, "loss_scale": 4.0, "retries": 0, "seconds": 75.1, "max_memory_gb": 16.47, "ce": 0.03646673623006791, "kl": 0.6661994516962295, "context": 0.36956646654289216}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 0.5255900621414185, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 16.47, "ce": 0.11584945549839176, "kl": 0.008037711209908593, "context": 0.11986831336980686}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 1.8391447067260742, "loss_scale": 4.0, "retries": 0, "seconds": 75.9, "max_memory_gb": 16.47, "ce": 0.12454358024115209, "kl": 0.03778899661392643, "context": 0.1434380780265201}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 0.7818291187286377, "loss_scale": 4.0, "retries": 0, "seconds": 76.2, "max_memory_gb": 16.47, "ce": 0.34851751499809325, "kl": 0.020121883702813648, "context": 0.35857845784630626}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 0.5142010450363159, "loss_scale": 4.0, "retries": 0, "seconds": 76.6, "max_memory_gb": 16.47, "ce": 0.0739532053557923, "kl": 0.09347408126632217, "context": 0.12069024598167744}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 0.060144100338220596, "loss_scale": 4.0, "retries": 0, "seconds": 77.0, "max_memory_gb": 16.47, "ce": 0.0020258083677617833, "kl": 0.003010268395428284, "context": 0.0035309425875311717}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 0.8705191612243652, "loss_scale": 4.0, "retries": 0, "seconds": 77.3, "max_memory_gb": 16.47, "ce": 0.06307820872461889, "kl": 0.07465846079867333, "context": 0.10040744283469394}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.6567627191543579, "loss_scale": 4.0, "retries": 0, "seconds": 77.7, "max_memory_gb": 16.47, "ce": 0.03844300241325982, "kl": 0.00752836384344846, "context": 0.04220718389842659}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 1.5584237575531006, "loss_scale": 4.0, "retries": 0, "seconds": 78.1, "max_memory_gb": 16.47, "ce": 0.009527946298476309, "kl": 0.3570271915086778, "context": 0.18804154571262188}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 0.11572178453207016, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 16.47, "ce": 0.005019896474550478, "kl": 0.0020912714326186688, "context": 0.0060655323177343234}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 0.25836363434791565, "loss_scale": 4.0, "retries": 0, "seconds": 78.8, "max_memory_gb": 16.47, "ce": 0.07002195087261498, "kl": 0.0471033600915689, "context": 0.0935736270621419}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.08919896930456161, "loss_scale": 4.0, "retries": 0, "seconds": 79.2, "max_memory_gb": 16.47, "ce": 0.0038312537071760744, "kl": 0.001783654992323136, "context": 0.004723081292468123}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 0.5937061309814453, "loss_scale": 4.0, "retries": 0, "seconds": 79.6, "max_memory_gb": 16.47, "ce": 0.16860436991555616, "kl": 0.04636991332517937, "context": 0.19178932750946842}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 1.0133178234100342, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 16.47, "ce": 0.46624342462746426, "kl": 0.11074703005579067, "context": 0.521616913494654}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.6531597375869751, "loss_scale": 4.0, "retries": 0, "seconds": 80.4, "max_memory_gb": 16.47, "ce": 0.11403090530075133, "kl": 0.11175189632922411, "context": 0.16990684706252068}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 1.4864791631698608, "loss_scale": 4.0, "retries": 0, "seconds": 80.7, "max_memory_gb": 16.47, "ce": 0.04915522175724618, "kl": 0.8859698821033817, "context": 0.49214015633333474}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 0.849037230014801, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 16.47, "ce": 0.2380294131289702, "kl": 0.030955231388361426, "context": 0.2535070267331321}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 3.607100009918213, "loss_scale": 4.0, "retries": 0, "seconds": 81.4, "max_memory_gb": 16.47, "ce": 0.4058263941842597, "kl": 0.12988696433603764, "context": 0.47076985659077764}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 1.3108117580413818, "loss_scale": 4.0, "retries": 0, "seconds": 81.8, "max_memory_gb": 16.47, "ce": 0.5340528840606567, "kl": 0.02390852714893299, "context": 0.5460071610868908}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 1.063254952430725, "loss_scale": 4.0, "retries": 0, "seconds": 82.2, "max_memory_gb": 16.47, "ce": 0.040259367786347866, "kl": 0.13043024099897593, "context": 0.10547448880970478}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 0.08914871513843536, "loss_scale": 4.0, "retries": 0, "seconds": 82.6, "max_memory_gb": 16.47, "ce": 0.0038588158786296844, "kl": 0.0045954680767863465, "context": 0.0061565499054268}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 0.7415753602981567, "loss_scale": 4.0, "retries": 0, "seconds": 83.0, "max_memory_gb": 16.47, "ce": 0.19583761971443892, "kl": 0.1829968011006713, "context": 0.2873360179364681}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.22523267567157745, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 16.47, "ce": 0.018487931505660526, "kl": 0.06847485355683602, "context": 0.05272535956464708}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 0.666539192199707, "loss_scale": 4.0, "retries": 0, "seconds": 83.7, "max_memory_gb": 16.47, "ce": 0.04661873294389807, "kl": 0.007183560617704643, "context": 0.05021051276708022}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 1.8918818235397339, "loss_scale": 4.0, "retries": 0, "seconds": 84.1, "max_memory_gb": 16.47, "ce": 0.00478581314382609, "kl": 0.3446688579933834, "context": 0.17712024180218577}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 1.2580595016479492, "loss_scale": 4.0, "retries": 0, "seconds": 84.5, "max_memory_gb": 16.47, "ce": 0.02413000285741873, "kl": 0.056485004231490166, "context": 0.052372505713719875}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 0.6261196136474609, "loss_scale": 4.0, "retries": 0, "seconds": 84.9, "max_memory_gb": 16.47, "ce": 0.0556555257353466, "kl": 0.0044502636446850374, "context": 0.05788065874367021}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 0.6913971304893494, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 16.47, "ce": 0.0538810086727608, "kl": 0.08219753256707918, "context": 0.0949797754874453}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 1.2869007587432861, "loss_scale": 4.0, "retries": 0, "seconds": 85.6, "max_memory_gb": 16.47, "ce": 0.007214898447273299, "kl": 0.11727036199954455, "context": 0.06585007975809276}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 0.7540601491928101, "loss_scale": 4.0, "retries": 0, "seconds": 86.0, "max_memory_gb": 16.47, "ce": 0.10156975369318388, "kl": 0.019935157732106745, "context": 0.11153733351966366}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 0.4910772740840912, "loss_scale": 4.0, "retries": 0, "seconds": 86.3, "max_memory_gb": 16.47, "ce": 0.06049316137796268, "kl": 0.006443053913244512, "context": 0.06371468855650164}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 1.4194769859313965, "loss_scale": 4.0, "retries": 0, "seconds": 86.7, "max_memory_gb": 16.47, "ce": 0.014695379315526225, "kl": 0.17346279649063945, "context": 0.1014267758000642}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 0.4714776873588562, "loss_scale": 4.0, "retries": 0, "seconds": 87.1, "max_memory_gb": 16.47, "ce": 0.0855556153546786, "kl": 0.016333155815118516, "context": 0.09372219443321228}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 1.5537902116775513, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 16.47, "ce": 0.3353754512500018, "kl": 0.22757658828049898, "context": 0.4491637524915859}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 0.8339364528656006, "loss_scale": 4.0, "retries": 0, "seconds": 87.8, "max_memory_gb": 16.47, "ce": 0.055992231995332986, "kl": 0.11839321695151739, "context": 0.1151888442109339}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 0.5641066431999207, "loss_scale": 4.0, "retries": 0, "seconds": 88.2, "max_memory_gb": 16.47, "ce": 0.02124386018840596, "kl": 0.00885069889773149, "context": 0.025669209222542122}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 0.7081896066665649, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 16.47, "ce": 0.017782746348530054, "kl": 0.1475083266850561, "context": 0.09153691027313471}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 0.3769172728061676, "loss_scale": 4.0, "retries": 0, "seconds": 89.0, "max_memory_gb": 16.47, "ce": 0.04113022590172477, "kl": 0.03511330056062434, "context": 0.05868687623296864}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 1.0042308568954468, "loss_scale": 4.0, "retries": 0, "seconds": 89.4, "max_memory_gb": 16.47, "ce": 0.22526582670980133, "kl": 0.026682250914745964, "context": 0.23860695547773503}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 0.37550005316734314, "loss_scale": 4.0, "retries": 0, "seconds": 89.8, "max_memory_gb": 16.47, "ce": 0.014762225182494149, "kl": 0.02047308747205534, "context": 0.0249987686984241}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 0.44837456941604614, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 16.47, "ce": 0.10213250410743058, "kl": 0.0246649350028747, "context": 0.11446497173164971}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.3780568540096283, "loss_scale": 4.0, "retries": 0, "seconds": 90.5, "max_memory_gb": 16.47, "ce": 0.03611299971817061, "kl": 0.013946773751740693, "context": 0.043086387173389085}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 0.5128703117370605, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 16.47, "ce": 0.12954806687776, "kl": 0.02059185587859247, "context": 0.13984399114269763}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 1.7161273956298828, "loss_scale": 4.0, "retries": 0, "seconds": 91.2, "max_memory_gb": 16.47, "ce": 0.13532035559182987, "kl": 0.027717215809389018, "context": 0.1491789638530463}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 0.6830705404281616, "loss_scale": 4.0, "retries": 0, "seconds": 91.6, "max_memory_gb": 16.47, "ce": 0.19186534800974187, "kl": 0.020527619912172668, "context": 0.20212915621232241}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 0.24240662157535553, "loss_scale": 4.0, "retries": 0, "seconds": 92.0, "max_memory_gb": 16.47, "ce": 0.008872782433172688, "kl": 0.01967269080614642, "context": 0.01870912784943357}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.08302410691976547, "loss_scale": 4.0, "retries": 0, "seconds": 92.3, "max_memory_gb": 16.47, "ce": 0.003994766273535788, "kl": 0.003012652858160436, "context": 0.005501092702616006}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 1.138001561164856, "loss_scale": 4.0, "retries": 0, "seconds": 92.7, "max_memory_gb": 16.47, "ce": 0.07862470201507676, "kl": 0.1288427395575127, "context": 0.14304606667428743}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 0.9014365673065186, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 16.47, "ce": 0.3761249631352257, "kl": 0.034601066072355025, "context": 0.3934254952473566}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.0575554184615612, "loss_scale": 4.0, "retries": 0, "seconds": 93.5, "max_memory_gb": 16.47, "ce": 0.0029902705718996003, "kl": 0.0010475618037162349, "context": 0.0035140514955855906}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 1.3495700359344482, "loss_scale": 4.0, "retries": 0, "seconds": 93.9, "max_memory_gb": 16.47, "ce": 0.045711888902587816, "kl": 0.43700347534468165, "context": 0.26421362462861}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.7220816612243652, "loss_scale": 4.0, "retries": 0, "seconds": 94.3, "max_memory_gb": 16.47, "ce": 0.03527180524542928, "kl": 0.032312741968780756, "context": 0.05142817588057369}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 1.5349928140640259, "loss_scale": 4.0, "retries": 0, "seconds": 94.6, "max_memory_gb": 16.47, "ce": 0.04839531797915697, "kl": 0.05487796082161367, "context": 0.07583429664373398}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 0.21956956386566162, "loss_scale": 4.0, "retries": 0, "seconds": 95.0, "max_memory_gb": 16.47, "ce": 0.01584162237122655, "kl": 0.007120237700291909, "context": 0.019401741272304207}
|
runs/single_internvl/replacement/L01_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/replacement/L01_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.1.mlp.gate_proj",
|
| 35 |
+
"model.language_model.layers.1.mlp.down_proj",
|
| 36 |
+
"model.language_model.layers.1.self_attn.q_proj",
|
| 37 |
+
"model.language_model.layers.1.self_attn.v_proj",
|
| 38 |
+
"model.language_model.layers.1.self_attn.o_proj",
|
| 39 |
+
"model.language_model.layers.1.self_attn.k_proj",
|
| 40 |
+
"model.language_model.layers.1.mlp.up_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/replacement/L01_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "replacement",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/replacement.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 1,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/replacement/L01_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.9593239426612854, "loss_scale": 4.0, "retries": 0, "seconds": 1.3, "max_memory_gb": 24.84, "pair": 1.0986123085021973, "sign": 0.7328193015418947, "keep": 1.137254401575774, "leak": 0.0, "replacement": 2.0336491465568542}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.8836205005645752, "loss_scale": 4.0, "retries": 0, "seconds": 2.4, "max_memory_gb": 24.84, "pair": 1.100912630558014, "sign": 0.18113453849218786, "keep": 1.3659147868456785, "leak": 4.569975430968043e-05, "replacement": 1.8744487464427948}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 2.0255348682403564, "loss_scale": 4.0, "retries": 0, "seconds": 3.5, "max_memory_gb": 24.84, "pair": 1.096558153629303, "sign": 0.1154714091680944, "keep": 2.063725635409355, "leak": 0.0001009879802040814, "replacement": 2.186181962490082}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.3091331720352173, "loss_scale": 4.0, "retries": 0, "seconds": 4.6, "max_memory_gb": 24.84, "pair": 1.0949964225292206, "sign": 0.03801619075238705, "keep": 0.8714177075889893, "leak": 5.002560396860645e-05, "replacement": 1.549725890159607}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.5621684193611145, "loss_scale": 4.0, "retries": 0, "seconds": 5.7, "max_memory_gb": 24.84, "pair": 1.1011099517345428, "sign": 0.11726495902985334, "keep": 0.02337648676621029, "leak": 6.079615431886509e-05, "replacement": 1.1714458763599396}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.7710880041122437, "loss_scale": 4.0, "retries": 0, "seconds": 6.7, "max_memory_gb": 24.84, "pair": 1.1010949909687042, "sign": 0.20000055711716413, "keep": 0.6361354511464015, "leak": 3.651940033932988e-05, "replacement": 1.5191721022129059}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 1.5538718700408936, "loss_scale": 4.0, "retries": 0, "seconds": 7.8, "max_memory_gb": 24.84, "pair": 1.0963671505451202, "sign": 0.08985515031963587, "keep": 1.5254834839870455, "leak": 7.902192828623811e-05, "replacement": 1.9040562808513641}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 1.6165627241134644, "loss_scale": 4.0, "retries": 0, "seconds": 8.9, "max_memory_gb": 24.88, "pair": 1.090606689453125, "sign": 0.20960095198825002, "keep": 1.4586929976940155, "leak": 0.00017665238647168735, "replacement": 1.9247978031635284}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 4.158328056335449, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 24.88, "pair": 1.0889300405979156, "sign": 0.5744876121170819, "keep": 1.380729965865612, "leak": 2.7532164441623763e-05, "replacement": 2.0665457248687744}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.6829278469085693, "loss_scale": 4.0, "retries": 0, "seconds": 11.1, "max_memory_gb": 24.88, "pair": 1.094979465007782, "sign": 0.0802498459815979, "keep": 0.09905106155201793, "leak": 9.233138098352356e-05, "replacement": 1.1846530139446259}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 1.3712434768676758, "loss_scale": 4.0, "retries": 0, "seconds": 12.2, "max_memory_gb": 24.88, "pair": 1.0926148891448975, "sign": 0.20463008247315884, "keep": 1.3226584108633688, "leak": 8.867381075106096e-05, "replacement": 1.8562813103199005}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.155125617980957, "loss_scale": 4.0, "retries": 0, "seconds": 13.3, "max_memory_gb": 24.88, "pair": 1.086828202009201, "sign": 0.07588598993606865, "keep": 0.9347169754619244, "leak": 0.00013547539482772208, "replacement": 1.592163622379303}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 0.9987849593162537, "loss_scale": 4.0, "retries": 0, "seconds": 14.3, "max_memory_gb": 24.88, "pair": 1.0809769034385681, "sign": 0.05775862652808428, "keep": 0.8571424642868806, "leak": 0.00036681828265727745, "replacement": 1.53851917386055}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 1.6895848512649536, "loss_scale": 4.0, "retries": 0, "seconds": 15.4, "max_memory_gb": 24.88, "pair": 1.0866911113262177, "sign": 0.26730811037123203, "keep": 1.6208590865135193, "leak": 0.0003212282181266346, "replacement": 2.030855029821396}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 1.566087245941162, "loss_scale": 4.0, "retries": 0, "seconds": 16.5, "max_memory_gb": 24.88, "pair": 1.0758999586105347, "sign": 0.19864547508768737, "keep": 1.008882045336577, "leak": 0.0009738420776557177, "replacement": 1.679907202720642}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 1.0744260549545288, "loss_scale": 4.0, "retries": 0, "seconds": 17.5, "max_memory_gb": 24.88, "pair": 1.0773667693138123, "sign": 0.30122254928573966, "keep": 0.37775913847144693, "leak": 0.0003974992396251764, "replacement": 1.4169570207595825}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 1.2805174589157104, "loss_scale": 4.0, "retries": 0, "seconds": 18.6, "max_memory_gb": 24.88, "pair": 1.0755679309368134, "sign": 0.7401372450403869, "keep": 0.8349960019113496, "leak": 0.0009162000933429226, "replacement": 1.8633635640144348}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.6800304651260376, "loss_scale": 4.0, "retries": 0, "seconds": 19.7, "max_memory_gb": 24.91, "pair": 1.0667639374732971, "sign": 0.1887393263168633, "keep": 1.0721882901561912, "leak": 0.0012781343248207122, "replacement": 1.6975472569465637}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 2.6647350788116455, "loss_scale": 4.0, "retries": 0, "seconds": 20.8, "max_memory_gb": 24.91, "pair": 1.034154251217842, "sign": 0.036760090151801705, "keep": 1.3786970674991608, "leak": 0.008728663691726979, "replacement": 1.74406498670578}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 0.9031949639320374, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 24.91, "pair": 1.0687974989414215, "sign": 0.03151711728423834, "keep": 0.0038333598313329276, "leak": 0.0008942735148593783, "replacement": 1.086696296930313}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 1.7205910682678223, "loss_scale": 4.0, "retries": 0, "seconds": 23.0, "max_memory_gb": 24.91, "pair": 1.038803517818451, "sign": 0.04534841806162149, "keep": 0.6147950096492423, "leak": 0.006071922078490388, "replacement": 1.3703932166099548}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 1.7196253538131714, "loss_scale": 4.0, "retries": 0, "seconds": 24.1, "max_memory_gb": 24.91, "pair": 1.0368816256523132, "sign": 0.0672085927799344, "keep": 0.7375379111617804, "leak": 0.0032519787055207416, "replacement": 1.440067857503891}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.0640406608581543, "loss_scale": 4.0, "retries": 0, "seconds": 25.3, "max_memory_gb": 24.91, "pair": 1.027148813009262, "sign": 0.03629625914618373, "keep": 1.2526440611909493, "leak": 0.0027307907585054636, "replacement": 1.6723016798496246}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 2.1688501834869385, "loss_scale": 4.0, "retries": 0, "seconds": 26.3, "max_memory_gb": 24.91, "pair": 0.9735378921031952, "sign": 0.016917557921260595, "keep": 0.6271993592381477, "leak": 0.008676064317114651, "replacement": 1.2977654039859772}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 2.0932400226593018, "loss_scale": 4.0, "retries": 0, "seconds": 27.4, "max_memory_gb": 24.91, "pair": 0.9956439137458801, "sign": 0.11262324918061495, "keep": 1.3639576763453078, "leak": 0.009263821542845108, "replacement": 1.7362503111362457}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 2.2887678146362305, "loss_scale": 4.0, "retries": 0, "seconds": 28.5, "max_memory_gb": 24.91, "pair": 0.9629136174917221, "sign": 0.06361226190347224, "keep": 0.6764476895332336, "leak": 0.015040741243865341, "replacement": 1.33670374751091}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 2.127100944519043, "loss_scale": 4.0, "retries": 0, "seconds": 29.5, "max_memory_gb": 24.91, "pair": 0.9679032266139984, "sign": 0.14422611729241908, "keep": 0.8377881149644963, "leak": 0.015461106318980455, "replacement": 1.4627755731344223}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 2.200202703475952, "loss_scale": 4.0, "retries": 0, "seconds": 30.6, "max_memory_gb": 24.91, "pair": 1.0048754215240479, "sign": 0.4199048785958439, "keep": 1.4303466156125069, "leak": 0.019662330771097913, "replacement": 1.9349167346954346}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 2.0573318004608154, "loss_scale": 4.0, "retries": 0, "seconds": 31.6, "max_memory_gb": 24.91, "pair": 1.0216076076030731, "sign": 0.7323151044547558, "keep": 0.5860957559198141, "leak": 0.005672101583513722, "replacement": 1.6822310388088226}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 2.2713782787323, "loss_scale": 4.0, "retries": 0, "seconds": 32.6, "max_memory_gb": 24.91, "pair": 0.9004113376140594, "sign": 0.011818741215392947, "keep": 0.6846532040217426, "leak": 0.020499725826084614, "replacement": 1.2537721991539001}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 3.529432535171509, "loss_scale": 4.0, "retries": 0, "seconds": 33.6, "max_memory_gb": 24.91, "pair": 0.9632610827684402, "sign": 0.21891301800496876, "keep": 1.3696066588163376, "leak": 0.026491371827432886, "replacement": 1.7641438245773315}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 3.1211459636688232, "loss_scale": 4.0, "retries": 0, "seconds": 34.7, "max_memory_gb": 24.91, "pair": 0.8854369521141052, "sign": 0.1204133415594697, "keep": 1.0019936636090279, "leak": 0.07520221918821335, "replacement": 1.4654409885406494}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 1.7438175678253174, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 24.91, "pair": 0.9792991578578949, "sign": 0.2037269426509738, "keep": 0.5346342120319605, "leak": 0.019816807907773182, "replacement": 1.353433907032013}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 1.8584084510803223, "loss_scale": 4.0, "retries": 0, "seconds": 36.9, "max_memory_gb": 24.91, "pair": 0.9075607806444168, "sign": 0.0335038504563272, "keep": 0.3647868746847962, "leak": 0.05572750326246023, "replacement": 1.1206380128860474}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 2.1501827239990234, "loss_scale": 4.0, "retries": 0, "seconds": 38.0, "max_memory_gb": 24.91, "pair": 0.9112263023853302, "sign": 0.06441305577754974, "keep": 0.5002845347626135, "leak": 0.031040637753903866, "replacement": 1.2013352662324905}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 1.9061715602874756, "loss_scale": 4.0, "retries": 0, "seconds": 39.1, "max_memory_gb": 24.91, "pair": 0.8837364464998245, "sign": 0.04627762152813375, "keep": 0.47674247436225414, "leak": 0.07598071033135056, "replacement": 1.1642416715621948}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 1.7615516185760498, "loss_scale": 4.0, "retries": 0, "seconds": 40.2, "max_memory_gb": 24.91, "pair": 0.9151197075843811, "sign": 0.21506958827376366, "keep": 0.9008815992856398, "leak": 0.05712219490123971, "replacement": 1.4873758554458618}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 2.0702669620513916, "loss_scale": 4.0, "retries": 0, "seconds": 41.3, "max_memory_gb": 24.91, "pair": 0.8231582939624786, "sign": 0.09208180801942945, "keep": 0.3433177024126053, "leak": 0.11076125130057335, "replacement": 1.0685483813285828}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 1.4943896532058716, "loss_scale": 4.0, "retries": 0, "seconds": 42.3, "max_memory_gb": 24.91, "pair": 0.9385769069194794, "sign": 1.3502098609460518, "keep": 0.3742325920102303, "leak": 0.038869730415171944, "replacement": 1.8105155676603317}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 1.884460687637329, "loss_scale": 4.0, "retries": 0, "seconds": 43.4, "max_memory_gb": 24.91, "pair": 0.8147080540657043, "sign": 0.021244398318231106, "keep": 0.28902948620088864, "leak": 0.10730892606079578, "replacement": 0.9966722130775452}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 1.861406922340393, "loss_scale": 4.0, "retries": 0, "seconds": 44.5, "max_memory_gb": 24.91, "pair": 0.9391740560531616, "sign": 0.862089367583394, "keep": 0.9869738142006099, "leak": 0.02878342568874359, "replacement": 1.8709015101194382}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 1.8960652351379395, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 24.91, "pair": 0.7361322194337845, "sign": 0.056262134574353695, "keep": 0.13470058426901232, "leak": 0.11910193227231503, "replacement": 0.8613890558481216}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 2.140575885772705, "loss_scale": 4.0, "retries": 0, "seconds": 46.6, "max_memory_gb": 24.91, "pair": 0.728646382689476, "sign": 0.15320531511679292, "keep": 0.38966699813317973, "leak": 0.410024621989578, "replacement": 1.102588728070259}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 2.8155951499938965, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 24.91, "pair": 0.7938971817493439, "sign": 0.1722112197894603, "keep": 1.0035638708795886, "leak": 0.12069942988455296, "replacement": 1.4119596034288406}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 2.78945255279541, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 24.91, "pair": 0.9258239567279816, "sign": 0.6827853759750724, "keep": 1.9325696126325056, "leak": 0.03271326544222575, "replacement": 2.241679832339287}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 2.1906940937042236, "loss_scale": 4.0, "retries": 0, "seconds": 49.9, "max_memory_gb": 24.91, "pair": 0.772977277636528, "sign": 0.2694443682557903, "keep": 0.5847670249640942, "leak": 0.10058769630268216, "replacement": 1.2252298891544342}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 1.8550363779067993, "loss_scale": 4.0, "retries": 0, "seconds": 51.0, "max_memory_gb": 24.91, "pair": 0.7494458630681038, "sign": 0.10781487240456045, "keep": 0.1399290138942888, "leak": 0.1479154495173134, "replacement": 0.9102966487407684}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 2.113006114959717, "loss_scale": 4.0, "retries": 0, "seconds": 52.1, "max_memory_gb": 24.91, "pair": 0.6056891828775406, "sign": 0.02908642269903794, "keep": 0.3788295137928799, "leak": 0.3252095356583595, "replacement": 0.890949547290802}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 2.433511257171631, "loss_scale": 4.0, "retries": 0, "seconds": 53.2, "max_memory_gb": 24.91, "pair": 0.6061426177620888, "sign": 0.05482121976092458, "keep": 0.2475166850927053, "leak": 0.1772209107875824, "replacement": 0.8016167879104614}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 1.9321521520614624, "loss_scale": 4.0, "retries": 0, "seconds": 54.2, "max_memory_gb": 24.91, "pair": 0.7025566250085831, "sign": 0.11669068038463593, "keep": 0.5151523369349889, "leak": 0.31896121241152287, "replacement": 1.0982184410095215}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 2.8775508403778076, "loss_scale": 4.0, "retries": 0, "seconds": 55.2, "max_memory_gb": 24.91, "pair": 0.6460344195365906, "sign": 0.034270408912561834, "keep": 1.1264112496719463, "leak": 0.12568371527595446, "replacement": 1.257796198129654}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 1.928645133972168, "loss_scale": 4.0, "retries": 0, "seconds": 56.3, "max_memory_gb": 24.91, "pair": 0.6637492924928665, "sign": 0.2721002446487546, "keep": 0.4724055406986736, "leak": 0.21040969155728817, "replacement": 1.0886046141386032}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 2.3556747436523438, "loss_scale": 4.0, "retries": 0, "seconds": 57.3, "max_memory_gb": 24.91, "pair": 0.5731634199619293, "sign": 0.09998900070786476, "keep": 0.4027402396313846, "leak": 0.4756558556109667, "replacement": 0.9434420168399811}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 2.57547664642334, "loss_scale": 4.0, "retries": 0, "seconds": 58.4, "max_memory_gb": 24.91, "pair": 0.6269023194909096, "sign": 0.16912469873204827, "keep": 0.35750041739083827, "leak": 0.41954235220327973, "replacement": 0.9951004981994629}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 1.2889399528503418, "loss_scale": 4.0, "retries": 0, "seconds": 59.4, "max_memory_gb": 24.91, "pair": 0.5369869694113731, "sign": 0.03741335088852793, "keep": 0.01811919348256197, "leak": 0.2997436113655567, "replacement": 0.6396891474723816}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 1.4326708316802979, "loss_scale": 4.0, "retries": 0, "seconds": 60.4, "max_memory_gb": 24.91, "pair": 0.4692816957831383, "sign": 0.01882627629674971, "keep": 0.11444142973050475, "leak": 0.3823040798306465, "replacement": 0.6314915716648102}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 1.4226309061050415, "loss_scale": 4.0, "retries": 0, "seconds": 61.5, "max_memory_gb": 24.91, "pair": 0.6088783219456673, "sign": 0.3788689957000315, "keep": 0.04120112683449406, "leak": 0.3841096750693396, "replacement": 0.914940819144249}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 1.3140082359313965, "loss_scale": 4.0, "retries": 0, "seconds": 62.6, "max_memory_gb": 24.91, "pair": 0.4019218981266022, "sign": 0.01698891445994377, "keep": 0.10793496236146893, "leak": 0.49546221643686295, "replacement": 0.5882494077086449}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 1.532514214515686, "loss_scale": 4.0, "retries": 0, "seconds": 63.6, "max_memory_gb": 24.91, "pair": 0.527865506708622, "sign": 0.044620986678637564, "keep": 0.030759991248487495, "leak": 0.28746331646107137, "replacement": 0.6374218240380287}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 1.5194724798202515, "loss_scale": 4.0, "retries": 0, "seconds": 64.7, "max_memory_gb": 24.91, "pair": 0.4796089790761471, "sign": 0.0640085821505636, "keep": 0.10910004400648177, "leak": 0.7658028483911039, "replacement": 0.757614016532898}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 1.5239684581756592, "loss_scale": 4.0, "retries": 0, "seconds": 65.8, "max_memory_gb": 24.91, "pair": 0.504695039242506, "sign": 0.24969042581506073, "keep": 0.1104187211021781, "leak": 0.9446919541805983, "replacement": 0.9209225922822952}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 1.0000324249267578, "loss_scale": 4.0, "retries": 0, "seconds": 66.9, "max_memory_gb": 24.91, "pair": 0.49943047016859055, "sign": 0.033917239401489496, "keep": 0.09513175894971937, "leak": 0.6745178150013089, "replacement": 0.7325844317674637}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 1.1779537200927734, "loss_scale": 4.0, "retries": 0, "seconds": 68.0, "max_memory_gb": 24.91, "pair": 0.4339828584343195, "sign": 0.14946638653054833, "keep": 0.10187394195236266, "leak": 0.1898023618850857, "replacement": 0.6071036159992218}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.860977053642273, "loss_scale": 4.0, "retries": 0, "seconds": 69.1, "max_memory_gb": 24.91, "pair": 0.3329704962670803, "sign": 0.018753668060526252, "keep": 0.04903025319799781, "leak": 0.5910621248185635, "replacement": 0.5146280005574226}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 3.2860939502716064, "loss_scale": 4.0, "retries": 0, "seconds": 70.2, "max_memory_gb": 24.91, "pair": 0.41611362993717194, "sign": 0.358366958796978, "keep": 0.3174682538956404, "leak": 2.0285084173083305, "replacement": 1.2611583769321442}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 1.528635859489441, "loss_scale": 4.0, "retries": 0, "seconds": 71.3, "max_memory_gb": 24.91, "pair": 0.4837736636400223, "sign": 0.13629374478477985, "keep": 0.35867646333639414, "leak": 0.24402592983096838, "replacement": 0.7922652587294579}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 0.919424831867218, "loss_scale": 4.0, "retries": 0, "seconds": 72.4, "max_memory_gb": 24.91, "pair": 0.24085511453449726, "sign": 0.021444848272949457, "keep": 0.1414259089215193, "leak": 1.071868747472763, "replacement": 0.5902576744556427}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 1.3649795055389404, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 24.91, "pair": 0.5065605565905571, "sign": 0.054017911199480295, "keep": 0.030905174504823663, "leak": 0.4684041514992714, "replacement": 0.6661231368780136}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 2.295872211456299, "loss_scale": 4.0, "retries": 0, "seconds": 74.6, "max_memory_gb": 24.91, "pair": 0.6220027878880501, "sign": 0.8636098445858806, "keep": 0.649433043319732, "leak": 0.8501729383133352, "replacement": 1.5910674929618835}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 2.2771055698394775, "loss_scale": 4.0, "retries": 0, "seconds": 75.7, "max_memory_gb": 24.91, "pair": 0.42841141670942307, "sign": 0.03860668194829486, "keep": 0.04521861298417207, "leak": 1.866131566464901, "replacement": 0.9368569552898407}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 1.298382043838501, "loss_scale": 4.0, "retries": 0, "seconds": 76.8, "max_memory_gb": 24.91, "pair": 0.199148740619421, "sign": 0.00781807373277843, "keep": 0.20065851504841703, "leak": 0.877310000360012, "replacement": 0.5227145403623581}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 1.6637517213821411, "loss_scale": 4.0, "retries": 0, "seconds": 77.9, "max_memory_gb": 24.91, "pair": 0.28785306960344315, "sign": 0.19592745508998632, "keep": 0.1429770067334175, "leak": 1.3263638615608215, "replacement": 0.7888962626457214}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.6548287868499756, "loss_scale": 4.0, "retries": 0, "seconds": 78.9, "max_memory_gb": 24.91, "pair": 0.49113987386226654, "sign": 0.12668264005333185, "keep": 0.5229754740921635, "leak": 0.4790707267820835, "replacement": 0.9357366114854813}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 1.998927116394043, "loss_scale": 4.0, "retries": 0, "seconds": 80.0, "max_memory_gb": 24.91, "pair": 0.36229724809527397, "sign": 0.2446999920066446, "keep": 0.23086390643766208, "leak": 1.5239223489770666, "replacement": 0.981059804558754}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.6441998481750488, "loss_scale": 4.0, "retries": 0, "seconds": 81.1, "max_memory_gb": 24.91, "pair": 0.7146413065493107, "sign": 0.5235644610365853, "keep": 0.8656133045442402, "leak": 0.2414648192934692, "replacement": 1.4695963859558105}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.0517312288284302, "loss_scale": 4.0, "retries": 0, "seconds": 82.2, "max_memory_gb": 24.91, "pair": 0.3612573556602001, "sign": 0.10788466990925372, "keep": 0.03121067295433022, "leak": 0.76962910592556, "replacement": 0.6232122778892517}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 2.7209420204162598, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 24.91, "pair": 0.18452558107674122, "sign": 0.008171118330210447, "keep": 0.010215282119133917, "leak": 2.032524287700653, "replacement": 0.7018498554825783}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 2.094033718109131, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 24.91, "pair": 0.4612255357205868, "sign": 0.6581724279094487, "keep": 0.002558058753493242, "leak": 0.4953908510506153, "replacement": 0.9154385477304459}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 1.4861345291137695, "loss_scale": 4.0, "retries": 0, "seconds": 85.4, "max_memory_gb": 24.91, "pair": 0.46879126876592636, "sign": 0.031184740364551544, "keep": 0.1327907797131047, "leak": 0.5021803677082062, "replacement": 0.6763241291046143}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 1.0156691074371338, "loss_scale": 4.0, "retries": 0, "seconds": 86.4, "max_memory_gb": 24.91, "pair": 0.5084546804428101, "sign": 0.10036893608048558, "keep": 0.06729683261710306, "leak": 0.6280559562146664, "replacement": 0.749301552772522}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 1.0555386543273926, "loss_scale": 4.0, "retries": 0, "seconds": 87.5, "max_memory_gb": 24.91, "pair": 0.2903808318078518, "sign": 0.06096653948770836, "keep": 0.16856108838692307, "leak": 1.2339978758245707, "replacement": 0.7136441022157669}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 3.1040585041046143, "loss_scale": 4.0, "retries": 0, "seconds": 88.6, "max_memory_gb": 24.91, "pair": 0.2593694142997265, "sign": 1.0351668803486973, "keep": 0.11068129900377244, "leak": 1.988919958472252, "replacement": 1.3295235186815262}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 1.183480978012085, "loss_scale": 4.0, "retries": 0, "seconds": 89.7, "max_memory_gb": 24.91, "pair": 0.29837838746607304, "sign": 0.03216676786541939, "keep": 0.3582564471871592, "leak": 1.085783015936613, "replacement": 0.7650357559323311}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 1.478989601135254, "loss_scale": 4.0, "retries": 0, "seconds": 90.8, "max_memory_gb": 24.91, "pair": 0.25139553658664227, "sign": 0.029099896550178528, "keep": 0.05490097764345592, "leak": 1.1824527978897095, "replacement": 0.589009165763855}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 3.204007863998413, "loss_scale": 4.0, "retries": 0, "seconds": 91.9, "max_memory_gb": 24.91, "pair": 0.29698001593351364, "sign": 0.051450155093334615, "keep": 0.11955138284892541, "leak": 0.7830079197883606, "replacement": 0.5782327800989151}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 2.198681592941284, "loss_scale": 4.0, "retries": 0, "seconds": 93.0, "max_memory_gb": 24.91, "pair": 0.31244171783328056, "sign": 0.2843579100444913, "keep": 0.3249598501715809, "leak": 0.3660430945456028, "replacement": 0.7086113691329956}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 1.8662866353988647, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 24.91, "pair": 0.4569992944598198, "sign": 0.09859325061552227, "keep": 0.019736574176931754, "leak": 0.5683194734156132, "replacement": 0.6582440808415413}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 2.1579151153564453, "loss_scale": 4.0, "retries": 0, "seconds": 95.1, "max_memory_gb": 24.91, "pair": 0.42954133450984955, "sign": 0.2302305530756712, "keep": 0.03827836864979872, "leak": 0.653175838291645, "replacement": 0.7270897626876831}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 1.3979181051254272, "loss_scale": 4.0, "retries": 0, "seconds": 96.2, "max_memory_gb": 24.91, "pair": 0.5724548101425171, "sign": 0.18656899756751955, "keep": 0.016759189938568397, "leak": 0.18955759145319462, "replacement": 0.7215082794427872}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 2.065509557723999, "loss_scale": 4.0, "retries": 0, "seconds": 97.2, "max_memory_gb": 24.91, "pair": 0.4241482503712177, "sign": 0.1060818734113127, "keep": 0.06110549397999421, "leak": 1.17609441280365, "replacement": 0.8017655164003372}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 4.381953239440918, "loss_scale": 4.0, "retries": 0, "seconds": 98.4, "max_memory_gb": 24.91, "pair": 0.576403483748436, "sign": 0.24230129946954548, "keep": 0.2939742944945465, "leak": 0.5023393221199512, "replacement": 0.9701261520385742}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 4.595095157623291, "loss_scale": 4.0, "retries": 0, "seconds": 99.5, "max_memory_gb": 24.91, "pair": 0.4428306110203266, "sign": 0.7479272527853027, "keep": 1.2177159899756589, "leak": 0.6894774107495323, "replacement": 1.5980216711759567}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.6285349130630493, "loss_scale": 4.0, "retries": 0, "seconds": 100.6, "max_memory_gb": 24.91, "pair": 0.3534163273870945, "sign": 0.0387694031232968, "keep": 0.18897054356057197, "leak": 0.5697673810645938, "replacement": 0.609728142619133}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 3.102586030960083, "loss_scale": 4.0, "retries": 0, "seconds": 101.7, "max_memory_gb": 24.91, "pair": 0.27160472236573696, "sign": 0.03475168254226446, "keep": 0.7331659314222634, "leak": 1.3338413685560226, "replacement": 0.9890238493680954}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 2.6145823001861572, "loss_scale": 4.0, "retries": 0, "seconds": 102.7, "max_memory_gb": 24.91, "pair": 0.4677175208926201, "sign": 0.30721723777242005, "keep": 1.4163293838500977, "leak": 0.6765773685765453, "replacement": 1.4986352026462555}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.8379371166229248, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 24.91, "pair": 0.39782610535621643, "sign": 0.11552224715705961, "keep": 0.9671072978526354, "leak": 0.5707640768960118, "replacement": 1.0818318873643875}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 1.2721278667449951, "loss_scale": 4.0, "retries": 0, "seconds": 104.9, "max_memory_gb": 24.91, "pair": 0.302354346960783, "sign": 0.017706738435663283, "keep": 0.03737335402001918, "leak": 0.9607368595898151, "replacement": 0.5700786113739014}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 1.0255733728408813, "loss_scale": 4.0, "retries": 0, "seconds": 106.0, "max_memory_gb": 24.91, "pair": 0.43252531811594963, "sign": 0.15593825647374615, "keep": 0.030070498430859516, "leak": 0.7909931689500809, "replacement": 0.7232779860496521}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 2.613483428955078, "loss_scale": 4.0, "retries": 0, "seconds": 107.1, "max_memory_gb": 24.91, "pair": 0.6475401818752289, "sign": 0.6941652958630584, "keep": 0.04238144040073166, "leak": 0.24268474825657904, "replacement": 1.0764847919344902}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 3.5690970420837402, "loss_scale": 4.0, "retries": 0, "seconds": 108.1, "max_memory_gb": 24.91, "pair": 0.542979009449482, "sign": 0.9291125233285129, "keep": 0.37865057843737304, "leak": 0.940795011818409, "replacement": 1.432059332728386}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 3.6809451580047607, "loss_scale": 4.0, "retries": 0, "seconds": 109.2, "max_memory_gb": 24.91, "pair": 0.1731370585039258, "sign": 0.010188606946030632, "keep": 0.03546982655825559, "leak": 3.00321476906538, "replacement": 0.9467699304223061}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 1.1764495372772217, "loss_scale": 4.0, "retries": 0, "seconds": 110.3, "max_memory_gb": 24.91, "pair": 0.2798560597002506, "sign": 0.024191282922402024, "keep": 0.0656227616709657, "leak": 0.3883320251479745, "replacement": 0.42184609547257423}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.8349543213844299, "loss_scale": 4.0, "retries": 0, "seconds": 111.4, "max_memory_gb": 24.91, "pair": 0.29803847521543503, "sign": 0.03369372908491641, "keep": 0.02818828178988042, "leak": 0.7264191210269928, "replacement": 0.5105842500925064}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 2.9172067642211914, "loss_scale": 4.0, "retries": 0, "seconds": 112.4, "max_memory_gb": 24.91, "pair": 0.4421111047267914, "sign": 0.39588362781796604, "keep": 0.0068165764496370684, "leak": 0.7245193719863892, "replacement": 0.8245910257101059}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 1.1476163864135742, "loss_scale": 4.0, "retries": 0, "seconds": 113.4, "max_memory_gb": 24.91, "pair": 0.25864996016025543, "sign": 0.022716065286658704, "keep": 0.10362828988581896, "leak": 1.3286185413599014, "replacement": 0.653976783156395}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 1.2531541585922241, "loss_scale": 4.0, "retries": 0, "seconds": 114.4, "max_memory_gb": 24.91, "pair": 0.33804773911833763, "sign": 0.03968671336770058, "keep": 0.06845818219153443, "leak": 0.7167413919232786, "replacement": 0.5713055282831192}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.9077298641204834, "loss_scale": 4.0, "retries": 0, "seconds": 115.5, "max_memory_gb": 24.91, "pair": 0.37590742111206055, "sign": 0.3811760365497321, "keep": 2.2467839462606207, "leak": 0.4252070225775242, "replacement": 1.7961892634630203}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 1.6107070446014404, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 24.91, "pair": 0.4323650151491165, "sign": 0.26637348672375083, "keep": 0.038253881124092004, "leak": 0.5542995585128665, "replacement": 0.7232535630464554}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 1.858938455581665, "loss_scale": 4.0, "retries": 0, "seconds": 117.7, "max_memory_gb": 24.91, "pair": 0.4138353392481804, "sign": 0.0919773387722671, "keep": 0.08554625511169434, "leak": 0.5896672746166587, "replacement": 0.6500139534473419}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 4.200831890106201, "loss_scale": 4.0, "retries": 0, "seconds": 118.8, "max_memory_gb": 24.91, "pair": 0.3625268414616585, "sign": 0.21520765614695847, "keep": 0.47288758493959904, "leak": 0.7359690815210342, "replacement": 0.8905667439103127}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 3.321861982345581, "loss_scale": 4.0, "retries": 0, "seconds": 119.9, "max_memory_gb": 24.91, "pair": 0.4086840348318219, "sign": 1.5195088152540848, "keep": 0.0030406703992866824, "leak": 1.8026190474629402, "replacement": 1.6206136047840118}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 4.054687976837158, "loss_scale": 4.0, "retries": 0, "seconds": 121.0, "max_memory_gb": 24.91, "pair": 0.2070870567113161, "sign": 0.010054697515442967, "keep": 0.005094756510516163, "leak": 2.1414314284920692, "replacement": 0.7500196397304535}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 1.0585137605667114, "loss_scale": 4.0, "retries": 0, "seconds": 122.1, "max_memory_gb": 24.91, "pair": 0.34578029811382294, "sign": 0.04156517377123237, "keep": 0.01277232525171712, "leak": 0.33465573098510504, "replacement": 0.4566129967570305}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.172142505645752, "loss_scale": 4.0, "retries": 0, "seconds": 123.1, "max_memory_gb": 24.91, "pair": 0.2706954590976238, "sign": 0.05027043307200074, "keep": 0.03894834592938423, "leak": 1.2099412754178047, "replacement": 0.6177901774644852}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 1.1659260988235474, "loss_scale": 4.0, "retries": 0, "seconds": 124.2, "max_memory_gb": 24.91, "pair": 0.24824467860162258, "sign": 0.012026338605210185, "keep": 0.02842355681650588, "leak": 0.8268828559666872, "replacement": 0.47519034892320633}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 4.187062740325928, "loss_scale": 4.0, "retries": 0, "seconds": 125.3, "max_memory_gb": 24.91, "pair": 0.44520123675465584, "sign": 0.7278933066409081, "keep": 0.8925373759229842, "leak": 0.6573405861854553, "replacement": 1.4197516217827797}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 6.019753932952881, "loss_scale": 4.0, "retries": 0, "seconds": 126.4, "max_memory_gb": 24.93, "pair": 0.30928342789411545, "sign": 0.38798998261336237, "keep": 1.0958164720836976, "leak": 1.5612314194440842, "replacement": 1.4414945244789124}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 1.532555341720581, "loss_scale": 4.0, "retries": 0, "seconds": 127.5, "max_memory_gb": 24.93, "pair": 0.47221649810671806, "sign": 0.35717565880622715, "keep": 0.48929161829255463, "leak": 0.7214965559542179, "replacement": 1.0758242756128311}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 1.544098138809204, "loss_scale": 4.0, "retries": 0, "seconds": 128.6, "max_memory_gb": 24.93, "pair": 0.44692740589380264, "sign": 0.2955343248322606, "keep": 0.017713204107622005, "leak": 0.426730003207922, "replacement": 0.7102336883544922}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 3.6731114387512207, "loss_scale": 4.0, "retries": 0, "seconds": 129.7, "max_memory_gb": 24.93, "pair": 0.47094255685806274, "sign": 0.8820948172360659, "keep": 1.7356908440124243, "leak": 0.47101191990259395, "replacement": 1.8975883275270462}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 1.8598378896713257, "loss_scale": 4.0, "retries": 0, "seconds": 130.8, "max_memory_gb": 24.93, "pair": 0.6006423756480217, "sign": 0.4185390619095415, "keep": 0.10404306649047612, "leak": 0.31466369703412056, "replacement": 0.940599337220192}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 1.21583092212677, "loss_scale": 4.0, "retries": 0, "seconds": 131.9, "max_memory_gb": 24.93, "pair": 0.41824057325720787, "sign": 0.10898036090657115, "keep": 0.04729924815569575, "leak": 0.6686311303637922, "replacement": 0.6635381430387497}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 3.7164225578308105, "loss_scale": 4.0, "retries": 0, "seconds": 133.0, "max_memory_gb": 24.93, "pair": 0.43956473004072905, "sign": 1.1730663797352463, "keep": 0.3180158889153972, "leak": 0.9120279811322689, "replacement": 1.4131128564476967}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.905483603477478, "loss_scale": 4.0, "retries": 0, "seconds": 134.0, "max_memory_gb": 24.93, "pair": 0.6137754581868649, "sign": 0.5760065759532154, "keep": 1.7199488950136583, "leak": 0.37657239101827145, "replacement": 1.8558962941169739}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 1.9457882642745972, "loss_scale": 4.0, "retries": 0, "seconds": 135.1, "max_memory_gb": 24.93, "pair": 0.4287236388772726, "sign": 0.25548437098041177, "keep": 0.3269997055940621, "leak": 1.0851105019100942, "replacement": 0.9912433177232742}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 3.3027384281158447, "loss_scale": 4.0, "retries": 0, "seconds": 136.2, "max_memory_gb": 24.93, "pair": 0.6699004136025906, "sign": 0.5140186174539849, "keep": 1.0309038301929832, "leak": 1.0580214275760227, "replacement": 1.7068670094013214}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 1.0694621801376343, "loss_scale": 4.0, "retries": 0, "seconds": 137.2, "max_memory_gb": 24.93, "pair": 0.2785356044769287, "sign": 0.010739287245087326, "keep": 0.024149430349325485, "leak": 0.7059322223067284, "replacement": 0.47246304154396057}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 3.8562347888946533, "loss_scale": 4.0, "retries": 0, "seconds": 138.3, "max_memory_gb": 24.93, "pair": 0.528052244335413, "sign": 0.9396564872004092, "keep": 1.265245403687004, "leak": 0.4326715711504221, "replacement": 1.7386710867285728}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 1.3259423971176147, "loss_scale": 4.0, "retries": 0, "seconds": 139.4, "max_memory_gb": 24.93, "pair": 0.40960243344306946, "sign": 0.09500663005746901, "keep": 0.1646482888609171, "leak": 0.4405415372457355, "replacement": 0.6495652869343758}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 1.9972788095474243, "loss_scale": 4.0, "retries": 0, "seconds": 140.5, "max_memory_gb": 24.93, "pair": 0.5518851242959499, "sign": 1.7759046335704625, "keep": 0.07064478460233659, "leak": 0.2774111714679748, "replacement": 1.5445126444101334}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 5.230543613433838, "loss_scale": 4.0, "retries": 0, "seconds": 141.5, "max_memory_gb": 24.93, "pair": 0.7281762957572937, "sign": 0.341538829728961, "keep": 0.3835207244119374, "leak": 0.5325461020693183, "replacement": 1.2238426059484482}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 1.2322065830230713, "loss_scale": 4.0, "retries": 0, "seconds": 142.6, "max_memory_gb": 24.93, "pair": 0.2905115634202957, "sign": 0.030296982440631837, "keep": 0.4390941962665238, "leak": 0.7045866325497627, "replacement": 0.7013538181781769}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 1.6812834739685059, "loss_scale": 4.0, "retries": 0, "seconds": 143.7, "max_memory_gb": 24.93, "pair": 0.5613155215978622, "sign": 0.49056821479462087, "keep": 0.8230503501254134, "leak": 0.2715425053611398, "replacement": 1.2860103994607925}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 2.724658250808716, "loss_scale": 4.0, "retries": 0, "seconds": 144.8, "max_memory_gb": 24.93, "pair": 0.37417713552713394, "sign": 0.3060669509577565, "keep": 0.5371280969261534, "leak": 0.17757286317646503, "replacement": 0.8401678279042244}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 1.26642906665802, "loss_scale": 4.0, "retries": 0, "seconds": 145.9, "max_memory_gb": 24.93, "pair": 0.26085035502910614, "sign": 0.03837483818642795, "keep": 0.1579753588885069, "leak": 0.7956230118870735, "replacement": 0.5579312145709991}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 4.883063316345215, "loss_scale": 4.0, "retries": 0, "seconds": 146.9, "max_memory_gb": 24.93, "pair": 0.25460266694426537, "sign": 0.22575442446395755, "keep": 0.37516088658594526, "leak": 1.6637781858444214, "replacement": 0.971004918217659}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 3.0485987663269043, "loss_scale": 4.0, "retries": 0, "seconds": 148.0, "max_memory_gb": 24.93, "pair": 0.2361427452415228, "sign": 0.019267166790086776, "keep": 0.03660562734876294, "leak": 1.8488797275349498, "replacement": 0.7262990698218346}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 2.8134753704071045, "loss_scale": 4.0, "retries": 0, "seconds": 149.2, "max_memory_gb": 24.93, "pair": 0.37436958961188793, "sign": 0.036338417092338204, "keep": 0.3120726686902344, "leak": 0.46869363193400204, "replacement": 0.6657485365867615}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 2.8296473026275635, "loss_scale": 4.0, "retries": 0, "seconds": 150.3, "max_memory_gb": 24.93, "pair": 0.35645992308855057, "sign": 0.7758504537050612, "keep": 0.027349790470907465, "leak": 1.3997674733400345, "replacement": 1.1080019026994705}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 1.3065638542175293, "loss_scale": 4.0, "retries": 0, "seconds": 151.4, "max_memory_gb": 24.93, "pair": 0.20971788465976715, "sign": 0.03199217893416062, "keep": 0.018716996773946448, "leak": 1.1309322640299797, "replacement": 0.5178055465221405}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 1.0335960388183594, "loss_scale": 4.0, "retries": 0, "seconds": 152.6, "max_memory_gb": 24.93, "pair": 0.33864186704158783, "sign": 0.2296614813967608, "keep": 0.01346135109270108, "leak": 0.8704527732916176, "replacement": 0.6778164803981781}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 2.8041460514068604, "loss_scale": 4.0, "retries": 0, "seconds": 153.6, "max_memory_gb": 24.93, "pair": 0.2906288616359234, "sign": 0.43369505659211427, "keep": 0.36443934338967665, "leak": 0.8499573171138763, "replacement": 0.902185395359993}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 4.285581111907959, "loss_scale": 4.0, "retries": 0, "seconds": 154.8, "max_memory_gb": 24.93, "pair": 0.4347127266228199, "sign": 0.14871118683367968, "keep": 1.4863541302574959, "leak": 0.21535199508070946, "replacement": 1.3060833737254143}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 3.385026454925537, "loss_scale": 4.0, "retries": 0, "seconds": 155.9, "max_memory_gb": 24.93, "pair": 0.4185349605977535, "sign": 0.6860082827042788, "keep": 0.10067321872338653, "leak": 0.6282849349081516, "replacement": 0.9689469635486603}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 0.9929678440093994, "loss_scale": 4.0, "retries": 0, "seconds": 157.0, "max_memory_gb": 24.93, "pair": 0.32683079689741135, "sign": 0.076187803177163, "keep": 0.02257899972209998, "leak": 0.7278052717447281, "replacement": 0.55816550552845}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 3.6484501361846924, "loss_scale": 4.0, "retries": 0, "seconds": 158.1, "max_memory_gb": 24.93, "pair": 0.45163427386432886, "sign": 0.17421144200488925, "keep": 0.021891324780881405, "leak": 1.554620198905468, "replacement": 0.9383406937122345}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 2.310969829559326, "loss_scale": 4.0, "retries": 0, "seconds": 159.2, "max_memory_gb": 24.93, "pair": 0.6591605395078659, "sign": 0.8994787838310003, "keep": 1.1398554370607599, "leak": 0.5147867724299431, "replacement": 1.8075242191553116}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 4.07835054397583, "loss_scale": 4.0, "retries": 0, "seconds": 160.3, "max_memory_gb": 24.93, "pair": 0.35081420093774796, "sign": 0.1657156510045752, "keep": 0.005783470040029215, "leak": 0.8096556700766087, "replacement": 0.638977661728859}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 1.3288289308547974, "loss_scale": 4.0, "retries": 0, "seconds": 161.4, "max_memory_gb": 24.93, "pair": 0.2891420852392912, "sign": 0.08687075390480459, "keep": 0.049861780455103144, "leak": 0.9175247028470039, "replacement": 0.5868895277380943}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 1.193445086479187, "loss_scale": 4.0, "retries": 0, "seconds": 162.4, "max_memory_gb": 24.93, "pair": 0.33917665481567383, "sign": 0.13270351674873382, "keep": 0.17403953755274415, "leak": 0.5453286473639309, "replacement": 0.6288803443312645}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 1.1364271640777588, "loss_scale": 4.0, "retries": 0, "seconds": 163.6, "max_memory_gb": 24.93, "pair": 0.34255581721663475, "sign": 0.02782122523058206, "keep": 0.017507838494566386, "leak": 0.6540457159280777, "replacement": 0.5287317782640457}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 1.0524426698684692, "loss_scale": 4.0, "retries": 0, "seconds": 164.7, "max_memory_gb": 24.93, "pair": 0.22804156690835953, "sign": 0.008357671438716352, "keep": 0.04537333872212912, "leak": 0.9813272580504417, "replacement": 0.5002388656139374}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 2.34171199798584, "loss_scale": 4.0, "retries": 0, "seconds": 165.8, "max_memory_gb": 24.93, "pair": 0.24884680286049843, "sign": 0.011378215276636183, "keep": 0.020689811548891157, "leak": 1.236374944448471, "replacement": 0.5739745423197746}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 1.9993990659713745, "loss_scale": 4.0, "retries": 0, "seconds": 166.8, "max_memory_gb": 24.93, "pair": 0.13861635141074657, "sign": 0.008888072101399302, "keep": 0.3424466662108898, "leak": 1.7980482876300812, "replacement": 0.763795793056488}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 1.184498906135559, "loss_scale": 4.0, "retries": 0, "seconds": 168.0, "max_memory_gb": 24.93, "pair": 0.21212404407560825, "sign": 0.00961054116487503, "keep": 0.45747234754401234, "leak": 0.7844105139374733, "replacement": 0.6417680978775024}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 1.985670804977417, "loss_scale": 4.0, "retries": 0, "seconds": 169.0, "max_memory_gb": 24.93, "pair": 0.30387186631560326, "sign": 0.006324557005427778, "keep": 0.03155219216580463, "leak": 0.48527682200074196, "replacement": 0.44412943720817566}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 3.8058700561523438, "loss_scale": 4.0, "retries": 0, "seconds": 170.1, "max_memory_gb": 24.93, "pair": 0.3134430325590074, "sign": 1.0892468402162194, "keep": 0.09617990866536275, "leak": 3.1508861798793077, "replacement": 1.69387786090374}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 2.5355281829833984, "loss_scale": 4.0, "retries": 0, "seconds": 171.2, "max_memory_gb": 24.93, "pair": 0.4485981948673725, "sign": 0.3766232007765211, "keep": 0.4648508028994911, "leak": 0.42905305279418826, "replacement": 0.976598471403122}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 1.3076965808868408, "loss_scale": 4.0, "retries": 0, "seconds": 172.2, "max_memory_gb": 24.99, "pair": 0.2940502427518368, "sign": 0.016818295931443572, "keep": 0.003081015922361985, "leak": 0.8783110864460468, "replacement": 0.5235776752233505}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 3.7664358615875244, "loss_scale": 4.0, "retries": 0, "seconds": 173.3, "max_memory_gb": 24.99, "pair": 0.22144816257059574, "sign": 0.1467078490240965, "keep": 0.5245042748283595, "leak": 1.0973334144800901, "replacement": 0.8313875608146191}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 1.729967474937439, "loss_scale": 4.0, "retries": 0, "seconds": 174.4, "max_memory_gb": 24.99, "pair": 0.41211695596575737, "sign": 0.15593015006743371, "keep": 0.2260102191180522, "leak": 0.3094064365141094, "replacement": 0.6804387643933296}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.83511221408844, "loss_scale": 4.0, "retries": 0, "seconds": 175.5, "max_memory_gb": 24.99, "pair": 0.2401308063417673, "sign": 0.037572961940895766, "keep": 0.2242560943054741, "leak": 1.0927541702985764, "replacement": 0.6442338675260544}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 12.363283157348633, "loss_scale": 4.0, "retries": 0, "seconds": 176.5, "max_memory_gb": 24.99, "pair": 0.9788151532411575, "sign": 0.6186982106883079, "keep": 0.9413886107504368, "leak": 1.1784784680930898, "replacement": 2.0534781143069267}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 5.020566940307617, "loss_scale": 4.0, "retries": 0, "seconds": 177.7, "max_memory_gb": 24.99, "pair": 0.5817694962024689, "sign": 0.3549208453623578, "keep": 0.2679030189756304, "leak": 0.39315043756505474, "replacement": 0.9914690032601357}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 1.1555222272872925, "loss_scale": 4.0, "retries": 0, "seconds": 178.8, "max_memory_gb": 24.99, "pair": 0.27640434354543686, "sign": 0.02555088623194024, "keep": 0.004771091381826409, "leak": 0.47965491749346256, "replacement": 0.41147907078266144}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 7.497646808624268, "loss_scale": 4.0, "retries": 0, "seconds": 179.8, "max_memory_gb": 24.99, "pair": 0.7848343215882778, "sign": 0.7235962064005435, "keep": 1.1378323092940263, "leak": 0.8868104907451198, "replacement": 1.9372511208057404}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 2.0417230129241943, "loss_scale": 4.0, "retries": 0, "seconds": 180.9, "max_memory_gb": 24.99, "pair": 0.3921002969145775, "sign": 0.07392324972897768, "keep": 0.08019361412152648, "leak": 0.40396696142852306, "replacement": 0.5701504498720169}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.8404585123062134, "loss_scale": 4.0, "retries": 0, "seconds": 182.0, "max_memory_gb": 24.99, "pair": 0.2766610737890005, "sign": 0.017286977090407163, "keep": 0.08653535446865135, "leak": 0.8852465599775314, "replacement": 0.5498838871717453}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.7634210586547852, "loss_scale": 4.0, "retries": 0, "seconds": 183.1, "max_memory_gb": 24.99, "pair": 0.46562105044722557, "sign": 0.16707195434719324, "keep": 0.21189073706045747, "leak": 0.3617795594036579, "replacement": 0.7455473095178604}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 1.6742719411849976, "loss_scale": 4.0, "retries": 0, "seconds": 184.2, "max_memory_gb": 24.99, "pair": 0.2743111327290535, "sign": 0.0064439522684551775, "keep": 0.1397808197998529, "leak": 1.0640602111816406, "replacement": 0.6134385764598846}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 1.982043743133545, "loss_scale": 4.0, "retries": 0, "seconds": 185.3, "max_memory_gb": 24.99, "pair": 0.3241423089057207, "sign": 0.14557083987165242, "keep": 0.09257607080314756, "leak": 0.254929818212986, "replacement": 0.5069482289254665}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 1.1445151567459106, "loss_scale": 4.0, "retries": 0, "seconds": 186.4, "max_memory_gb": 24.99, "pair": 0.15897092781960964, "sign": 0.00411990872817114, "keep": 0.03367975656874478, "leak": 0.741442084312439, "replacement": 0.36323128640651703}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.906279444694519, "loss_scale": 4.0, "retries": 0, "seconds": 187.4, "max_memory_gb": 24.99, "pair": 0.23009561188519, "sign": 0.09614683655672707, "keep": 0.12514976819511503, "leak": 0.7172860510181636, "replacement": 0.5200654119253159}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 1.2286524772644043, "loss_scale": 4.0, "retries": 0, "seconds": 188.5, "max_memory_gb": 24.99, "pair": 0.2366097755730152, "sign": 0.01990867091808468, "keep": 0.0803719137495591, "leak": 0.37574900686740875, "replacement": 0.38068731874227524}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 4.2647576332092285, "loss_scale": 4.0, "retries": 0, "seconds": 189.8, "max_memory_gb": 24.99, "pair": 0.3731683362275362, "sign": 0.2722278961446136, "keep": 1.306439084932208, "leak": 0.6584702329710126, "replacement": 1.3271193578839302}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 1.0375542640686035, "loss_scale": 4.0, "retries": 0, "seconds": 190.9, "max_memory_gb": 24.99, "pair": 0.23104071244597435, "sign": 0.03964825233560987, "keep": 0.020476743578910828, "leak": 0.45497722338768654, "replacement": 0.37484750896692276}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 1.6760914325714111, "loss_scale": 4.0, "retries": 0, "seconds": 192.0, "max_memory_gb": 24.99, "pair": 0.43331005051732063, "sign": 0.4676265332964249, "keep": 0.00913747777667595, "leak": 0.3693727634381503, "replacement": 0.7640352845191956}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 3.7500810623168945, "loss_scale": 4.0, "retries": 0, "seconds": 193.1, "max_memory_gb": 24.99, "pair": 0.2140225013718009, "sign": 0.09733506082557142, "keep": 0.2732184684209642, "leak": 1.5806738324463367, "replacement": 0.7944677174091339}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 1.8318873643875122, "loss_scale": 4.0, "retries": 0, "seconds": 194.1, "max_memory_gb": 24.99, "pair": 0.6923937350511551, "sign": 0.5702927092206664, "keep": 0.6247464066691464, "leak": 0.20215080212801695, "replacement": 1.3404510393738747}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 0.9772827625274658, "loss_scale": 4.0, "retries": 0, "seconds": 195.2, "max_memory_gb": 24.99, "pair": 0.42059263959527016, "sign": 0.04382743185851723, "keep": 0.07488440105225891, "leak": 0.481336077908054, "replacement": 0.6002825796604156}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 2.469465970993042, "loss_scale": 4.0, "retries": 0, "seconds": 196.2, "max_memory_gb": 24.99, "pair": 0.18559584766626358, "sign": 0.015292140538804233, "keep": 0.06957277865149081, "leak": 2.0368454456329346, "replacement": 0.7372396886348724}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 4.148369789123535, "loss_scale": 4.0, "retries": 0, "seconds": 197.3, "max_memory_gb": 24.99, "pair": 0.4791129529476166, "sign": 0.31713391258381307, "keep": 0.024831723087118007, "leak": 1.418541144579649, "replacement": 1.00473103672266}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 0.8816785216331482, "loss_scale": 4.0, "retries": 0, "seconds": 198.4, "max_memory_gb": 24.99, "pair": 0.32513774931430817, "sign": 0.08334699302213266, "keep": 0.010504899315037619, "leak": 0.7418855167925358, "replacement": 0.5575350672006607}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 1.3782780170440674, "loss_scale": 4.0, "retries": 0, "seconds": 199.6, "max_memory_gb": 24.99, "pair": 0.28254637494683266, "sign": 0.008981412567663938, "keep": 0.020976212013920303, "leak": 0.7028708010911942, "replacement": 0.4732428789138794}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 1.3064557313919067, "loss_scale": 4.0, "retries": 0, "seconds": 200.7, "max_memory_gb": 24.99, "pair": 0.3010259550064802, "sign": 0.03992174629820511, "keep": 0.14196350472047925, "leak": 0.8695204071700573, "replacement": 0.6093486919999123}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 2.22420334815979, "loss_scale": 4.0, "retries": 0, "seconds": 201.8, "max_memory_gb": 24.99, "pair": 0.43661223724484444, "sign": 0.02706250798655674, "keep": 0.8571585751910789, "leak": 0.4008036106824875, "replacement": 0.9789236634969711}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 1.1716539859771729, "loss_scale": 4.0, "retries": 0, "seconds": 202.9, "max_memory_gb": 24.99, "pair": 0.29500249586999416, "sign": 0.009634711313992739, "keep": 0.46586935652885586, "leak": 0.779317220672965, "replacement": 0.7275838479399681}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 2.3495354652404785, "loss_scale": 4.0, "retries": 0, "seconds": 203.9, "max_memory_gb": 24.99, "pair": 0.5926392152905464, "sign": 0.8545813377131708, "keep": 0.7122217144424212, "leak": 0.2811112813651562, "replacement": 1.4463184960186481}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 3.7598488330841064, "loss_scale": 4.0, "retries": 0, "seconds": 205.0, "max_memory_gb": 24.99, "pair": 0.33589346520602703, "sign": 0.3942911250051111, "keep": 0.8193914893781766, "leak": 1.607711043174504, "replacement": 1.3446625173091888}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 1.4050366878509521, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 24.99, "pair": 0.19082124903798103, "sign": 0.008124373620375991, "keep": 0.01612045877845958, "leak": 1.1482974141836166, "replacement": 0.49001801759004593}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 8.395153045654297, "loss_scale": 4.0, "retries": 0, "seconds": 207.1, "max_memory_gb": 24.99, "pair": 0.5257293079048395, "sign": 0.4240441604051739, "keep": 0.5857828713913023, "leak": 0.8973417282104492, "replacement": 1.2549782916903496}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.2581760883331299, "loss_scale": 4.0, "retries": 0, "seconds": 208.1, "max_memory_gb": 24.99, "pair": 0.5972996801137924, "sign": 0.488755667116493, "keep": 1.1125326175242662, "leak": 0.5088561633601785, "replacement": 1.5251579135656357}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 1.3761824369430542, "loss_scale": 4.0, "retries": 0, "seconds": 209.1, "max_memory_gb": 24.99, "pair": 0.31209372356534004, "sign": 0.032937017269432545, "keep": 0.03414852600872109, "leak": 0.6709902733564377, "replacement": 0.5133840590715408}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 2.062492847442627, "loss_scale": 4.0, "retries": 0, "seconds": 210.2, "max_memory_gb": 24.99, "pair": 0.5228137522935867, "sign": 0.3200417053885758, "keep": 0.024076367411907995, "leak": 0.41804844699800014, "replacement": 0.7993848994374275}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 2.010516405105591, "loss_scale": 4.0, "retries": 0, "seconds": 211.2, "max_memory_gb": 24.99, "pair": 0.39275528863072395, "sign": 0.07077908486826345, "keep": 0.010677798869437538, "leak": 1.005004070699215, "replacement": 0.6847347468137741}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 1.2690080404281616, "loss_scale": 4.0, "retries": 0, "seconds": 212.4, "max_memory_gb": 24.99, "pair": 0.14749355614185333, "sign": 0.00850968511076644, "keep": 0.024042396398954224, "leak": 1.1161699146032333, "replacement": 0.44281207397580147}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.2366952896118164, "loss_scale": 4.0, "retries": 0, "seconds": 213.5, "max_memory_gb": 24.99, "pair": 0.2860703580081463, "sign": 0.03664178459439427, "keep": 0.009499028892605565, "leak": 0.8494780138134956, "replacement": 0.5215102657675743}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 2.5155980587005615, "loss_scale": 4.0, "retries": 0, "seconds": 214.6, "max_memory_gb": 24.99, "pair": 0.23835106939077377, "sign": 0.08098439080640674, "keep": 0.0960849372568191, "leak": 0.9375838525593281, "replacement": 0.5612816959619522}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 3.3235890865325928, "loss_scale": 4.0, "retries": 0, "seconds": 215.7, "max_memory_gb": 24.99, "pair": 0.17599445302039385, "sign": 0.034568350500194356, "keep": 0.5687340884032892, "leak": 1.573245532810688, "replacement": 0.870957039296627}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 2.677016019821167, "loss_scale": 4.0, "retries": 0, "seconds": 216.8, "max_memory_gb": 24.99, "pair": 0.44903623778373003, "sign": 0.3678488750301767, "keep": 1.2037163924542256, "leak": 1.039696453139186, "replacement": 1.494742974638939}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 2.8060665130615234, "loss_scale": 4.0, "retries": 0, "seconds": 217.9, "max_memory_gb": 24.99, "pair": 0.42836498096585274, "sign": 0.2014997557271272, "keep": 0.19323230828376836, "leak": 0.42390753142535686, "replacement": 0.7317078933119774}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 1.2002732753753662, "loss_scale": 4.0, "retries": 0, "seconds": 219.0, "max_memory_gb": 24.99, "pair": 0.3927311599254608, "sign": 0.14990553091047332, "keep": 0.007730083510068653, "leak": 0.7046972066164017, "replacement": 0.6477232575416565}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 1.0474112033843994, "loss_scale": 4.0, "retries": 0, "seconds": 220.1, "max_memory_gb": 24.99, "pair": 0.15404482372105122, "sign": 0.04516937187872827, "keep": 0.01009875995805487, "leak": 0.8938828110694885, "replacement": 0.40514960139989853}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 0.8149201273918152, "loss_scale": 4.0, "retries": 0, "seconds": 221.2, "max_memory_gb": 24.99, "pair": 0.3718704879283905, "sign": 0.028529081493616104, "keep": 0.053709254949353635, "leak": 0.8863686682889238, "replacement": 0.6345818117260933}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 0.8895628452301025, "loss_scale": 4.0, "retries": 0, "seconds": 222.2, "max_memory_gb": 24.99, "pair": 0.245581628754735, "sign": 0.008125437656417489, "keep": 0.0030012432289368007, "leak": 0.5750560015439987, "replacement": 0.39490897208452225}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 1.2300633192062378, "loss_scale": 4.0, "retries": 0, "seconds": 223.3, "max_memory_gb": 24.99, "pair": 0.26778467185795307, "sign": 0.03904356784187257, "keep": 0.04004437103003511, "leak": 0.4835614189505577, "replacement": 0.4282190129160881}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 1.1300227642059326, "loss_scale": 4.0, "retries": 0, "seconds": 224.4, "max_memory_gb": 24.99, "pair": 0.48043570667505264, "sign": 0.1702093841158785, "keep": 0.08346565888086843, "leak": 0.3058642678661272, "replacement": 0.6837393417954445}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.9903798699378967, "loss_scale": 4.0, "retries": 0, "seconds": 225.5, "max_memory_gb": 24.99, "pair": 0.16263315454125404, "sign": 0.004163047473412007, "keep": 0.030736499807517248, "leak": 0.657559422776103, "replacement": 0.3444727808237076}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 2.0491628646850586, "loss_scale": 4.0, "retries": 0, "seconds": 226.6, "max_memory_gb": 24.99, "pair": 0.49101484939455986, "sign": 0.3926476505585015, "keep": 0.3188402026571566, "leak": 0.9482863899320364, "replacement": 1.0838303864002228}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 2.835543632507324, "loss_scale": 4.0, "retries": 0, "seconds": 227.7, "max_memory_gb": 24.99, "pair": 0.5375588834285736, "sign": 0.49868948257062584, "keep": 1.3273123729718463, "leak": 0.45124703738838434, "replacement": 1.5633715391159058}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 0.9926357269287109, "loss_scale": 4.0, "retries": 0, "seconds": 228.8, "max_memory_gb": 24.99, "pair": 0.49684442579746246, "sign": 0.2651340625016019, "keep": 0.09437710233032703, "leak": 0.6011285916902125, "replacement": 0.8268821537494659}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 6.977097988128662, "loss_scale": 4.0, "retries": 0, "seconds": 230.0, "max_memory_gb": 24.99, "pair": 0.7231140099465847, "sign": 0.24852213641861454, "keep": 0.1210326119326055, "leak": 1.378215602133423, "replacement": 1.252445325255394}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 2.227473497390747, "loss_scale": 4.0, "retries": 0, "seconds": 231.1, "max_memory_gb": 24.99, "pair": 0.4637271389365196, "sign": 0.3504747621482238, "keep": 0.541645475444966, "leak": 0.5293718059547246, "replacement": 1.042130209505558}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 1.433350920677185, "loss_scale": 4.0, "retries": 0, "seconds": 232.2, "max_memory_gb": 24.99, "pair": 0.4955456294119358, "sign": 0.3860320364474319, "keep": 1.2428011884794614, "leak": 0.0678311123047024, "replacement": 1.326919972896576}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 2.8206026554107666, "loss_scale": 4.0, "retries": 0, "seconds": 233.3, "max_memory_gb": 24.99, "pair": 0.7728674449026585, "sign": 0.8214359041303396, "keep": 1.3611158429516763, "leak": 0.07518915338636134, "replacement": 1.8829406052827835}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 0.9174391627311707, "loss_scale": 4.0, "retries": 0, "seconds": 234.4, "max_memory_gb": 24.99, "pair": 0.17493557929992676, "sign": 0.030091025168076158, "keep": 0.01958649525840883, "leak": 0.8337089195847511, "replacement": 0.40820156782865524}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 1.2611151933670044, "loss_scale": 4.0, "retries": 0, "seconds": 235.5, "max_memory_gb": 24.99, "pair": 0.21983375772833824, "sign": 0.022175160353071988, "keep": 0.0006954239488550229, "leak": 0.9773836508393288, "replacement": 0.4756149649620056}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 8.273880004882812, "loss_scale": 4.0, "retries": 0, "seconds": 236.6, "max_memory_gb": 24.99, "pair": 1.1362624764442444, "sign": 1.8333762288093567, "keep": 0.4450632850639522, "leak": 0.5772938574664295, "replacement": 2.4198057055473328}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 0.973691463470459, "loss_scale": 4.0, "retries": 0, "seconds": 237.7, "max_memory_gb": 24.99, "pair": 0.3146728128194809, "sign": 0.0845340231899172, "keep": 0.0047265142275136895, "leak": 0.512618824839592, "replacement": 0.4874577820301056}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 1.4642994403839111, "loss_scale": 4.0, "retries": 0, "seconds": 238.8, "max_memory_gb": 24.99, "pair": 0.3964524269104004, "sign": 0.055544751056004316, "keep": 0.06559567411204625, "leak": 0.23486266657710075, "replacement": 0.5157383158802986}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.8963322639465332, "loss_scale": 4.0, "retries": 0, "seconds": 239.9, "max_memory_gb": 24.99, "pair": 0.3693454675376415, "sign": 0.17790049209725112, "keep": 0.012877494329586625, "leak": 0.7536231745034456, "replacement": 0.6531402617692947}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 1.9718118906021118, "loss_scale": 4.0, "retries": 0, "seconds": 241.0, "max_memory_gb": 24.99, "pair": 0.2758192718029022, "sign": 0.029670532152522355, "keep": 0.056898416793046636, "leak": 0.5292431870475411, "replacement": 0.4514145404100418}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 5.309587478637695, "loss_scale": 4.0, "retries": 0, "seconds": 242.1, "max_memory_gb": 24.99, "pair": 0.5834416970610619, "sign": 0.09785410697804764, "keep": 0.018058152605590294, "leak": 0.7232361994683743, "replacement": 0.8222068697214127}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 1.5152078866958618, "loss_scale": 4.0, "retries": 0, "seconds": 243.2, "max_memory_gb": 24.99, "pair": 0.39440203458070755, "sign": 0.04787377826869488, "keep": 0.008997296172310598, "leak": 0.3068964993581176, "replacement": 0.49956168234348297}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 1.34420645236969, "loss_scale": 4.0, "retries": 0, "seconds": 244.3, "max_memory_gb": 24.99, "pair": 0.2567629776895046, "sign": 0.03627667436376214, "keep": 0.003438861433096463, "leak": 0.8082209415733814, "replacement": 0.478675976395607}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 6.642570972442627, "loss_scale": 4.0, "retries": 0, "seconds": 245.4, "max_memory_gb": 24.99, "pair": 0.2684350134804845, "sign": 0.036483674542978406, "keep": 0.8457033586746547, "leak": 0.9200051221996546, "replacement": 0.9395297914743423}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 1.3736262321472168, "loss_scale": 4.0, "retries": 0, "seconds": 246.4, "max_memory_gb": 24.99, "pair": 0.23158452659845352, "sign": 0.016892935382202268, "keep": 0.05601030691605047, "leak": 0.566243963316083, "replacement": 0.40959715098142624}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 1.0417975187301636, "loss_scale": 4.0, "retries": 0, "seconds": 247.5, "max_memory_gb": 24.99, "pair": 0.23069331422448158, "sign": 0.05388856143690646, "keep": 0.0017416144314665871, "leak": 1.0159521028399467, "replacement": 0.5124964267015457}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 1.7731822729110718, "loss_scale": 4.0, "retries": 0, "seconds": 248.6, "max_memory_gb": 24.99, "pair": 0.3830191530287266, "sign": 0.1205246530007571, "keep": 0.1103319058883585, "leak": 0.9129511440114584, "replacement": 0.7266852110624313}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 39.756385803222656, "loss_scale": 4.0, "retries": 0, "seconds": 249.6, "max_memory_gb": 24.99, "pair": 1.0833229720592499, "sign": 0.649001732468605, "keep": 0.9139517678577249, "leak": 0.510363171575591, "replacement": 1.992390513420105}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 1.6890212297439575, "loss_scale": 4.0, "retries": 0, "seconds": 250.7, "max_memory_gb": 24.99, "pair": 0.5734516717493534, "sign": 0.7052554795518517, "keep": 0.03385873494198677, "leak": 0.44706688076257706, "replacement": 1.054775483906269}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 2.685453176498413, "loss_scale": 4.0, "retries": 0, "seconds": 251.8, "max_memory_gb": 24.99, "pair": 0.20572122558951378, "sign": 0.02415109591674991, "keep": 0.05330219061579555, "leak": 1.7820310518145561, "replacement": 0.6899556368589401}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 1.0131847858428955, "loss_scale": 4.0, "retries": 0, "seconds": 252.9, "max_memory_gb": 24.99, "pair": 0.27292584627866745, "sign": 0.00877245661104098, "keep": 0.012286557350307703, "leak": 0.9153244532644749, "replacement": 0.5122864693403244}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 2.6371700763702393, "loss_scale": 4.0, "retries": 0, "seconds": 254.0, "max_memory_gb": 24.99, "pair": 0.22101532109081745, "sign": 0.0248621697537601, "keep": 0.05236437737039523, "leak": 0.7801490761339664, "replacement": 0.454665869474411}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 9.997612953186035, "loss_scale": 4.0, "retries": 0, "seconds": 255.1, "max_memory_gb": 24.99, "pair": 1.0767176058143377, "sign": 0.6876704229507595, "keep": 1.3044942498563614, "leak": 2.022232748568058, "replacement": 2.578358009457588}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 1.1501895189285278, "loss_scale": 4.0, "retries": 0, "seconds": 256.2, "max_memory_gb": 24.99, "pair": 0.34980154410004616, "sign": 0.10414957979810424, "keep": 0.005787847787132705, "leak": 0.19838217552751303, "replacement": 0.4543657973408699}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 5.138010501861572, "loss_scale": 4.0, "retries": 0, "seconds": 257.2, "max_memory_gb": 24.99, "pair": 0.4382684752345085, "sign": 0.42636554641649127, "keep": 0.14265052443442983, "leak": 0.3259022943675518, "replacement": 0.8042520508170128}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 1.977783203125, "loss_scale": 4.0, "retries": 0, "seconds": 258.3, "max_memory_gb": 24.99, "pair": 0.24123477563261986, "sign": 0.01903329137712717, "keep": 0.034335087053477764, "leak": 0.7060666233301163, "replacement": 0.44443561881780624}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 2.5061237812042236, "loss_scale": 4.0, "retries": 0, "seconds": 259.4, "max_memory_gb": 24.99, "pair": 0.6328091993927956, "sign": 1.4336691470816731, "keep": 0.16068989378982224, "leak": 0.44574308009350716, "replacement": 1.5414244532585144}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 1.1575303077697754, "loss_scale": 4.0, "retries": 0, "seconds": 260.4, "max_memory_gb": 24.99, "pair": 0.40914881229400635, "sign": 0.33875303337117657, "keep": 0.1996708688857325, "leak": 0.5939733497798443, "replacement": 0.8268541246652603}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 6.019176959991455, "loss_scale": 4.0, "retries": 0, "seconds": 261.6, "max_memory_gb": 24.99, "pair": 0.9024174474179745, "sign": 0.6317698795464821, "keep": 0.3639980854893565, "leak": 0.500407219864428, "replacement": 1.5254032388329506}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 1.487187385559082, "loss_scale": 4.0, "retries": 0, "seconds": 262.6, "max_memory_gb": 24.99, "pair": 0.42722487077116966, "sign": 0.050640727393329144, "keep": 0.004231036638657315, "leak": 0.6532794162631035, "replacement": 0.617980606853962}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 1.2078847885131836, "loss_scale": 4.0, "retries": 0, "seconds": 263.7, "max_memory_gb": 24.99, "pair": 0.24766818061470985, "sign": 0.021910291514359415, "keep": 0.059002846508519724, "leak": 0.9007081463932991, "replacement": 0.513301782310009}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 2.4074032306671143, "loss_scale": 4.0, "retries": 0, "seconds": 264.8, "max_memory_gb": 24.99, "pair": 0.3866242989897728, "sign": 0.18010614393278956, "keep": 0.026819340884685516, "leak": 1.1639584824442863, "replacement": 0.7810766249895096}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 4.175290584564209, "loss_scale": 4.0, "retries": 0, "seconds": 265.9, "max_memory_gb": 24.99, "pair": 0.6567663475871086, "sign": 0.40703338012099266, "keep": 0.7492658756673336, "leak": 0.3459197459742427, "replacement": 1.3213958963751793}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.8467075824737549, "loss_scale": 4.0, "retries": 0, "seconds": 267.1, "max_memory_gb": 24.99, "pair": 0.22003832645714283, "sign": 0.006694875308312476, "keep": 0.006287017838985776, "leak": 0.5727590546011925, "replacement": 0.369719035923481}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 3.2035775184631348, "loss_scale": 4.0, "retries": 0, "seconds": 268.2, "max_memory_gb": 24.99, "pair": 0.26987424492836, "sign": 0.02557332639116794, "keep": 0.29556808683423696, "leak": 0.8317966461181641, "replacement": 0.6383941173553467}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 0.8486275672912598, "loss_scale": 4.0, "retries": 0, "seconds": 269.2, "max_memory_gb": 24.99, "pair": 0.24245628714561462, "sign": 0.036714589688926935, "keep": 0.04881329067211482, "leak": 0.6852042805403471, "replacement": 0.45652129501104355}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 1.6075814962387085, "loss_scale": 4.0, "retries": 0, "seconds": 270.4, "max_memory_gb": 24.99, "pair": 0.2608849871903658, "sign": 0.08294130954891443, "keep": 0.05676294007571414, "leak": 0.8112590275704861, "replacement": 0.5335518717765808}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 1.9267163276672363, "loss_scale": 4.0, "retries": 0, "seconds": 271.4, "max_memory_gb": 24.99, "pair": 0.4951863959431648, "sign": 0.04435571865178645, "keep": 0.11805228857338079, "leak": 0.2158895842730999, "replacement": 0.6303628012537956}
|
runs/single_internvl/replacement/L02_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/replacement/L02_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.2.mlp.up_proj",
|
| 35 |
+
"model.language_model.layers.2.mlp.down_proj",
|
| 36 |
+
"model.language_model.layers.2.self_attn.o_proj",
|
| 37 |
+
"model.language_model.layers.2.mlp.gate_proj",
|
| 38 |
+
"model.language_model.layers.2.self_attn.k_proj",
|
| 39 |
+
"model.language_model.layers.2.self_attn.v_proj",
|
| 40 |
+
"model.language_model.layers.2.self_attn.q_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
runs/single_internvl/replacement/L02_s13/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"behavior": "replacement",
|
| 3 |
+
"seed": 13,
|
| 4 |
+
"model": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 5 |
+
"backbone": "internvl",
|
| 6 |
+
"train": "/data/OPSD/assets/v3/replacement.jsonl",
|
| 7 |
+
"items": 1000,
|
| 8 |
+
"steps": 250,
|
| 9 |
+
"epochs": 1,
|
| 10 |
+
"steps_per_epoch": 250,
|
| 11 |
+
"groups_per_step": 4,
|
| 12 |
+
"lr": 5e-05,
|
| 13 |
+
"scheduler": "cosine",
|
| 14 |
+
"warmup_steps": 16,
|
| 15 |
+
"final_fraction": 0.1,
|
| 16 |
+
"weight_decay": 0.0,
|
| 17 |
+
"clip": 1.0,
|
| 18 |
+
"initial_loss_scale": 4.0,
|
| 19 |
+
"adapter": {
|
| 20 |
+
"layer": 2,
|
| 21 |
+
"rank": 16,
|
| 22 |
+
"trainable_parameters": 1212416
|
| 23 |
+
}
|
| 24 |
+
}
|
runs/single_internvl/replacement/L02_s13/steps.jsonl
ADDED
|
@@ -0,0 +1,250 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"step": 1, "steps": 250, "lr": 3.125e-06, "grad_norm": 0.9583612084388733, "loss_scale": 4.0, "retries": 0, "seconds": 1.1, "max_memory_gb": 24.6, "pair": 1.0986123085021973, "sign": 0.7328193015418947, "keep": 1.137254401575774, "leak": 0.0, "replacement": 2.0336491465568542}
|
| 2 |
+
{"step": 2, "steps": 250, "lr": 6.25e-06, "grad_norm": 0.9831684231758118, "loss_scale": 4.0, "retries": 0, "seconds": 2.1, "max_memory_gb": 24.6, "pair": 1.096964567899704, "sign": 0.1813456095987931, "keep": 1.368505568461842, "leak": 2.480725449061083e-05, "replacement": 1.8718963265419006}
|
| 3 |
+
{"step": 3, "steps": 250, "lr": 9.375000000000001e-06, "grad_norm": 2.420527696609497, "loss_scale": 4.0, "retries": 0, "seconds": 3.1, "max_memory_gb": 24.6, "pair": 1.0907756984233856, "sign": 0.11447693407535553, "keep": 2.064627856016159, "leak": 2.4098124242755148e-05, "replacement": 2.1803340315818787}
|
| 4 |
+
{"step": 4, "steps": 250, "lr": 1.25e-05, "grad_norm": 1.5746196508407593, "loss_scale": 4.0, "retries": 0, "seconds": 4.1, "max_memory_gb": 24.6, "pair": 1.0951945185661316, "sign": 0.03809042857028544, "keep": 0.8673102994653163, "leak": 6.120599465475607e-05, "replacement": 1.5479102432727814}
|
| 5 |
+
{"step": 5, "steps": 250, "lr": 1.5625e-05, "grad_norm": 0.6794475317001343, "loss_scale": 4.0, "retries": 0, "seconds": 5.1, "max_memory_gb": 24.6, "pair": 1.097258299589157, "sign": 0.11683737323619425, "keep": 0.023240134498337284, "leak": 1.527200950590668e-05, "replacement": 1.1673009097576141}
|
| 6 |
+
{"step": 6, "steps": 250, "lr": 1.8750000000000002e-05, "grad_norm": 0.9000606536865234, "loss_scale": 4.0, "retries": 0, "seconds": 6.1, "max_memory_gb": 24.6, "pair": 1.0976634919643402, "sign": 0.19926619064062834, "keep": 0.6332892416394316, "leak": 4.5745800861141106e-05, "replacement": 1.513952612876892}
|
| 7 |
+
{"step": 7, "steps": 250, "lr": 2.1875e-05, "grad_norm": 1.8879647254943848, "loss_scale": 4.0, "retries": 0, "seconds": 7.0, "max_memory_gb": 24.6, "pair": 1.0944688618183136, "sign": 0.08967955759726465, "keep": 1.524530688933737, "leak": 5.184311477535175e-05, "replacement": 1.9015869200229645}
|
| 8 |
+
{"step": 8, "steps": 250, "lr": 2.5e-05, "grad_norm": 2.045976161956787, "loss_scale": 4.0, "retries": 0, "seconds": 8.0, "max_memory_gb": 24.64, "pair": 1.0944971144199371, "sign": 0.20976255694404244, "keep": 1.4593652412295341, "leak": 0.00013179336474422598, "replacement": 1.929093986749649}
|
| 9 |
+
{"step": 9, "steps": 250, "lr": 2.8125000000000003e-05, "grad_norm": 3.6866044998168945, "loss_scale": 4.0, "retries": 0, "seconds": 9.0, "max_memory_gb": 24.64, "pair": 1.100559264421463, "sign": 0.5766326216980815, "keep": 1.3853222876787186, "leak": 7.79694858010771e-07, "replacement": 2.0815370678901672}
|
| 10 |
+
{"step": 10, "steps": 250, "lr": 3.125e-05, "grad_norm": 0.8040004968643188, "loss_scale": 4.0, "retries": 0, "seconds": 10.0, "max_memory_gb": 24.64, "pair": 1.092244178056717, "sign": 0.08064881293103099, "keep": 0.09780377201968804, "leak": 0.00015258977146004327, "replacement": 1.1815086007118225}
|
| 11 |
+
{"step": 11, "steps": 250, "lr": 3.4375e-05, "grad_norm": 1.7379133701324463, "loss_scale": 4.0, "retries": 0, "seconds": 11.0, "max_memory_gb": 24.64, "pair": 1.085358202457428, "sign": 0.20401584031060338, "keep": 1.3154912420432083, "leak": 9.137156212091213e-05, "replacement": 1.8451346158981323}
|
| 12 |
+
{"step": 12, "steps": 250, "lr": 3.7500000000000003e-05, "grad_norm": 1.3809114694595337, "loss_scale": 4.0, "retries": 0, "seconds": 12.0, "max_memory_gb": 24.64, "pair": 1.0844092965126038, "sign": 0.07545977435074747, "keep": 0.9349518977105618, "leak": 0.0001359367652327137, "replacement": 1.5896491706371307}
|
| 13 |
+
{"step": 13, "steps": 250, "lr": 4.0625000000000005e-05, "grad_norm": 1.308713674545288, "loss_scale": 4.0, "retries": 0, "seconds": 13.0, "max_memory_gb": 24.64, "pair": 1.0782186090946198, "sign": 0.057775964960455894, "keep": 0.8533943975780858, "leak": 0.0004578279385896167, "replacement": 1.5339182913303375}
|
| 14 |
+
{"step": 14, "steps": 250, "lr": 4.375e-05, "grad_norm": 1.943528413772583, "loss_scale": 4.0, "retries": 0, "seconds": 13.9, "max_memory_gb": 24.64, "pair": 1.0801407396793365, "sign": 0.26642381865531206, "keep": 1.6105240881443024, "leak": 0.0007258682817337103, "replacement": 2.0187961757183075}
|
| 15 |
+
{"step": 15, "steps": 250, "lr": 4.6875e-05, "grad_norm": 1.9395681619644165, "loss_scale": 4.0, "retries": 0, "seconds": 14.9, "max_memory_gb": 24.64, "pair": 1.069372683763504, "sign": 0.19765539490617812, "keep": 0.993158270164713, "leak": 0.002023095468757674, "replacement": 1.6652851998806}
|
| 16 |
+
{"step": 16, "steps": 250, "lr": 5e-05, "grad_norm": 1.1997666358947754, "loss_scale": 4.0, "retries": 0, "seconds": 15.9, "max_memory_gb": 24.64, "pair": 1.0717414915561676, "sign": 0.3019342829938978, "keep": 0.3747153547301423, "leak": 0.0006249765574466437, "replacement": 1.4102225601673126}
|
| 17 |
+
{"step": 17, "steps": 250, "lr": 5e-05, "grad_norm": 2.976576566696167, "loss_scale": 4.0, "retries": 0, "seconds": 16.9, "max_memory_gb": 24.64, "pair": 1.0833787322044373, "sign": 0.7484222706407309, "keep": 0.831612793088425, "leak": 0.0014533174035022967, "replacement": 1.8737595081329346}
|
| 18 |
+
{"step": 18, "steps": 250, "lr": 4.999797225177623e-05, "grad_norm": 1.921366572380066, "loss_scale": 4.0, "retries": 0, "seconds": 17.9, "max_memory_gb": 24.66, "pair": 1.0613307058811188, "sign": 0.18896084697917104, "keep": 1.056324827921344, "leak": 0.0017677610530881793, "replacement": 1.6844155490398407}
|
| 19 |
+
{"step": 19, "steps": 250, "lr": 4.999188937259492e-05, "grad_norm": 3.1113827228546143, "loss_scale": 4.0, "retries": 0, "seconds": 18.9, "max_memory_gb": 24.66, "pair": 1.0263586938381195, "sign": 0.03659565723501146, "keep": 1.3573853373527527, "leak": 0.01062368307611905, "replacement": 1.7260051369667053}
|
| 20 |
+
{"step": 20, "steps": 250, "lr": 4.998175245886032e-05, "grad_norm": 1.1257145404815674, "loss_scale": 4.0, "retries": 0, "seconds": 19.9, "max_memory_gb": 24.66, "pair": 1.0584298968315125, "sign": 0.03150372998788953, "keep": 0.003740171483514132, "leak": 0.0019348568766872631, "replacement": 1.0765355229377747}
|
| 21 |
+
{"step": 21, "steps": 250, "lr": 4.996756333769319e-05, "grad_norm": 2.07155442237854, "loss_scale": 4.0, "retries": 0, "seconds": 20.9, "max_memory_gb": 24.66, "pair": 1.0229685604572296, "sign": 0.04505528323352337, "keep": 0.5889426099347475, "leak": 0.008522697258740664, "replacement": 1.3420981466770172}
|
| 22 |
+
{"step": 22, "steps": 250, "lr": 4.994932456660157e-05, "grad_norm": 2.077216625213623, "loss_scale": 4.0, "retries": 0, "seconds": 21.9, "max_memory_gb": 24.66, "pair": 1.0253704190254211, "sign": 0.0668171988800168, "keep": 0.7205783311801497, "leak": 0.004258246503013652, "replacement": 1.420132726430893}
|
| 23 |
+
{"step": 23, "steps": 250, "lr": 4.992703943301973e-05, "grad_norm": 2.3394546508789062, "loss_scale": 4.0, "retries": 0, "seconds": 22.9, "max_memory_gb": 24.66, "pair": 1.0084541141986847, "sign": 0.0357923770789057, "keep": 1.2176667890962563, "leak": 0.004282996931578964, "replacement": 1.6362545192241669}
|
| 24 |
+
{"step": 24, "steps": 250, "lr": 4.9900711953715675e-05, "grad_norm": 2.5804288387298584, "loss_scale": 4.0, "retries": 0, "seconds": 23.8, "max_memory_gb": 24.66, "pair": 0.9505388140678406, "sign": 0.016616527223959565, "keep": 0.5954832471907139, "leak": 0.012381731648929417, "replacement": 1.2596841752529144}
|
| 25 |
+
{"step": 25, "steps": 250, "lr": 4.987034687406713e-05, "grad_norm": 2.387911796569824, "loss_scale": 4.0, "retries": 0, "seconds": 24.8, "max_memory_gb": 24.66, "pair": 0.9822705686092377, "sign": 0.11136306449770927, "keep": 1.347866703494219, "leak": 0.013590652961283922, "replacement": 1.7152830958366394}
|
| 26 |
+
{"step": 26, "steps": 250, "lr": 4.9835949667206215e-05, "grad_norm": 2.6690642833709717, "loss_scale": 4.0, "retries": 0, "seconds": 25.8, "max_memory_gb": 24.66, "pair": 0.9420401453971863, "sign": 0.06373409216757864, "keep": 0.6415662467479706, "leak": 0.02197742486896459, "replacement": 1.3001846671104431}
|
| 27 |
+
{"step": 27, "steps": 250, "lr": 4.9797526533032945e-05, "grad_norm": 2.5473341941833496, "loss_scale": 4.0, "retries": 0, "seconds": 26.8, "max_memory_gb": 24.66, "pair": 0.9418675452470779, "sign": 0.14459697343409061, "keep": 0.791927744721761, "leak": 0.022170746000483632, "replacement": 1.4156726151704788}
|
| 28 |
+
{"step": 28, "steps": 250, "lr": 4.9755084397097716e-05, "grad_norm": 1.9068999290466309, "loss_scale": 4.0, "retries": 0, "seconds": 27.8, "max_memory_gb": 24.66, "pair": 0.9878570735454559, "sign": 0.42101318994536996, "keep": 1.409144975244999, "leak": 0.027624098816886544, "replacement": 1.9098421037197113}
|
| 29 |
+
{"step": 29, "steps": 250, "lr": 4.970863090935304e-05, "grad_norm": 2.3723690509796143, "loss_scale": 4.0, "retries": 0, "seconds": 28.8, "max_memory_gb": 24.66, "pair": 1.0086399167776108, "sign": 0.735750287771225, "keep": 0.5285235373303294, "leak": 0.008001487716683187, "replacement": 1.6427772343158722}
|
| 30 |
+
{"step": 30, "steps": 250, "lr": 4.965817444277468e-05, "grad_norm": 2.865506887435913, "loss_scale": 4.0, "retries": 0, "seconds": 29.8, "max_memory_gb": 24.66, "pair": 0.8498270213603973, "sign": 0.011454866966232657, "keep": 0.6413502853247337, "leak": 0.04144103452563286, "replacement": 1.1865897923707962}
|
| 31 |
+
{"step": 31, "steps": 250, "lr": 4.9603724091852456e-05, "grad_norm": 3.16275954246521, "loss_scale": 4.0, "retries": 0, "seconds": 30.8, "max_memory_gb": 24.66, "pair": 0.9178417921066284, "sign": 0.21522133611142635, "keep": 1.3149241246283054, "leak": 0.040764375356957316, "replacement": 1.6931056380271912}
|
| 32 |
+
{"step": 32, "steps": 250, "lr": 4.9545289670951025e-05, "grad_norm": 3.521577835083008, "loss_scale": 4.0, "retries": 0, "seconds": 31.8, "max_memory_gb": 24.66, "pair": 0.8373127579689026, "sign": 0.11897367518395185, "keep": 0.920768640935421, "leak": 0.10850454494357109, "replacement": 1.3843100368976593}
|
| 33 |
+
{"step": 33, "steps": 250, "lr": 4.948288171254089e-05, "grad_norm": 1.975196361541748, "loss_scale": 4.0, "retries": 0, "seconds": 32.8, "max_memory_gb": 24.66, "pair": 0.9441412687301636, "sign": 0.1999318622983992, "keep": 0.5188692957162857, "leak": 0.03205039264867082, "replacement": 1.311554491519928}
|
| 34 |
+
{"step": 34, "steps": 250, "lr": 4.941651146529998e-05, "grad_norm": 2.288963794708252, "loss_scale": 4.0, "retries": 0, "seconds": 33.8, "max_memory_gb": 24.66, "pair": 0.858988881111145, "sign": 0.0324381769169122, "keep": 0.30815399949642597, "leak": 0.09200108703225851, "replacement": 1.0522852540016174}
|
| 35 |
+
{"step": 35, "steps": 250, "lr": 4.9346190892086174e-05, "grad_norm": 2.4940593242645264, "loss_scale": 4.0, "retries": 0, "seconds": 34.8, "max_memory_gb": 24.66, "pair": 0.8584348857402802, "sign": 0.06355391722172499, "keep": 0.4384362563723698, "leak": 0.05551717570051551, "replacement": 1.1233092546463013}
|
| 36 |
+
{"step": 36, "steps": 250, "lr": 4.927193266778101e-05, "grad_norm": 2.2380595207214355, "loss_scale": 4.0, "retries": 0, "seconds": 35.8, "max_memory_gb": 24.66, "pair": 0.7981594502925873, "sign": 0.0437262614723295, "keep": 0.4316729459242197, "leak": 0.14465833315625787, "replacement": 1.0720236748456955}
|
| 37 |
+
{"step": 37, "steps": 250, "lr": 4.919375017700515e-05, "grad_norm": 2.0182392597198486, "loss_scale": 4.0, "retries": 0, "seconds": 36.7, "max_memory_gb": 24.66, "pair": 0.8672910183668137, "sign": 0.21615763986483216, "keep": 0.8887733938754536, "leak": 0.10831619708187645, "replacement": 1.4468356370925903}
|
| 38 |
+
{"step": 38, "steps": 250, "lr": 4.911165751170588e-05, "grad_norm": 2.4127581119537354, "loss_scale": 4.0, "retries": 0, "seconds": 37.7, "max_memory_gb": 24.66, "pair": 0.7364848554134369, "sign": 0.09165746159851551, "keep": 0.26575225219130516, "leak": 0.19899964146316051, "replacement": 0.9649396389722824}
|
| 39 |
+
{"step": 39, "steps": 250, "lr": 4.902566946861708e-05, "grad_norm": 1.644340991973877, "loss_scale": 4.0, "retries": 0, "seconds": 38.7, "max_memory_gb": 24.66, "pair": 0.9048540443181992, "sign": 1.3901672773063183, "keep": 0.3717589436346316, "leak": 0.08477264270186424, "replacement": 1.8070103526115417}
|
| 40 |
+
{"step": 40, "steps": 250, "lr": 4.89358015465922e-05, "grad_norm": 2.481245756149292, "loss_scale": 4.0, "retries": 0, "seconds": 39.7, "max_memory_gb": 24.66, "pair": 0.7080312073230743, "sign": 0.0190753978677094, "keep": 0.2544433439907152, "leak": 0.1743155950680375, "replacement": 0.8883694708347321}
|
| 41 |
+
{"step": 41, "steps": 250, "lr": 4.884206994381078e-05, "grad_norm": 2.1692371368408203, "loss_scale": 4.0, "retries": 0, "seconds": 40.7, "max_memory_gb": 24.66, "pair": 0.9046269804239273, "sign": 0.8865694310516119, "keep": 1.0307181058451533, "leak": 0.044736487325280905, "replacement": 1.874454841017723}
|
| 42 |
+
{"step": 42, "steps": 250, "lr": 4.874449155485866e-05, "grad_norm": 1.7763490676879883, "loss_scale": 4.0, "retries": 0, "seconds": 41.6, "max_memory_gb": 24.66, "pair": 0.6533639580011368, "sign": 0.05498817120678723, "keep": 0.0974711838789517, "leak": 0.24752528965473175, "replacement": 0.7914749383926392}
|
| 43 |
+
{"step": 43, "steps": 250, "lr": 4.864308396768294e-05, "grad_norm": 2.024191379547119, "loss_scale": 4.0, "retries": 0, "seconds": 42.6, "max_memory_gb": 24.66, "pair": 0.6400236021727324, "sign": 0.15295484708622098, "keep": 0.31538343837019056, "leak": 0.698502330109477, "replacement": 1.0488183051347733}
|
| 44 |
+
{"step": 44, "steps": 250, "lr": 4.853786546042184e-05, "grad_norm": 2.752363681793213, "loss_scale": 4.0, "retries": 0, "seconds": 43.6, "max_memory_gb": 24.66, "pair": 0.6713006719946861, "sign": 0.17985882493667305, "keep": 0.9986480736843077, "leak": 0.2586264666169882, "replacement": 1.3252107203006744}
|
| 45 |
+
{"step": 45, "steps": 250, "lr": 4.842885499811012e-05, "grad_norm": 2.703026056289673, "loss_scale": 4.0, "retries": 0, "seconds": 44.6, "max_memory_gb": 24.66, "pair": 0.9554942101240158, "sign": 0.7460218775086105, "keep": 1.9431093638995662, "leak": 0.06456891939160414, "replacement": 2.3162019848823547}
|
| 46 |
+
{"step": 46, "steps": 250, "lr": 4.831607222926078e-05, "grad_norm": 2.1827712059020996, "loss_scale": 4.0, "retries": 0, "seconds": 45.6, "max_memory_gb": 24.66, "pair": 0.6985536515712738, "sign": 0.2693364628357813, "keep": 0.5921103209257126, "leak": 0.15887169539928436, "replacement": 1.168994978070259}
|
| 47 |
+
{"step": 47, "steps": 250, "lr": 4.8199537482323545e-05, "grad_norm": 1.7091879844665527, "loss_scale": 4.0, "retries": 0, "seconds": 46.7, "max_memory_gb": 24.66, "pair": 0.6582672856748104, "sign": 0.10962767276214436, "keep": 0.10021453895024024, "leak": 0.2652213447727263, "replacement": 0.8294937014579773}
|
| 48 |
+
{"step": 48, "steps": 250, "lr": 4.807927176202073e-05, "grad_norm": 1.8036415576934814, "loss_scale": 4.0, "retries": 0, "seconds": 47.7, "max_memory_gb": 24.66, "pair": 0.45369047671556473, "sign": 0.02889237756608054, "keep": 0.32379862444940954, "leak": 0.6704546511173248, "replacement": 0.7976496517658234}
|
| 49 |
+
{"step": 49, "steps": 250, "lr": 4.795529674556129e-05, "grad_norm": 1.8968255519866943, "loss_scale": 4.0, "retries": 0, "seconds": 48.8, "max_memory_gb": 24.66, "pair": 0.4546547085046768, "sign": 0.04574205307289958, "keep": 0.1768071377009619, "leak": 0.38017382472753525, "replacement": 0.6609727591276169}
|
| 50 |
+
{"step": 50, "steps": 250, "lr": 4.782763477873361e-05, "grad_norm": 1.9136351346969604, "loss_scale": 4.0, "retries": 0, "seconds": 49.8, "max_memory_gb": 24.66, "pair": 0.5578595697879791, "sign": 0.10361083783209324, "keep": 0.43794060830623494, "leak": 0.5914833173155785, "replacement": 0.9765061289072037}
|
| 51 |
+
{"step": 51, "steps": 250, "lr": 4.769630887187782e-05, "grad_norm": 2.5538735389709473, "loss_scale": 4.0, "retries": 0, "seconds": 50.9, "max_memory_gb": 24.66, "pair": 0.5149668827652931, "sign": 0.03198935999535024, "keep": 1.1336502969243156, "leak": 0.2812798556078633, "replacement": 1.1681066900491714}
|
| 52 |
+
{"step": 52, "steps": 250, "lr": 4.756134269573832e-05, "grad_norm": 1.756796956062317, "loss_scale": 4.0, "retries": 0, "seconds": 51.9, "max_memory_gb": 24.66, "pair": 0.5423272550106049, "sign": 0.26681355526670814, "keep": 0.4081090726540424, "leak": 0.42363209649920464, "replacement": 0.9856965988874435}
|
| 53 |
+
{"step": 53, "steps": 250, "lr": 4.7422760577197226e-05, "grad_norm": 3.481926441192627, "loss_scale": 4.0, "retries": 0, "seconds": 52.8, "max_memory_gb": 24.66, "pair": 0.5006474480032921, "sign": 0.0958057168754749, "keep": 0.4421991645358503, "leak": 0.6694219466298819, "replacement": 0.9370054006576538}
|
| 54 |
+
{"step": 54, "steps": 250, "lr": 4.728058749488963e-05, "grad_norm": 2.298646926879883, "loss_scale": 4.0, "retries": 0, "seconds": 53.8, "max_memory_gb": 24.66, "pair": 0.5048634707927704, "sign": 0.16976024548057467, "keep": 0.3467544511659071, "leak": 0.7618602174334228, "replacement": 0.9535858482122421}
|
| 55 |
+
{"step": 55, "steps": 250, "lr": 4.713484907470133e-05, "grad_norm": 1.108834147453308, "loss_scale": 4.0, "retries": 0, "seconds": 54.8, "max_memory_gb": 24.66, "pair": 0.41196709126234055, "sign": 0.03490909386891872, "keep": 0.0133626511451439, "leak": 0.5329079367220402, "replacement": 0.5693299397826195}
|
| 56 |
+
{"step": 56, "steps": 250, "lr": 4.6985571585149876e-05, "grad_norm": 0.8909606337547302, "loss_scale": 4.0, "retries": 0, "seconds": 55.7, "max_memory_gb": 24.66, "pair": 0.3465124163776636, "sign": 0.01532230875454843, "keep": 0.103188322391361, "leak": 0.7490449994802475, "replacement": 0.5930289775133133}
|
| 57 |
+
{"step": 57, "steps": 250, "lr": 4.6832781932649884e-05, "grad_norm": 1.3335822820663452, "loss_scale": 4.0, "retries": 0, "seconds": 56.7, "max_memory_gb": 24.66, "pair": 0.4817569926381111, "sign": 0.39958301186561584, "keep": 0.02182653116506117, "leak": 0.8005374024888994, "replacement": 0.8925961330533028}
|
| 58 |
+
{"step": 58, "steps": 250, "lr": 4.667650765666325e-05, "grad_norm": 1.002448320388794, "loss_scale": 4.0, "retries": 0, "seconds": 57.7, "max_memory_gb": 24.66, "pair": 0.24914909340441227, "sign": 0.01223599398508668, "keep": 0.07959034198211157, "leak": 1.0670633763074875, "replacement": 0.5618281066417694}
|
| 59 |
+
{"step": 59, "steps": 250, "lr": 4.651677692473538e-05, "grad_norm": 1.2724955081939697, "loss_scale": 4.0, "retries": 0, "seconds": 58.7, "max_memory_gb": 24.66, "pair": 0.42641347646713257, "sign": 0.03751878038747236, "keep": 0.03095673782081576, "leak": 0.48933885409496725, "replacement": 0.5829859524965286}
|
| 60 |
+
{"step": 60, "steps": 250, "lr": 4.635361852741808e-05, "grad_norm": 2.2295453548431396, "loss_scale": 4.0, "retries": 0, "seconds": 59.7, "max_memory_gb": 24.66, "pair": 0.4791686711832881, "sign": 0.0631140394252725, "keep": 0.10639960586559027, "leak": 1.1441137964029622, "replacement": 0.8499539345502853}
|
| 61 |
+
{"step": 61, "steps": 250, "lr": 4.6187061873080264e-05, "grad_norm": 3.6485376358032227, "loss_scale": 4.0, "retries": 0, "seconds": 60.7, "max_memory_gb": 24.66, "pair": 0.437805213034153, "sign": 0.2419138951227069, "keep": 0.08167383540421724, "leak": 1.8879885794594884, "replacement": 1.0715962201356888}
|
| 62 |
+
{"step": 62, "steps": 250, "lr": 4.601713698260727e-05, "grad_norm": 1.4375462532043457, "loss_scale": 4.0, "retries": 0, "seconds": 61.6, "max_memory_gb": 24.66, "pair": 0.4585406444966793, "sign": 0.03353908518329263, "keep": 0.07708484054455766, "leak": 1.0326422676444054, "replacement": 0.7720131576061249}
|
| 63 |
+
{"step": 63, "steps": 250, "lr": 4.5843874483989746e-05, "grad_norm": 1.1096991300582886, "loss_scale": 4.0, "retries": 0, "seconds": 62.6, "max_memory_gb": 24.66, "pair": 0.3737536892294884, "sign": 0.1423791740089655, "keep": 0.1175845805555582, "leak": 0.23508719308301806, "replacement": 0.5625073462724686}
|
| 64 |
+
{"step": 64, "steps": 250, "lr": 4.566730560680316e-05, "grad_norm": 0.6844059824943542, "loss_scale": 4.0, "retries": 0, "seconds": 63.6, "max_memory_gb": 24.66, "pair": 0.2666590139269829, "sign": 0.01634377404116094, "keep": 0.03513277415186167, "leak": 0.7809305414557457, "replacement": 0.4876299351453781}
|
| 65 |
+
{"step": 65, "steps": 250, "lr": 4.548746217657878e-05, "grad_norm": 3.0959537029266357, "loss_scale": 4.0, "retries": 0, "seconds": 64.6, "max_memory_gb": 24.66, "pair": 0.40972965210676193, "sign": 0.3660194796975702, "keep": 0.3073175959289074, "leak": 2.0613107439130545, "replacement": 1.261725902557373}
|
| 66 |
+
{"step": 66, "steps": 250, "lr": 4.5304376609067396e-05, "grad_norm": 1.402345061302185, "loss_scale": 4.0, "retries": 0, "seconds": 65.6, "max_memory_gb": 24.66, "pair": 0.4641815945506096, "sign": 0.13505766831804067, "keep": 0.3417984559782212, "leak": 0.28400288335978985, "replacement": 0.7736103758215904}
|
| 67 |
+
{"step": 67, "steps": 250, "lr": 4.5118081904396504e-05, "grad_norm": 6.1691670417785645, "loss_scale": 4.0, "retries": 0, "seconds": 66.6, "max_memory_gb": 24.66, "pair": 0.21390890888869762, "sign": 0.020173318102024496, "keep": 0.1177414103076444, "leak": 1.1401393413543701, "replacement": 0.5679011195898056}
|
| 68 |
+
{"step": 68, "steps": 250, "lr": 4.4928611641122225e-05, "grad_norm": 1.4553927183151245, "loss_scale": 4.0, "retries": 0, "seconds": 67.6, "max_memory_gb": 24.66, "pair": 0.5638898760080338, "sign": 0.07239087345078588, "keep": 0.01944893612380838, "leak": 0.4855845794081688, "replacement": 0.731205940246582}
|
| 69 |
+
{"step": 69, "steps": 250, "lr": 4.473599997017701e-05, "grad_norm": 2.3669536113739014, "loss_scale": 4.0, "retries": 0, "seconds": 68.6, "max_memory_gb": 24.66, "pair": 0.5674436669796705, "sign": 0.8514063064940274, "keep": 0.5781269040890038, "leak": 1.006084989523515, "replacement": 1.5337315499782562}
|
| 70 |
+
{"step": 70, "steps": 250, "lr": 4.4540281608714053e-05, "grad_norm": 2.122206687927246, "loss_scale": 4.0, "retries": 0, "seconds": 69.6, "max_memory_gb": 24.66, "pair": 0.4220619569532573, "sign": 0.036985050624934956, "keep": 0.03786211875558365, "leak": 1.8019311428070068, "replacement": 0.9099683314561844}
|
| 71 |
+
{"step": 71, "steps": 250, "lr": 4.434149183384977e-05, "grad_norm": 1.2520971298217773, "loss_scale": 4.0, "retries": 0, "seconds": 70.5, "max_memory_gb": 24.66, "pair": 0.18568865582346916, "sign": 0.007623986341059208, "keep": 0.16326946272647547, "leak": 0.8296209499239922, "replacement": 0.4785406142473221}
|
| 72 |
+
{"step": 72, "steps": 250, "lr": 4.413966647630534e-05, "grad_norm": 1.9414350986480713, "loss_scale": 4.0, "retries": 0, "seconds": 71.5, "max_memory_gb": 24.66, "pair": 0.28387574199587107, "sign": 0.19949716818518937, "keep": 0.10217262338846922, "leak": 1.4600576683878899, "replacement": 0.7997250258922577}
|
| 73 |
+
{"step": 73, "steps": 250, "lr": 4.3934841913948323e-05, "grad_norm": 1.6074433326721191, "loss_scale": 4.0, "retries": 0, "seconds": 72.5, "max_memory_gb": 24.66, "pair": 0.5129876472055912, "sign": 0.13282848987728357, "keep": 0.5463389169781294, "leak": 0.3680935800075531, "replacement": 0.9445948228240013}
|
| 74 |
+
{"step": 74, "steps": 250, "lr": 4.372705506523583e-05, "grad_norm": 1.9558775424957275, "loss_scale": 4.0, "retries": 0, "seconds": 73.5, "max_memory_gb": 24.66, "pair": 0.3874411843717098, "sign": 0.25732381641864777, "keep": 0.220347783520765, "leak": 1.4530226937495172, "replacement": 0.9895326644182205}
|
| 75 |
+
{"step": 75, "steps": 250, "lr": 4.351634338256017e-05, "grad_norm": 1.7386842966079712, "loss_scale": 4.0, "retries": 0, "seconds": 74.5, "max_memory_gb": 24.66, "pair": 0.7576713971793652, "sign": 0.524545130552724, "keep": 0.8487801133305766, "leak": 0.24184099296689965, "replacement": 1.504794254899025}
|
| 76 |
+
{"step": 76, "steps": 250, "lr": 4.330274484549825e-05, "grad_norm": 1.0755178928375244, "loss_scale": 4.0, "retries": 0, "seconds": 75.5, "max_memory_gb": 24.66, "pair": 0.34665507078170776, "sign": 0.10477638116572052, "keep": 0.025347669317852706, "leak": 0.7690584026277065, "replacement": 0.6039817184209824}
|
| 77 |
+
{"step": 77, "steps": 250, "lr": 4.308629795396599e-05, "grad_norm": 2.1013667583465576, "loss_scale": 4.0, "retries": 0, "seconds": 76.4, "max_memory_gb": 24.66, "pair": 0.21090046875178814, "sign": 0.00877962401136756, "keep": 0.008565520895899681, "leak": 1.6523834317922592, "replacement": 0.6326688826084137}
|
| 78 |
+
{"step": 78, "steps": 250, "lr": 4.286704172127894e-05, "grad_norm": 2.1413016319274902, "loss_scale": 4.0, "retries": 0, "seconds": 77.4, "max_memory_gb": 24.66, "pair": 0.4950392171740532, "sign": 0.7009562803432345, "keep": 0.0021382997510954738, "leak": 0.4661720935255289, "replacement": 0.9631294906139374}
|
| 79 |
+
{"step": 79, "steps": 250, "lr": 4.264501566712033e-05, "grad_norm": 1.385023593902588, "loss_scale": 4.0, "retries": 0, "seconds": 78.4, "max_memory_gb": 24.66, "pair": 0.5097723230719566, "sign": 0.03510914300568402, "keep": 0.1283182966021741, "leak": 0.45784442126750946, "replacement": 0.7059471607208252}
|
| 80 |
+
{"step": 80, "steps": 250, "lr": 4.2420259810417894e-05, "grad_norm": 1.0412802696228027, "loss_scale": 4.0, "retries": 0, "seconds": 79.3, "max_memory_gb": 24.66, "pair": 0.5504274368286133, "sign": 0.10364799248054624, "keep": 0.047161212577293554, "leak": 0.6270403228700161, "replacement": 0.7825921177864075}
|
| 81 |
+
{"step": 81, "steps": 250, "lr": 4.219281466213066e-05, "grad_norm": 0.9949823021888733, "loss_scale": 4.0, "retries": 0, "seconds": 80.3, "max_memory_gb": 24.66, "pair": 0.31056103855371475, "sign": 0.06323168851668015, "keep": 0.18131924141198397, "leak": 1.0818393509835005, "replacement": 0.703296348452568}
|
| 82 |
+
{"step": 82, "steps": 250, "lr": 4.1962721217947135e-05, "grad_norm": 2.794419050216675, "loss_scale": 4.0, "retries": 0, "seconds": 81.3, "max_memory_gb": 24.66, "pair": 0.2753683142364025, "sign": 1.0118075896752998, "keep": 0.09059798822272569, "leak": 1.768195554614067, "replacement": 1.268620029091835}
|
| 83 |
+
{"step": 83, "steps": 250, "lr": 4.173002095089599e-05, "grad_norm": 0.9002965688705444, "loss_scale": 4.0, "retries": 0, "seconds": 82.3, "max_memory_gb": 24.66, "pair": 0.33017314970493317, "sign": 0.03234747762326151, "keep": 0.3640815284452401, "leak": 0.8088610209524632, "replacement": 0.7306028828024864}
|
| 84 |
+
{"step": 84, "steps": 250, "lr": 4.149475580387084e-05, "grad_norm": 1.1176789999008179, "loss_scale": 4.0, "retries": 0, "seconds": 83.3, "max_memory_gb": 24.66, "pair": 0.2998217474669218, "sign": 0.030811138567514718, "keep": 0.048419292419453086, "leak": 0.9583147317171097, "replacement": 0.5790156498551369}
|
| 85 |
+
{"step": 85, "steps": 250, "lr": 4.1256968182070275e-05, "grad_norm": 4.599983215332031, "loss_scale": 4.0, "retries": 0, "seconds": 84.3, "max_memory_gb": 24.66, "pair": 0.3382658585906029, "sign": 0.06270264345221221, "keep": 0.0768634229204963, "leak": 0.7978003211319447, "replacement": 0.6074989587068558}
|
| 86 |
+
{"step": 86, "steps": 250, "lr": 4.101670094535453e-05, "grad_norm": 2.0579488277435303, "loss_scale": 4.0, "retries": 0, "seconds": 85.2, "max_memory_gb": 24.66, "pair": 0.3559749610722065, "sign": 0.3135653231292963, "keep": 0.39646017155610025, "leak": 0.2673886101692915, "replacement": 0.7778348848223686}
|
| 87 |
+
{"step": 87, "steps": 250, "lr": 4.077399740052027e-05, "grad_norm": 1.648735761642456, "loss_scale": 4.0, "retries": 0, "seconds": 86.2, "max_memory_gb": 24.66, "pair": 0.4878396540880203, "sign": 0.09880043694283813, "keep": 0.015030964277684689, "leak": 0.4194616936329112, "replacement": 0.649620771408081}
|
| 88 |
+
{"step": 88, "steps": 250, "lr": 4.0528901293494775e-05, "grad_norm": 3.226145029067993, "loss_scale": 4.0, "retries": 0, "seconds": 87.2, "max_memory_gb": 24.66, "pair": 0.4518614485859871, "sign": 0.2189642246812582, "keep": 0.028454314916530166, "leak": 0.5987624377012253, "replacement": 0.7252613455057144}
|
| 89 |
+
{"step": 89, "steps": 250, "lr": 4.0281456801451004e-05, "grad_norm": 1.6483436822891235, "loss_scale": 4.0, "retries": 0, "seconds": 88.1, "max_memory_gb": 24.66, "pair": 0.5994590073823929, "sign": 0.1955470060929656, "keep": 0.018005205628810472, "leak": 0.11220102908555418, "replacement": 0.7342853993177414}
|
| 90 |
+
{"step": 90, "steps": 250, "lr": 4.0031708524844884e-05, "grad_norm": 1.990323781967163, "loss_scale": 4.0, "retries": 0, "seconds": 89.1, "max_memory_gb": 24.66, "pair": 0.5011741667985916, "sign": 0.133425181847997, "keep": 0.06518608616897836, "leak": 1.0415531545877457, "replacement": 0.8608680963516235}
|
| 91 |
+
{"step": 91, "steps": 250, "lr": 3.977970147937635e-05, "grad_norm": 2.691159963607788, "loss_scale": 4.0, "retries": 0, "seconds": 90.1, "max_memory_gb": 24.66, "pair": 0.5535736158490181, "sign": 0.22438134881667793, "keep": 0.19863681607239414, "leak": 0.2605085727991536, "replacement": 0.8302098512649536}
|
| 92 |
+
{"step": 92, "steps": 250, "lr": 3.952548108787553e-05, "grad_norm": 3.736649513244629, "loss_scale": 4.0, "retries": 0, "seconds": 91.1, "max_memory_gb": 24.66, "pair": 0.5214594006538391, "sign": 0.7962845609290525, "keep": 1.3222555919705883, "leak": 0.6176423830911517, "replacement": 1.7351400628685951}
|
| 93 |
+
{"step": 93, "steps": 250, "lr": 3.9269093172115496e-05, "grad_norm": 1.5101782083511353, "loss_scale": 4.0, "retries": 0, "seconds": 92.1, "max_memory_gb": 24.66, "pair": 0.38712257519364357, "sign": 0.04169383563566953, "keep": 0.17010751937050372, "leak": 0.4959554527304135, "replacement": 0.6170121282339096}
|
| 94 |
+
{"step": 94, "steps": 250, "lr": 3.901058394455323e-05, "grad_norm": 2.7410333156585693, "loss_scale": 4.0, "retries": 0, "seconds": 93.1, "max_memory_gb": 24.66, "pair": 0.3056482672691345, "sign": 0.03770825994433835, "keep": 0.811612605350092, "leak": 1.1731159687042236, "replacement": 1.0235877260565758}
|
| 95 |
+
{"step": 95, "steps": 250, "lr": 3.875e-05, "grad_norm": 2.2016639709472656, "loss_scale": 4.0, "retries": 0, "seconds": 94.0, "max_memory_gb": 24.66, "pair": 0.4971481002867222, "sign": 0.313059814972803, "keep": 1.5178969078697264, "leak": 0.6209772012662143, "replacement": 1.5678707957267761}
|
| 96 |
+
{"step": 96, "steps": 250, "lr": 3.8487388307223005e-05, "grad_norm": 1.5700916051864624, "loss_scale": 4.0, "retries": 0, "seconds": 95.0, "max_memory_gb": 24.66, "pair": 0.41828399524092674, "sign": 0.12330188206396997, "keep": 1.0207501631230116, "leak": 0.5149951588828117, "replacement": 1.1190588399767876}
|
| 97 |
+
{"step": 97, "steps": 250, "lr": 3.822279620047943e-05, "grad_norm": 1.4360700845718384, "loss_scale": 4.0, "retries": 0, "seconds": 96.0, "max_memory_gb": 24.66, "pair": 0.3381325751543045, "sign": 0.01881279144436121, "keep": 0.06589419210649794, "leak": 0.8009067960083485, "replacement": 0.5807127729058266}
|
| 98 |
+
{"step": 98, "steps": 250, "lr": 3.795627137098479e-05, "grad_norm": 1.1904908418655396, "loss_scale": 4.0, "retries": 0, "seconds": 97.0, "max_memory_gb": 24.66, "pair": 0.4598841778934002, "sign": 0.17253416345920414, "keep": 0.02691756257399902, "leak": 0.7387669384479523, "replacement": 0.7443018034100533}
|
| 99 |
+
{"step": 99, "steps": 250, "lr": 3.7687861858316844e-05, "grad_norm": 2.367474317550659, "loss_scale": 4.0, "retries": 0, "seconds": 98.0, "max_memory_gb": 24.66, "pair": 0.6849384009838104, "sign": 0.7379139505792409, "keep": 0.039800777860818926, "leak": 0.21708634495735168, "replacement": 1.128067396581173}
|
| 100 |
+
{"step": 100, "steps": 250, "lr": 3.741761604175668e-05, "grad_norm": 3.215153217315674, "loss_scale": 4.0, "retries": 0, "seconds": 99.0, "max_memory_gb": 24.66, "pair": 0.5360939688980579, "sign": 0.9248852665768936, "keep": 0.5062148390570655, "leak": 0.8768063690513372, "replacement": 1.4708456173539162}
|
| 101 |
+
{"step": 101, "steps": 250, "lr": 3.714558263156872e-05, "grad_norm": 4.146965980529785, "loss_scale": 4.0, "retries": 0, "seconds": 100.0, "max_memory_gb": 24.66, "pair": 0.18346290895715356, "sign": 0.010901302623096853, "keep": 0.04563795323338127, "leak": 3.3531845584511757, "replacement": 1.0500286892056465}
|
| 102 |
+
{"step": 102, "steps": 250, "lr": 3.6871810660220904e-05, "grad_norm": 1.3754444122314453, "loss_scale": 4.0, "retries": 0, "seconds": 100.9, "max_memory_gb": 24.66, "pair": 0.2928492836654186, "sign": 0.0260408038739115, "keep": 0.07221900322474539, "leak": 0.34172313765157014, "replacement": 0.42740996554493904}
|
| 103 |
+
{"step": 103, "steps": 250, "lr": 3.659634947354686e-05, "grad_norm": 0.8876057863235474, "loss_scale": 4.0, "retries": 0, "seconds": 101.9, "max_memory_gb": 24.66, "pair": 0.3244105502963066, "sign": 0.03569386212620884, "keep": 0.02410866920126864, "leak": 0.6111603602766991, "replacement": 0.5071019157767296}
|
| 104 |
+
{"step": 104, "steps": 250, "lr": 3.631924872185169e-05, "grad_norm": 2.6864821910858154, "loss_scale": 4.0, "retries": 0, "seconds": 102.8, "max_memory_gb": 24.66, "pair": 0.5094882622361183, "sign": 0.4579635856789537, "keep": 0.004726395145553397, "leak": 0.5714577063918114, "replacement": 0.8836976885795593}
|
| 105 |
+
{"step": 105, "steps": 250, "lr": 3.6040558350962674e-05, "grad_norm": 1.3168597221374512, "loss_scale": 4.0, "retries": 0, "seconds": 103.8, "max_memory_gb": 24.66, "pair": 0.2621566243469715, "sign": 0.02317306841723621, "keep": 0.14227793784812093, "leak": 1.2781480550765991, "replacement": 0.6644191294908524}
|
| 106 |
+
{"step": 106, "steps": 250, "lr": 3.576032859322687e-05, "grad_norm": 1.2692667245864868, "loss_scale": 4.0, "retries": 0, "seconds": 104.8, "max_memory_gb": 24.66, "pair": 0.33026277646422386, "sign": 0.039815488969907165, "keep": 0.04928145610028878, "leak": 0.7930160553660244, "replacement": 0.573065273463726}
|
| 107 |
+
{"step": 107, "steps": 250, "lr": 3.5478609958457056e-05, "grad_norm": 2.390712022781372, "loss_scale": 4.0, "retries": 0, "seconds": 105.8, "max_memory_gb": 24.66, "pair": 0.39411743730306625, "sign": 0.36402428639121354, "keep": 2.362560058766576, "leak": 0.37404681742191315, "replacement": 1.8509212732315063}
|
| 108 |
+
{"step": 108, "steps": 250, "lr": 3.519545322482755e-05, "grad_norm": 1.5443437099456787, "loss_scale": 4.0, "retries": 0, "seconds": 106.8, "max_memory_gb": 24.66, "pair": 0.4689839705824852, "sign": 0.28420941391959786, "keep": 0.03314842368126847, "leak": 0.5026003571692854, "replacement": 0.7533130049705505}
|
| 109 |
+
{"step": 109, "steps": 250, "lr": 3.491090942972179e-05, "grad_norm": 1.5381715297698975, "loss_scale": 4.0, "retries": 0, "seconds": 107.7, "max_memory_gb": 24.66, "pair": 0.37417547032237053, "sign": 0.08339214499574155, "keep": 0.07928919931873679, "leak": 0.5927829146385193, "replacement": 0.6037118583917618}
|
| 110 |
+
{"step": 110, "steps": 250, "lr": 3.4625029860533135e-05, "grad_norm": 7.4922308921813965, "loss_scale": 4.0, "retries": 0, "seconds": 108.7, "max_memory_gb": 24.66, "pair": 0.39331139624118805, "sign": 0.24781543284188956, "keep": 0.587587489746511, "leak": 0.6681324290111661, "replacement": 0.9780459553003311}
|
| 111 |
+
{"step": 111, "steps": 250, "lr": 3.433786604542056e-05, "grad_norm": 3.350646495819092, "loss_scale": 4.0, "retries": 0, "seconds": 109.7, "max_memory_gb": 24.66, "pair": 0.42238283809274435, "sign": 1.5694062771799508, "keep": 0.0015500471621407996, "leak": 1.6441380381584167, "replacement": 1.6188954263925552}
|
| 112 |
+
{"step": 112, "steps": 250, "lr": 3.404946974402104e-05, "grad_norm": 3.925124406814575, "loss_scale": 4.0, "retries": 0, "seconds": 110.7, "max_memory_gb": 24.66, "pair": 0.21175910625606775, "sign": 0.010140179190784693, "keep": 0.002822799331624992, "leak": 2.0168115496635437, "replacement": 0.7224435061216354}
|
| 113 |
+
{"step": 113, "steps": 250, "lr": 3.375989293812019e-05, "grad_norm": 1.180708408355713, "loss_scale": 4.0, "retries": 0, "seconds": 111.7, "max_memory_gb": 24.66, "pair": 0.34355632215738297, "sign": 0.043174868798814714, "keep": 0.011776210449170321, "leak": 0.31978112272918224, "replacement": 0.4509771540760994}
|
| 114 |
+
{"step": 114, "steps": 250, "lr": 3.3469187822282766e-05, "grad_norm": 1.2147942781448364, "loss_scale": 4.0, "retries": 0, "seconds": 112.7, "max_memory_gb": 24.66, "pair": 0.26552725210785866, "sign": 0.05251597240567207, "keep": 0.03622193168848753, "leak": 1.2170369625091553, "replacement": 0.6141554415225983}
|
| 115 |
+
{"step": 115, "steps": 250, "lr": 3.317740679444499e-05, "grad_norm": 1.351759672164917, "loss_scale": 4.0, "retries": 0, "seconds": 113.6, "max_memory_gb": 24.66, "pair": 0.242323974147439, "sign": 0.011544771492481232, "keep": 0.035530563028260076, "leak": 0.7979355305433273, "replacement": 0.46534550935029984}
|
| 116 |
+
{"step": 116, "steps": 250, "lr": 3.2884602446470056e-05, "grad_norm": 3.791081428527832, "loss_scale": 4.0, "retries": 0, "seconds": 114.6, "max_memory_gb": 24.66, "pair": 0.5313710719347, "sign": 0.7918064665864222, "keep": 0.996148562396229, "leak": 0.5595980286598206, "replacement": 1.5652481019496918}
|
| 117 |
+
{"step": 117, "steps": 250, "lr": 3.259082755466876e-05, "grad_norm": 6.322779655456543, "loss_scale": 4.0, "retries": 0, "seconds": 115.6, "max_memory_gb": 24.68, "pair": 0.3275335542857647, "sign": 0.43122464464977384, "keep": 1.134960435287212, "leak": 1.8033497482538223, "replacement": 1.5614635348320007}
|
| 118 |
+
{"step": 118, "steps": 250, "lr": 3.229613507028686e-05, "grad_norm": 1.2590782642364502, "loss_scale": 4.0, "retries": 0, "seconds": 116.6, "max_memory_gb": 24.68, "pair": 0.45944229513406754, "sign": 0.33394566958304495, "keep": 0.4274619113996323, "leak": 0.6744770845398307, "replacement": 1.0087653696537018}
|
| 119 |
+
{"step": 119, "steps": 250, "lr": 3.2000578109961004e-05, "grad_norm": 1.2258919477462769, "loss_scale": 4.0, "retries": 0, "seconds": 117.6, "max_memory_gb": 24.68, "pair": 0.45102881640195847, "sign": 0.3040920081548393, "keep": 0.0182412240847043, "leak": 0.4504522867500782, "replacement": 0.7248084992170334}
|
| 120 |
+
{"step": 120, "steps": 250, "lr": 3.1704209946144616e-05, "grad_norm": 2.987356662750244, "loss_scale": 4.0, "retries": 0, "seconds": 118.6, "max_memory_gb": 24.68, "pair": 0.5235935673117638, "sign": 0.9000303689390421, "keep": 1.8368452274880838, "leak": 0.37226338870823383, "replacement": 1.9850972592830658}
|
| 121 |
+
{"step": 121, "steps": 250, "lr": 3.140708399750594e-05, "grad_norm": 1.6891783475875854, "loss_scale": 4.0, "retries": 0, "seconds": 119.6, "max_memory_gb": 24.68, "pair": 0.5993961170315742, "sign": 0.4473472014069557, "keep": 0.09137027709857648, "leak": 0.34318349370732903, "replacement": 0.9545507207512856}
|
| 122 |
+
{"step": 122, "steps": 250, "lr": 3.1109253819299606e-05, "grad_norm": 1.0295237302780151, "loss_scale": 4.0, "retries": 0, "seconds": 120.6, "max_memory_gb": 24.68, "pair": 0.4124405672773719, "sign": 0.10593914374476299, "keep": 0.039784901764505776, "leak": 0.6791188009083271, "replacement": 0.6550822854042053}
|
| 123 |
+
{"step": 123, "steps": 250, "lr": 3.081077309371361e-05, "grad_norm": 6.203326225280762, "loss_scale": 4.0, "retries": 0, "seconds": 121.6, "max_memory_gb": 24.68, "pair": 0.4529771786183119, "sign": 1.3413779526017606, "keep": 0.42224278289359063, "leak": 0.7961041256785393, "replacement": 1.5338135808706284}
|
| 124 |
+
{"step": 124, "steps": 250, "lr": 3.051169562019339e-05, "grad_norm": 1.7249661684036255, "loss_scale": 4.0, "retries": 0, "seconds": 122.5, "max_memory_gb": 24.68, "pair": 0.6190766207873821, "sign": 0.586926179472357, "keep": 1.6236588235769887, "leak": 0.27785826226387655, "replacement": 1.7938337177038193}
|
| 125 |
+
{"step": 125, "steps": 250, "lr": 3.021207530574477e-05, "grad_norm": 2.7465410232543945, "loss_scale": 4.0, "retries": 0, "seconds": 123.5, "max_memory_gb": 24.68, "pair": 0.4314076229929924, "sign": 0.25371776800602674, "keep": 0.5426819508502376, "leak": 0.7791642630490978, "replacement": 1.0243985652923584}
|
| 126 |
+
{"step": 126, "steps": 250, "lr": 2.991196615521758e-05, "grad_norm": 2.471228837966919, "loss_scale": 4.0, "retries": 0, "seconds": 124.5, "max_memory_gb": 24.68, "pair": 0.679855614900589, "sign": 0.5176421091891825, "keep": 0.958994921296835, "leak": 0.9730621133930981, "replacement": 1.6614396572113037}
|
| 127 |
+
{"step": 127, "steps": 250, "lr": 2.9611422261571536e-05, "grad_norm": 1.1127760410308838, "loss_scale": 4.0, "retries": 0, "seconds": 125.4, "max_memory_gb": 24.68, "pair": 0.30372118204832077, "sign": 0.011344788043061271, "keep": 0.02300305589915297, "leak": 0.5256867595016956, "replacement": 0.4523167908191681}
|
| 128 |
+
{"step": 128, "steps": 250, "lr": 2.931049779612634e-05, "grad_norm": 3.1657896041870117, "loss_scale": 4.0, "retries": 0, "seconds": 126.4, "max_memory_gb": 24.68, "pair": 0.5646733194589615, "sign": 0.9768932670122012, "keep": 1.4034409972082358, "leak": 0.339621027931571, "replacement": 1.839745618402958}
|
| 129 |
+
{"step": 129, "steps": 250, "lr": 2.9009246998797557e-05, "grad_norm": 1.6516714096069336, "loss_scale": 4.0, "retries": 0, "seconds": 127.4, "max_memory_gb": 24.68, "pair": 0.4304255470633507, "sign": 0.09408192103728652, "keep": 0.15641345363110304, "leak": 0.38970046024769545, "replacement": 0.6530983299016953}
|
| 130 |
+
{"step": 130, "steps": 250, "lr": 2.8707724168320238e-05, "grad_norm": 2.420579671859741, "loss_scale": 4.0, "retries": 0, "seconds": 128.3, "max_memory_gb": 24.68, "pair": 0.6045572534203529, "sign": 1.8358284700661898, "keep": 0.06546512347995304, "leak": 0.2914478639140725, "replacement": 1.6280660033226013}
|
| 131 |
+
{"step": 131, "steps": 250, "lr": 2.840598365246184e-05, "grad_norm": 3.4105288982391357, "loss_scale": 4.0, "retries": 0, "seconds": 129.3, "max_memory_gb": 24.68, "pair": 0.6524209529161453, "sign": 0.2768017053604126, "keep": 0.2665095748088788, "leak": 0.4002143112011254, "replacement": 1.024130180478096}
|
| 132 |
+
{"step": 132, "steps": 250, "lr": 2.8104079838226383e-05, "grad_norm": 1.5018386840820312, "loss_scale": 4.0, "retries": 0, "seconds": 130.3, "max_memory_gb": 24.68, "pair": 0.27325224317610264, "sign": 0.03154206706676632, "keep": 0.4081881655129109, "leak": 0.8655259497463703, "replacement": 0.7094988524913788}
|
| 133 |
+
{"step": 133, "steps": 250, "lr": 2.7802067142051446e-05, "grad_norm": 1.5678316354751587, "loss_scale": 4.0, "retries": 0, "seconds": 131.3, "max_memory_gb": 24.68, "pair": 0.5532828718423843, "sign": 0.4709280231036246, "keep": 0.7809847589232959, "leak": 0.24742108257487416, "replacement": 1.2410945035517216}
|
| 134 |
+
{"step": 134, "steps": 250, "lr": 2.7500000000000004e-05, "grad_norm": 2.126420497894287, "loss_scale": 4.0, "retries": 0, "seconds": 132.2, "max_memory_gb": 24.68, "pair": 0.40391374565660954, "sign": 0.3420941570948344, "keep": 0.6075877867606323, "leak": 0.18452776968479156, "replacement": 0.9248866587877274}
|
| 135 |
+
{"step": 135, "steps": 250, "lr": 2.7197932857948566e-05, "grad_norm": 2.141313314437866, "loss_scale": 4.0, "retries": 0, "seconds": 133.2, "max_memory_gb": 24.68, "pair": 0.2745865248143673, "sign": 0.04515840986277908, "keep": 0.1493565067066811, "leak": 0.8897400200366974, "replacement": 0.5942789614200592}
|
| 136 |
+
{"step": 136, "steps": 250, "lr": 2.689592016177363e-05, "grad_norm": 5.485594749450684, "loss_scale": 4.0, "retries": 0, "seconds": 134.1, "max_memory_gb": 24.68, "pair": 0.3970496505498886, "sign": 0.528050844091922, "keep": 0.8786217436099832, "leak": 1.012985222041607, "replacement": 1.3536321893334389}
|
| 137 |
+
{"step": 137, "steps": 250, "lr": 2.659401634753817e-05, "grad_norm": 3.6370160579681396, "loss_scale": 4.0, "retries": 0, "seconds": 135.2, "max_memory_gb": 24.68, "pair": 0.23684372566640377, "sign": 0.019436235481407493, "keep": 0.035330946106114425, "leak": 1.8963804300874472, "replacement": 0.738322414457798}
|
| 138 |
+
{"step": 138, "steps": 250, "lr": 2.6292275831679764e-05, "grad_norm": 3.2138564586639404, "loss_scale": 4.0, "retries": 0, "seconds": 136.1, "max_memory_gb": 24.68, "pair": 0.35114831291139126, "sign": 0.036172071122564375, "keep": 0.4469276072923094, "leak": 0.4187017120420933, "replacement": 0.6973735764622688}
|
| 139 |
+
{"step": 139, "steps": 250, "lr": 2.5990753001202445e-05, "grad_norm": 2.92756986618042, "loss_scale": 4.0, "retries": 0, "seconds": 137.1, "max_memory_gb": 24.68, "pair": 0.40564409643411636, "sign": 0.8267145385907497, "keep": 0.02258693582552951, "leak": 1.3506988361477852, "replacement": 1.167969599366188}
|
| 140 |
+
{"step": 140, "steps": 250, "lr": 2.5689502203873672e-05, "grad_norm": 1.2596755027770996, "loss_scale": 4.0, "retries": 0, "seconds": 138.1, "max_memory_gb": 24.68, "pair": 0.22285116836428642, "sign": 0.033618945570196956, "keep": 0.024103640324028675, "leak": 0.9406469240784645, "replacement": 0.48687418550252914}
|
| 141 |
+
{"step": 141, "steps": 250, "lr": 2.538857773842847e-05, "grad_norm": 0.9429095983505249, "loss_scale": 4.0, "retries": 0, "seconds": 139.2, "max_memory_gb": 24.68, "pair": 0.3478705547749996, "sign": 0.21518962061963975, "keep": 0.009758755291841226, "leak": 0.8158927707772818, "replacement": 0.6643179208040237}
|
| 142 |
+
{"step": 142, "steps": 250, "lr": 2.5088033844782426e-05, "grad_norm": 2.900627851486206, "loss_scale": 4.0, "retries": 0, "seconds": 140.1, "max_memory_gb": 24.68, "pair": 0.3412773013114929, "sign": 0.5030464074807242, "keep": 0.4116782279616018, "leak": 0.7105276361107826, "replacement": 0.9762715101242065}
|
| 143 |
+
{"step": 143, "steps": 250, "lr": 2.4787924694255237e-05, "grad_norm": 3.2071032524108887, "loss_scale": 4.0, "retries": 0, "seconds": 141.1, "max_memory_gb": 24.68, "pair": 0.48483793437480927, "sign": 0.16131535754539073, "keep": 1.6388544037181418, "leak": 0.21299783140420914, "replacement": 1.4381722509860992}
|
| 144 |
+
{"step": 144, "steps": 250, "lr": 2.4488304379806627e-05, "grad_norm": 3.4348576068878174, "loss_scale": 4.0, "retries": 0, "seconds": 142.1, "max_memory_gb": 24.68, "pair": 0.5065350495278835, "sign": 0.8060922706499696, "keep": 0.1742171193473041, "leak": 0.5699788816273212, "replacement": 1.1391844302415848}
|
| 145 |
+
{"step": 145, "steps": 250, "lr": 2.4189226906286388e-05, "grad_norm": 1.0660526752471924, "loss_scale": 4.0, "retries": 0, "seconds": 143.1, "max_memory_gb": 24.68, "pair": 0.31241314485669136, "sign": 0.0787529197987169, "keep": 0.021484358804627846, "leak": 0.7568880170583725, "replacement": 0.5517537891864777}
|
| 146 |
+
{"step": 146, "steps": 250, "lr": 2.3890746180700395e-05, "grad_norm": 2.3744094371795654, "loss_scale": 4.0, "retries": 0, "seconds": 144.1, "max_memory_gb": 24.68, "pair": 0.41263275034725666, "sign": 0.14270207902882248, "keep": 0.01848960356437601, "leak": 1.5713757574558258, "replacement": 0.8860725313425064}
|
| 147 |
+
{"step": 147, "steps": 250, "lr": 2.359291600249407e-05, "grad_norm": 2.6210482120513916, "loss_scale": 4.0, "retries": 0, "seconds": 145.1, "max_memory_gb": 24.68, "pair": 0.6525727733969688, "sign": 0.8666579904966056, "keep": 1.049173059524037, "leak": 0.6075557470321655, "replacement": 1.7623772099614143}
|
| 148 |
+
{"step": 148, "steps": 250, "lr": 2.3295790053855392e-05, "grad_norm": 4.772956848144531, "loss_scale": 4.0, "retries": 0, "seconds": 146.1, "max_memory_gb": 24.68, "pair": 0.43165119737386703, "sign": 0.2871141789946705, "keep": 0.007964545569734582, "leak": 0.778653210029006, "replacement": 0.7738538905978203}
|
| 149 |
+
{"step": 149, "steps": 250, "lr": 2.299942189003901e-05, "grad_norm": 1.5993372201919556, "loss_scale": 4.0, "retries": 0, "seconds": 147.0, "max_memory_gb": 24.68, "pair": 0.3110211957246065, "sign": 0.10131788207218051, "keep": 0.04651652264328732, "leak": 0.9160260334610939, "replacement": 0.613944910466671}
|
| 150 |
+
{"step": 150, "steps": 250, "lr": 2.2703864929713144e-05, "grad_norm": 4.407200813293457, "loss_scale": 4.0, "retries": 0, "seconds": 148.0, "max_memory_gb": 24.68, "pair": 0.3587878979742527, "sign": 0.14255786139983684, "keep": 0.19251474435441196, "leak": 0.5277333895210177, "replacement": 0.6582575440406799}
|
| 151 |
+
{"step": 151, "steps": 250, "lr": 2.2409172445331257e-05, "grad_norm": 1.1575767993927002, "loss_scale": 4.0, "retries": 0, "seconds": 149.0, "max_memory_gb": 24.68, "pair": 0.3693622350692749, "sign": 0.0308405130635947, "keep": 0.017062029532098677, "leak": 0.5678703337907791, "replacement": 0.5352811068296432}
|
| 152 |
+
{"step": 152, "steps": 250, "lr": 2.211539755352996e-05, "grad_norm": 1.0181719064712524, "loss_scale": 4.0, "retries": 0, "seconds": 150.0, "max_memory_gb": 24.68, "pair": 0.24353107437491417, "sign": 0.008667381131090224, "keep": 0.04199328434651761, "leak": 0.8922502733767033, "replacement": 0.4919239580631256}
|
| 153 |
+
{"step": 153, "steps": 250, "lr": 2.1822593205555016e-05, "grad_norm": 2.3358850479125977, "loss_scale": 4.0, "retries": 0, "seconds": 151.0, "max_memory_gb": 24.68, "pair": 0.2439179103821516, "sign": 0.010678358259610832, "keep": 0.01679060560763901, "leak": 1.1095673143863678, "replacement": 0.5350442230701447}
|
| 154 |
+
{"step": 154, "steps": 250, "lr": 2.1530812177717236e-05, "grad_norm": 2.964175224304199, "loss_scale": 4.0, "retries": 0, "seconds": 152.0, "max_memory_gb": 24.68, "pair": 0.18215937726199627, "sign": 0.01367277029203251, "keep": 0.7884106142446399, "leak": 1.1250727325677872, "replacement": 0.8644692525267601}
|
| 155 |
+
{"step": 155, "steps": 250, "lr": 2.1240107061879815e-05, "grad_norm": 1.2480281591415405, "loss_scale": 4.0, "retries": 0, "seconds": 153.0, "max_memory_gb": 24.68, "pair": 0.21841161511838436, "sign": 0.010050420998595655, "keep": 0.4437708201029409, "leak": 0.7753098905086517, "replacement": 0.6391497254371643}
|
| 156 |
+
{"step": 156, "steps": 250, "lr": 2.095053025597896e-05, "grad_norm": 1.8952786922454834, "loss_scale": 4.0, "retries": 0, "seconds": 154.0, "max_memory_gb": 24.68, "pair": 0.29259344562888145, "sign": 0.006144313374534249, "keep": 0.029084746095350056, "leak": 0.42083728313446045, "replacement": 0.41541730612516403}
|
| 157 |
+
{"step": 157, "steps": 250, "lr": 2.066213395457945e-05, "grad_norm": 4.378449440002441, "loss_scale": 4.0, "retries": 0, "seconds": 154.9, "max_memory_gb": 24.68, "pair": 0.33126164227724075, "sign": 1.153975584777072, "keep": 0.13409005863650236, "leak": 3.069933071732521, "replacement": 1.742777705192566}
|
| 158 |
+
{"step": 158, "steps": 250, "lr": 2.0374970139466877e-05, "grad_norm": 2.2127127647399902, "loss_scale": 4.0, "retries": 0, "seconds": 155.9, "max_memory_gb": 24.68, "pair": 0.5135011933743954, "sign": 0.41975810602889396, "keep": 0.5128259602856815, "leak": 0.28320625447668135, "replacement": 1.0505947545170784}
|
| 159 |
+
{"step": 159, "steps": 250, "lr": 2.0089090570278213e-05, "grad_norm": 1.195952296257019, "loss_scale": 4.0, "retries": 0, "seconds": 156.9, "max_memory_gb": 24.74, "pair": 0.30005433037877083, "sign": 0.017511889338493347, "keep": 0.0034001142885244917, "leak": 0.799357570707798, "replacement": 0.5103497058153152}
|
| 160 |
+
{"step": 160, "steps": 250, "lr": 1.9804546775172457e-05, "grad_norm": 3.452138662338257, "loss_scale": 4.0, "retries": 0, "seconds": 157.9, "max_memory_gb": 24.74, "pair": 0.2846377184614539, "sign": 0.20410702704975847, "keep": 0.6169904968701303, "leak": 1.1810213346034288, "replacement": 0.9904417842626572}
|
| 161 |
+
{"step": 161, "steps": 250, "lr": 1.9521390041542946e-05, "grad_norm": 1.8808056116104126, "loss_scale": 4.0, "retries": 0, "seconds": 158.8, "max_memory_gb": 24.74, "pair": 0.4421350695192814, "sign": 0.17133392090909183, "keep": 0.24806637929850694, "leak": 0.25223573739640415, "replacement": 0.7148941606283188}
|
| 162 |
+
{"step": 162, "steps": 250, "lr": 1.923967140677313e-05, "grad_norm": 1.7440357208251953, "loss_scale": 4.0, "retries": 0, "seconds": 159.8, "max_memory_gb": 24.74, "pair": 0.2578346636146307, "sign": 0.03973015281371772, "keep": 0.23291424014587392, "leak": 0.8154453560709953, "replacement": 0.598018191754818}
|
| 163 |
+
{"step": 163, "steps": 250, "lr": 1.8959441649037335e-05, "grad_norm": 5.919051647186279, "loss_scale": 4.0, "retries": 0, "seconds": 160.8, "max_memory_gb": 24.74, "pair": 0.6913284696638584, "sign": 0.3652959226164967, "keep": 0.5121570862829685, "leak": 0.30195926735177636, "replacement": 1.205544799566269}
|
| 164 |
+
{"step": 164, "steps": 250, "lr": 1.868075127814832e-05, "grad_norm": 3.13069486618042, "loss_scale": 4.0, "retries": 0, "seconds": 161.8, "max_memory_gb": 24.74, "pair": 0.5458705686032772, "sign": 0.3196977456100285, "keep": 0.25766932958504185, "leak": 0.38526895456016064, "replacement": 0.9308713227510452}
|
| 165 |
+
{"step": 165, "steps": 250, "lr": 1.8403650526453143e-05, "grad_norm": 1.197651982307434, "loss_scale": 4.0, "retries": 0, "seconds": 162.8, "max_memory_gb": 24.74, "pair": 0.2818778473883867, "sign": 0.026033568079583347, "keep": 0.0037100082299730275, "leak": 0.42449150048196316, "replacement": 0.4028725177049637}
|
| 166 |
+
{"step": 166, "steps": 250, "lr": 1.812818933977912e-05, "grad_norm": 7.111949443817139, "loss_scale": 4.0, "retries": 0, "seconds": 163.8, "max_memory_gb": 24.74, "pair": 0.7320393733680248, "sign": 0.6860975371673703, "keep": 1.051590766030131, "leak": 0.7363628547172993, "replacement": 1.7849742472171783}
|
| 167 |
+
{"step": 167, "steps": 250, "lr": 1.785441736843128e-05, "grad_norm": 2.0594003200531006, "loss_scale": 4.0, "retries": 0, "seconds": 164.8, "max_memory_gb": 24.74, "pair": 0.4206007868051529, "sign": 0.08581647975370288, "keep": 0.08115256065502763, "leak": 0.36355146393179893, "replacement": 0.5949731767177582}
|
| 168 |
+
{"step": 168, "steps": 250, "lr": 1.7582383958243314e-05, "grad_norm": 0.8969761729240417, "loss_scale": 4.0, "retries": 0, "seconds": 165.7, "max_memory_gb": 24.74, "pair": 0.29294706881046295, "sign": 0.01769259970751591, "keep": 0.08450165075919358, "leak": 0.7336046695709229, "replacement": 0.5274453461170197}
|
| 169 |
+
{"step": 169, "steps": 250, "lr": 1.7312138141683164e-05, "grad_norm": 1.9390813112258911, "loss_scale": 4.0, "retries": 0, "seconds": 166.8, "max_memory_gb": 24.74, "pair": 0.44401348009705544, "sign": 0.16360813425853848, "keep": 0.2113099144771695, "leak": 0.3709780741482973, "replacement": 0.7242170348763466}
|
| 170 |
+
{"step": 170, "steps": 250, "lr": 1.704372862901521e-05, "grad_norm": 1.8429088592529297, "loss_scale": 4.0, "retries": 0, "seconds": 167.8, "max_memory_gb": 24.74, "pair": 0.3056911528110504, "sign": 0.006597108149435371, "keep": 0.10903658702954999, "leak": 0.8796000331640244, "replacement": 0.583407998085022}
|
| 171 |
+
{"step": 171, "steps": 250, "lr": 1.6777203799520573e-05, "grad_norm": 2.103221893310547, "loss_scale": 4.0, "retries": 0, "seconds": 168.8, "max_memory_gb": 24.74, "pair": 0.335673613473773, "sign": 0.18224081593507435, "keep": 0.09951791533808318, "leak": 0.2915319986641407, "replacement": 0.5494359731674194}
|
| 172 |
+
{"step": 172, "steps": 250, "lr": 1.6512611692777008e-05, "grad_norm": 0.8600014448165894, "loss_scale": 4.0, "retries": 0, "seconds": 169.8, "max_memory_gb": 24.74, "pair": 0.14692682772874832, "sign": 0.004003676702268422, "keep": 0.028461400303058326, "leak": 0.7182058077305555, "replacement": 0.34271081164479256}
|
| 173 |
+
{"step": 173, "steps": 250, "lr": 1.6250000000000005e-05, "grad_norm": 0.925463080406189, "loss_scale": 4.0, "retries": 0, "seconds": 170.7, "max_memory_gb": 24.74, "pair": 0.24233432998880744, "sign": 0.08960155304521322, "keep": 0.12419226591009647, "leak": 0.6210681625525467, "replacement": 0.5044982954859734}
|
| 174 |
+
{"step": 174, "steps": 250, "lr": 1.598941605544678e-05, "grad_norm": 1.3093831539154053, "loss_scale": 4.0, "retries": 0, "seconds": 171.7, "max_memory_gb": 24.74, "pair": 0.2097362782806158, "sign": 0.01917744503589347, "keep": 0.0707409664169063, "leak": 0.3770945370197296, "replacement": 0.34896911308169365}
|
| 175 |
+
{"step": 175, "steps": 250, "lr": 1.5730906827884506e-05, "grad_norm": 4.3768391609191895, "loss_scale": 4.0, "retries": 0, "seconds": 172.7, "max_memory_gb": 24.74, "pair": 0.38066601008176804, "sign": 0.29151175590232015, "keep": 1.5519672585651278, "leak": 0.362284058064688, "replacement": 1.3929765224456787}
|
| 176 |
+
{"step": 176, "steps": 250, "lr": 1.5474518912124476e-05, "grad_norm": 1.190477728843689, "loss_scale": 4.0, "retries": 0, "seconds": 173.7, "max_memory_gb": 24.74, "pair": 0.2009126003831625, "sign": 0.033548048842931166, "keep": 0.018340546783292666, "leak": 0.3932546186260879, "replacement": 0.3251705542206764}
|
| 177 |
+
{"step": 177, "steps": 250, "lr": 1.5220298520623646e-05, "grad_norm": 1.9467300176620483, "loss_scale": 4.0, "retries": 0, "seconds": 174.6, "max_memory_gb": 24.74, "pair": 0.3811505474150181, "sign": 0.4797499429550953, "keep": 0.008828286909192684, "leak": 0.3197923987172544, "replacement": 0.7053877487778664}
|
| 178 |
+
{"step": 178, "steps": 250, "lr": 1.4968291475155121e-05, "grad_norm": 6.643339157104492, "loss_scale": 4.0, "retries": 0, "seconds": 175.6, "max_memory_gb": 24.74, "pair": 0.323183988686651, "sign": 0.1982637852197513, "keep": 0.7369928188272752, "leak": 1.2258214393659728, "replacement": 1.0972676277160645}
|
| 179 |
+
{"step": 179, "steps": 250, "lr": 1.4718543198548998e-05, "grad_norm": 1.8378536701202393, "loss_scale": 4.0, "retries": 0, "seconds": 176.6, "max_memory_gb": 24.74, "pair": 0.6455861814320087, "sign": 0.5480197243159637, "keep": 0.5969122966125724, "leak": 0.17126962868496776, "replacement": 1.2608695775270462}
|
| 180 |
+
{"step": 180, "steps": 250, "lr": 1.4471098706505227e-05, "grad_norm": 1.1774799823760986, "loss_scale": 4.0, "retries": 0, "seconds": 177.5, "max_memory_gb": 24.74, "pair": 0.42974692955613136, "sign": 0.04521738493349403, "keep": 0.07555981492623687, "leak": 0.46087250020354986, "replacement": 0.6053536534309387}
|
| 181 |
+
{"step": 181, "steps": 250, "lr": 1.4226002599479743e-05, "grad_norm": 3.349534273147583, "loss_scale": 4.0, "retries": 0, "seconds": 178.5, "max_memory_gb": 24.74, "pair": 0.1607363298535347, "sign": 0.014400601154193282, "keep": 0.06311604578513652, "leak": 2.156872808933258, "replacement": 0.7387128621339798}
|
| 182 |
+
{"step": 182, "steps": 250, "lr": 1.3983299054645471e-05, "grad_norm": 2.1869895458221436, "loss_scale": 4.0, "retries": 0, "seconds": 179.5, "max_memory_gb": 24.74, "pair": 0.48240770399570465, "sign": 0.30007244797889143, "keep": 0.04563671709911432, "leak": 0.9386080056428909, "replacement": 0.8899142816662788}
|
| 183 |
+
{"step": 183, "steps": 250, "lr": 1.3743031817929724e-05, "grad_norm": 1.005779504776001, "loss_scale": 4.0, "retries": 0, "seconds": 180.5, "max_memory_gb": 24.74, "pair": 0.31044916063547134, "sign": 0.08587539324071258, "keep": 0.006576593546924414, "leak": 0.7186725661158562, "replacement": 0.5363432988524437}
|
| 184 |
+
{"step": 184, "steps": 250, "lr": 1.3505244196129164e-05, "grad_norm": 1.51499342918396, "loss_scale": 4.0, "retries": 0, "seconds": 181.5, "max_memory_gb": 24.74, "pair": 0.3093450926244259, "sign": 0.009474072372540832, "keep": 0.014866826171783032, "leak": 0.5942841041833162, "replacement": 0.47008658945560455}
|
| 185 |
+
{"step": 185, "steps": 250, "lr": 1.3269979049104014e-05, "grad_norm": 1.5799096822738647, "loss_scale": 4.0, "retries": 0, "seconds": 182.5, "max_memory_gb": 24.74, "pair": 0.2782104145735502, "sign": 0.04021154245128855, "keep": 0.10650027316296473, "leak": 0.806831508903997, "replacement": 0.5532741993665695}
|
| 186 |
+
{"step": 186, "steps": 250, "lr": 1.3037278782052863e-05, "grad_norm": 2.3505427837371826, "loss_scale": 4.0, "retries": 0, "seconds": 183.5, "max_memory_gb": 24.74, "pair": 0.3946960009634495, "sign": 0.02587278315331787, "keep": 0.9038320262850448, "leak": 0.4328845962882042, "replacement": 0.9677695631980896}
|
| 187 |
+
{"step": 187, "steps": 250, "lr": 1.2807185337869343e-05, "grad_norm": 1.2628318071365356, "loss_scale": 4.0, "retries": 0, "seconds": 184.5, "max_memory_gb": 24.74, "pair": 0.27524733915925026, "sign": 0.009416723623871803, "keep": 0.43667835829546675, "leak": 0.7982598748058081, "replacement": 0.6978598535060883}
|
| 188 |
+
{"step": 188, "steps": 250, "lr": 1.2579740189582115e-05, "grad_norm": 2.129549503326416, "loss_scale": 4.0, "retries": 0, "seconds": 185.4, "max_memory_gb": 24.74, "pair": 0.6088897809386253, "sign": 0.851779750577407, "keep": 0.8403266891627936, "leak": 0.2092035785317421, "replacement": 1.5072439312934875}
|
| 189 |
+
{"step": 189, "steps": 250, "lr": 1.2354984332879668e-05, "grad_norm": 2.501600980758667, "loss_scale": 4.0, "retries": 0, "seconds": 186.4, "max_memory_gb": 24.74, "pair": 0.2646272834390402, "sign": 0.3826061834115535, "keep": 0.9298709113791119, "leak": 1.0335824297235376, "replacement": 1.1792614609003067}
|
| 190 |
+
{"step": 190, "steps": 250, "lr": 1.2132958278721064e-05, "grad_norm": 2.083996295928955, "loss_scale": 4.0, "retries": 0, "seconds": 187.4, "max_memory_gb": 24.74, "pair": 0.14123068377375603, "sign": 0.006985537242144346, "keep": 0.010601666159345768, "leak": 1.134407415986061, "replacement": 0.43362613767385483}
|
| 191 |
+
{"step": 191, "steps": 250, "lr": 1.1913702046034016e-05, "grad_norm": 11.4720458984375, "loss_scale": 4.0, "retries": 0, "seconds": 188.3, "max_memory_gb": 24.74, "pair": 0.4142060950398445, "sign": 0.3206815130542964, "keep": 0.4568268754044311, "leak": 0.8636348620057106, "replacement": 1.0188689678907394}
|
| 192 |
+
{"step": 192, "steps": 250, "lr": 1.1697255154501756e-05, "grad_norm": 1.3229846954345703, "loss_scale": 4.0, "retries": 0, "seconds": 189.3, "max_memory_gb": 24.74, "pair": 0.5774802044034004, "sign": 0.4806948087643832, "keep": 1.0591058093123138, "leak": 0.5161787131801248, "replacement": 1.4764251708984375}
|
| 193 |
+
{"step": 193, "steps": 250, "lr": 1.1483656617439842e-05, "grad_norm": 1.1804660558700562, "loss_scale": 4.0, "retries": 0, "seconds": 190.3, "max_memory_gb": 24.74, "pair": 0.2750133126974106, "sign": 0.03187513304874301, "keep": 0.02539148191544882, "leak": 0.7256048172712326, "replacement": 0.4850478321313858}
|
| 194 |
+
{"step": 194, "steps": 250, "lr": 1.1272944934764177e-05, "grad_norm": 1.7860713005065918, "loss_scale": 4.0, "retries": 0, "seconds": 191.3, "max_memory_gb": 24.74, "pair": 0.48775939643383026, "sign": 0.3252329397946596, "keep": 0.022208700433111517, "leak": 0.45614401675993577, "replacement": 0.7755162119865417}
|
| 195 |
+
{"step": 195, "steps": 250, "lr": 1.1065158086051684e-05, "grad_norm": 3.5462958812713623, "loss_scale": 4.0, "retries": 0, "seconds": 192.3, "max_memory_gb": 24.74, "pair": 0.34055744111537933, "sign": 0.06319637046544813, "keep": 0.009815837467613164, "leak": 1.089040668681264, "replacement": 0.6493237316608429}
|
| 196 |
+
{"step": 196, "steps": 250, "lr": 1.086033352369467e-05, "grad_norm": 1.5107567310333252, "loss_scale": 4.0, "retries": 0, "seconds": 193.3, "max_memory_gb": 24.74, "pair": 0.1270954292267561, "sign": 0.007908565632533282, "keep": 0.02212886786890067, "leak": 1.138126090168953, "replacement": 0.4266456700861454}
|
| 197 |
+
{"step": 197, "steps": 250, "lr": 1.0658508166150224e-05, "grad_norm": 1.7415881156921387, "loss_scale": 4.0, "retries": 0, "seconds": 194.3, "max_memory_gb": 24.74, "pair": 0.2336271032691002, "sign": 0.03188267751829699, "keep": 0.006007691714330576, "leak": 1.043013647198677, "replacement": 0.5133256986737251}
|
| 198 |
+
{"step": 198, "steps": 250, "lr": 1.0459718391285954e-05, "grad_norm": 3.5624310970306396, "loss_scale": 4.0, "retries": 0, "seconds": 195.3, "max_memory_gb": 24.74, "pair": 0.23703095503151417, "sign": 0.08420685806777328, "keep": 0.1411397090087121, "leak": 1.037549201399088, "replacement": 0.6090915352106094}
|
| 199 |
+
{"step": 199, "steps": 250, "lr": 1.0264000029822999e-05, "grad_norm": 4.891290664672852, "loss_scale": 4.0, "retries": 0, "seconds": 196.2, "max_memory_gb": 24.74, "pair": 0.17690121615305543, "sign": 0.03408937591302674, "keep": 0.5517067907203455, "leak": 1.89957071444951, "replacement": 0.9446919932961464}
|
| 200 |
+
{"step": 200, "steps": 250, "lr": 1.0071388358877775e-05, "grad_norm": 2.9174036979675293, "loss_scale": 4.0, "retries": 0, "seconds": 197.2, "max_memory_gb": 24.74, "pair": 0.4843905568122864, "sign": 0.38059647026238963, "keep": 1.5242545399378287, "leak": 0.8041630471125245, "replacement": 1.6378567814826965}
|
| 201 |
+
{"step": 201, "steps": 250, "lr": 9.881918095603501e-06, "grad_norm": 2.371588945388794, "loss_scale": 4.0, "retries": 0, "seconds": 198.2, "max_memory_gb": 24.74, "pair": 0.4327060580253601, "sign": 0.18774729873985052, "keep": 0.16135402317740954, "leak": 0.374745465349406, "replacement": 0.7009431049227715}
|
| 202 |
+
{"step": 202, "steps": 250, "lr": 9.695623390932612e-06, "grad_norm": 2.046698570251465, "loss_scale": 4.0, "retries": 0, "seconds": 199.2, "max_memory_gb": 24.74, "pair": 0.3701499179005623, "sign": 0.15056091744918376, "keep": 0.006144828402284475, "leak": 0.7333631739020348, "replacement": 0.631843589246273}
|
| 203 |
+
{"step": 203, "steps": 250, "lr": 9.51253782342123e-06, "grad_norm": 1.3998466730117798, "loss_scale": 4.0, "retries": 0, "seconds": 200.1, "max_memory_gb": 24.74, "pair": 0.16567211411893368, "sign": 0.0700190979114268, "keep": 0.0185822104831459, "leak": 0.7682474255561829, "replacement": 0.40203462541103363}
|
| 204 |
+
{"step": 204, "steps": 250, "lr": 9.332694393196846e-06, "grad_norm": 1.1246370077133179, "loss_scale": 4.0, "retries": 0, "seconds": 201.1, "max_memory_gb": 24.74, "pair": 0.36870966851711273, "sign": 0.02887213509529829, "keep": 0.04513326473534107, "leak": 0.8907244328584056, "replacement": 0.6283934786915779}
|
| 205 |
+
{"step": 205, "steps": 250, "lr": 9.156125516010256e-06, "grad_norm": 1.0618175268173218, "loss_scale": 4.0, "retries": 0, "seconds": 202.1, "max_memory_gb": 24.74, "pair": 0.2240634672343731, "sign": 0.007742583460640162, "keep": 0.002498421774362214, "leak": 0.5581760108470917, "replacement": 0.36872797831892967}
|
| 206 |
+
{"step": 206, "steps": 250, "lr": 8.982863017392733e-06, "grad_norm": 1.7629729509353638, "loss_scale": 4.0, "retries": 0, "seconds": 203.1, "max_memory_gb": 24.74, "pair": 0.26132189668715, "sign": 0.04295516636921093, "keep": 0.023978435427125078, "leak": 0.5144080258905888, "replacement": 0.423390693962574}
|
| 207 |
+
{"step": 207, "steps": 250, "lr": 8.81293812691974e-06, "grad_norm": 1.0684878826141357, "loss_scale": 4.0, "retries": 0, "seconds": 204.1, "max_memory_gb": 24.74, "pair": 0.4828185476362705, "sign": 0.17439320450648665, "keep": 0.08673075627211801, "leak": 0.2956607041414827, "replacement": 0.6872957237064838}
|
| 208 |
+
{"step": 208, "steps": 250, "lr": 8.646381472581927e-06, "grad_norm": 0.9857134222984314, "loss_scale": 4.0, "retries": 0, "seconds": 205.1, "max_memory_gb": 24.74, "pair": 0.14013178646564484, "sign": 0.004056560166645795, "keep": 0.028732059182402736, "leak": 0.6279943622648716, "replacement": 0.31352468580007553}
|
| 209 |
+
{"step": 209, "steps": 250, "lr": 8.483223075264622e-06, "grad_norm": 2.091301918029785, "loss_scale": 4.0, "retries": 0, "seconds": 206.1, "max_memory_gb": 24.74, "pair": 0.48070348612964153, "sign": 0.3862815306056291, "keep": 0.2951915383746382, "leak": 0.8960195276886225, "replacement": 1.0454448387026787}
|
| 210 |
+
{"step": 210, "steps": 250, "lr": 8.323492343336747e-06, "grad_norm": 1.6871013641357422, "loss_scale": 4.0, "retries": 0, "seconds": 207.0, "max_memory_gb": 24.74, "pair": 0.5287983790040016, "sign": 0.49231750273611397, "keep": 1.2032946277159908, "leak": 0.425346240401268, "replacement": 1.4829409569501877}
|
| 211 |
+
{"step": 211, "steps": 250, "lr": 8.167218067350122e-06, "grad_norm": 1.0033769607543945, "loss_scale": 4.0, "retries": 0, "seconds": 208.0, "max_memory_gb": 24.74, "pair": 0.506950818002224, "sign": 0.2647515068529174, "keep": 0.09285984933376312, "leak": 0.5255058240145445, "replacement": 0.8171329498291016}
|
| 212 |
+
{"step": 212, "steps": 250, "lr": 8.01442841485013e-06, "grad_norm": 3.4949138164520264, "loss_scale": 4.0, "retries": 0, "seconds": 209.0, "max_memory_gb": 24.74, "pair": 0.5569616360589862, "sign": 0.16003151988843456, "keep": 0.040282426634803414, "leak": 1.1970049431547523, "replacement": 0.9563698470592499}
|
| 213 |
+
{"step": 213, "steps": 250, "lr": 7.865150925298678e-06, "grad_norm": 2.9296116828918457, "loss_scale": 4.0, "retries": 0, "seconds": 210.0, "max_memory_gb": 24.74, "pair": 0.4737689904868603, "sign": 0.33381319395266473, "keep": 0.47312572484042903, "leak": 0.49495464749634266, "replacement": 1.000977136194706}
|
| 214 |
+
{"step": 214, "steps": 250, "lr": 7.719412505110375e-06, "grad_norm": 1.617417573928833, "loss_scale": 4.0, "retries": 0, "seconds": 211.1, "max_memory_gb": 24.74, "pair": 0.46893009543418884, "sign": 0.38097140251193196, "keep": 1.190083509997521, "leak": 0.07369629386812449, "replacement": 1.2728816121816635}
|
| 215 |
+
{"step": 215, "steps": 250, "lr": 7.57723942280278e-06, "grad_norm": 2.5690321922302246, "loss_scale": 4.0, "retries": 0, "seconds": 212.1, "max_memory_gb": 24.74, "pair": 0.7520340923219919, "sign": 0.8124584667384624, "keep": 1.335634930688002, "leak": 0.06989107791201832, "replacement": 1.8435536064207554}
|
| 216 |
+
{"step": 216, "steps": 250, "lr": 7.438657304261687e-06, "grad_norm": 1.4995585680007935, "loss_scale": 4.0, "retries": 0, "seconds": 213.1, "max_memory_gb": 24.74, "pair": 0.18960301019251347, "sign": 0.03982136072590947, "keep": 0.021589665633655386, "leak": 0.8255067840218544, "replacement": 0.42668522894382477}
|
| 217 |
+
{"step": 217, "steps": 250, "lr": 7.30369112812219e-06, "grad_norm": 1.5815818309783936, "loss_scale": 4.0, "retries": 0, "seconds": 214.1, "max_memory_gb": 24.74, "pair": 0.21437561139464378, "sign": 0.022580822667805478, "keep": 0.000701454293448478, "leak": 0.9977532885968685, "replacement": 0.47545506805181503}
|
| 218 |
+
{"step": 218, "steps": 250, "lr": 7.172365221266403e-06, "grad_norm": 4.834105968475342, "loss_scale": 4.0, "retries": 0, "seconds": 215.1, "max_memory_gb": 24.74, "pair": 0.9167595282196999, "sign": 1.685480386018753, "keep": 0.27066416945308447, "leak": 0.20418442549271276, "replacement": 1.9458779394626617}
|
| 219 |
+
{"step": 219, "steps": 250, "lr": 7.0447032544387154e-06, "grad_norm": 1.0190497636795044, "loss_scale": 4.0, "retries": 0, "seconds": 216.1, "max_memory_gb": 24.74, "pair": 0.2967536374926567, "sign": 0.08112943809828721, "keep": 0.0055534764906042255, "leak": 0.42936646938323975, "replacement": 0.44743672758340836}
|
| 220 |
+
{"step": 220, "steps": 250, "lr": 6.920728237979274e-06, "grad_norm": 1.7751972675323486, "loss_scale": 4.0, "retries": 0, "seconds": 217.2, "max_memory_gb": 24.74, "pair": 0.38288741558790207, "sign": 0.05504863930400461, "keep": 0.06588912999541208, "leak": 0.17596310563385487, "replacement": 0.4873470738530159}
|
| 221 |
+
{"step": 221, "steps": 250, "lr": 6.800462517676456e-06, "grad_norm": 0.9419345855712891, "loss_scale": 4.0, "retries": 0, "seconds": 218.1, "max_memory_gb": 24.74, "pair": 0.35291279293596745, "sign": 0.18416035926202312, "keep": 0.008232416468672454, "leak": 0.7376070022583008, "replacement": 0.6335109248757362}
|
| 222 |
+
{"step": 222, "steps": 250, "lr": 6.683927770739222e-06, "grad_norm": 1.2969467639923096, "loss_scale": 4.0, "retries": 0, "seconds": 219.1, "max_memory_gb": 24.74, "pair": 0.2707657655701041, "sign": 0.0322465953649953, "keep": 0.05516422617256467, "leak": 0.47926873131655157, "replacement": 0.4342883676290512}
|
| 223 |
+
{"step": 223, "steps": 250, "lr": 6.571145001889892e-06, "grad_norm": 4.133554935455322, "loss_scale": 4.0, "retries": 0, "seconds": 220.1, "max_memory_gb": 24.74, "pair": 0.48890840262174606, "sign": 0.06695785454940051, "keep": 0.007497886304008716, "leak": 0.5132066905498505, "replacement": 0.6544379591941833}
|
| 224 |
+
{"step": 224, "steps": 250, "lr": 6.46213453957817e-06, "grad_norm": 1.5987920761108398, "loss_scale": 4.0, "retries": 0, "seconds": 221.1, "max_memory_gb": 24.74, "pair": 0.39486249536275864, "sign": 0.04846268240362406, "keep": 0.008707286287972238, "leak": 0.30649596940202173, "replacement": 0.5000714883208275}
|
| 225 |
+
{"step": 225, "steps": 250, "lr": 6.356916032317063e-06, "grad_norm": 2.07977032661438, "loss_scale": 4.0, "retries": 0, "seconds": 222.1, "max_memory_gb": 24.74, "pair": 0.24058892577886581, "sign": 0.03514283709228039, "keep": 0.0025526075423840666, "leak": 0.865639753639698, "replacement": 0.4758465774357319}
|
| 226 |
+
{"step": 226, "steps": 250, "lr": 6.255508445141349e-06, "grad_norm": 2.5339818000793457, "loss_scale": 4.0, "retries": 0, "seconds": 223.1, "max_memory_gb": 24.74, "pair": 0.2159843109548092, "sign": 0.04118296655360609, "keep": 1.090549261556589, "leak": 0.7499679075554013, "replacement": 0.969342440366745}
|
| 227 |
+
{"step": 227, "steps": 250, "lr": 6.157930056189228e-06, "grad_norm": 1.8237663507461548, "loss_scale": 4.0, "retries": 0, "seconds": 224.0, "max_memory_gb": 24.74, "pair": 0.22300690785050392, "sign": 0.016372972051613033, "keep": 0.08424234494077609, "leak": 0.5470219943672419, "replacement": 0.4100700579583645}
|
| 228 |
+
{"step": 228, "steps": 250, "lr": 6.064198453407796e-06, "grad_norm": 1.1488839387893677, "loss_scale": 4.0, "retries": 0, "seconds": 225.0, "max_memory_gb": 24.74, "pair": 0.2196543626487255, "sign": 0.05304894014261663, "keep": 0.0013933365426055389, "leak": 1.0136514455080032, "replacement": 0.5002883672714233}
|
| 229 |
+
{"step": 229, "steps": 250, "lr": 5.974330531382935e-06, "grad_norm": 1.9709510803222656, "loss_scale": 4.0, "retries": 0, "seconds": 226.0, "max_memory_gb": 24.74, "pair": 0.3890041373670101, "sign": 0.12932838289998472, "keep": 0.12085858813910022, "leak": 1.0145957876520697, "replacement": 0.7677465677261353}
|
| 230 |
+
{"step": 230, "steps": 250, "lr": 5.888342488294125e-06, "grad_norm": 18.176668167114258, "loss_scale": 4.0, "retries": 0, "seconds": 227.0, "max_memory_gb": 24.74, "pair": 0.8744138032197952, "sign": 0.5127525217831135, "keep": 0.6344066113888402, "leak": 0.27520023472607136, "replacement": 1.5167934149503708}
|
| 231 |
+
{"step": 231, "steps": 250, "lr": 5.806249822994849e-06, "grad_norm": 2.004164695739746, "loss_scale": 4.0, "retries": 0, "seconds": 227.9, "max_memory_gb": 24.74, "pair": 0.6058543808758259, "sign": 0.7472716895863414, "keep": 0.03558024297899465, "leak": 0.5152346715331078, "replacement": 1.12608902156353}
|
| 232 |
+
{"step": 232, "steps": 250, "lr": 5.728067332218994e-06, "grad_norm": 3.038922071456909, "loss_scale": 4.0, "retries": 0, "seconds": 228.9, "max_memory_gb": 24.74, "pair": 0.1928755436092615, "sign": 0.023541983246104792, "keep": 0.04706697358051315, "leak": 1.8473247811198235, "replacement": 0.6900112107396126}
|
| 233 |
+
{"step": 233, "steps": 250, "lr": 5.65380910791383e-06, "grad_norm": 1.165693998336792, "loss_scale": 4.0, "retries": 0, "seconds": 229.9, "max_memory_gb": 24.74, "pair": 0.24169407598674297, "sign": 0.00825476658064872, "keep": 0.011380698531866074, "leak": 0.8753743432462215, "replacement": 0.4703553840517998}
|
| 234 |
+
{"step": 234, "steps": 250, "lr": 5.583488534700022e-06, "grad_norm": 1.892722725868225, "loss_scale": 4.0, "retries": 0, "seconds": 230.8, "max_memory_gb": 24.74, "pair": 0.23441801965236664, "sign": 0.028082188218832016, "keep": 0.14596780325427972, "leak": 0.6646902710199356, "replacement": 0.48761559277772903}
|
| 235 |
+
{"step": 235, "steps": 250, "lr": 5.517118287459118e-06, "grad_norm": 8.428054809570312, "loss_scale": 4.0, "retries": 0, "seconds": 231.8, "max_memory_gb": 24.74, "pair": 0.7598515395075083, "sign": 0.47231031709816307, "keep": 0.9348353873810993, "leak": 1.784026812762022, "replacement": 1.9094310104846954}
|
| 236 |
+
{"step": 236, "steps": 250, "lr": 5.454710329048978e-06, "grad_norm": 1.3471378087997437, "loss_scale": 4.0, "retries": 0, "seconds": 232.8, "max_memory_gb": 24.74, "pair": 0.33639834355562925, "sign": 0.11192536179441959, "keep": 0.006353531843842575, "leak": 0.151479653082788, "replacement": 0.43340769596397877}
|
| 237 |
+
{"step": 237, "steps": 250, "lr": 5.3962759081475436e-06, "grad_norm": 3.346125602722168, "loss_scale": 4.0, "retries": 0, "seconds": 233.8, "max_memory_gb": 24.74, "pair": 0.450339263305068, "sign": 0.46552169462665915, "keep": 0.19974413250020007, "leak": 0.4377942308783531, "replacement": 0.8924207240343094}
|
| 238 |
+
{"step": 238, "steps": 250, "lr": 5.341825557225321e-06, "grad_norm": 0.9563127160072327, "loss_scale": 4.0, "retries": 0, "seconds": 234.7, "max_memory_gb": 24.74, "pair": 0.22309228964149952, "sign": 0.01841454254463315, "keep": 0.03445546329021454, "leak": 0.6894334144890308, "replacement": 0.4218856580555439}
|
| 239 |
+
{"step": 239, "steps": 250, "lr": 5.291369090646961e-06, "grad_norm": 2.5931529998779297, "loss_scale": 4.0, "retries": 0, "seconds": 235.7, "max_memory_gb": 24.74, "pair": 0.6473049819469452, "sign": 1.4842074606567621, "keep": 0.16969206732755993, "leak": 0.5069979461841285, "replacement": 1.6010043025016785}
|
| 240 |
+
{"step": 240, "steps": 250, "lr": 5.244915602902289e-06, "grad_norm": 1.3975422382354736, "loss_scale": 4.0, "retries": 0, "seconds": 236.7, "max_memory_gb": 24.74, "pair": 0.4015575051307678, "sign": 0.3671450054971501, "keep": 0.1927434224053286, "leak": 0.5408448986709118, "replacement": 0.8167129382491112}
|
| 241 |
+
{"step": 241, "steps": 250, "lr": 5.20247346696706e-06, "grad_norm": 4.938520908355713, "loss_scale": 4.0, "retries": 0, "seconds": 237.8, "max_memory_gb": 24.74, "pair": 0.8988118954002857, "sign": 0.6350676498841494, "keep": 0.3276834362145564, "leak": 0.35816960463125724, "replacement": 1.4697298556566238}
|
| 242 |
+
{"step": 242, "steps": 250, "lr": 5.164050332793787e-06, "grad_norm": 1.6443690061569214, "loss_scale": 4.0, "retries": 0, "seconds": 238.7, "max_memory_gb": 24.74, "pair": 0.44303503073751926, "sign": 0.048998785903677344, "keep": 0.004161012405802467, "leak": 0.615154180675745, "replacement": 0.6234034523367882}
|
| 243 |
+
{"step": 243, "steps": 250, "lr": 5.129653125932873e-06, "grad_norm": 0.9997146725654602, "loss_scale": 4.0, "retries": 0, "seconds": 239.8, "max_memory_gb": 24.74, "pair": 0.23803298361599445, "sign": 0.02163693925831467, "keep": 0.049687670019920915, "leak": 0.8157923296093941, "replacement": 0.47764337062835693}
|
| 244 |
+
{"step": 244, "steps": 250, "lr": 5.099288046284329e-06, "grad_norm": 2.7651679515838623, "loss_scale": 4.0, "retries": 0, "seconds": 240.7, "max_memory_gb": 24.74, "pair": 0.40322357043623924, "sign": 0.22297428781166673, "keep": 0.052469597198069096, "leak": 1.2197195496410131, "replacement": 0.8458753824234009}
|
| 245 |
+
{"step": 245, "steps": 250, "lr": 5.072960566980277e-06, "grad_norm": 3.420501470565796, "loss_scale": 4.0, "retries": 0, "seconds": 241.7, "max_memory_gb": 24.74, "pair": 0.6719750314950943, "sign": 0.3994084168225527, "keep": 0.7294289264827967, "leak": 0.21866723615676165, "replacement": 1.2910604923963547}
|
| 246 |
+
{"step": 246, "steps": 250, "lr": 5.050675433398435e-06, "grad_norm": 0.9370740056037903, "loss_scale": 4.0, "retries": 0, "seconds": 242.7, "max_memory_gb": 24.74, "pair": 0.2067292369902134, "sign": 0.006454047746956348, "keep": 0.004825383817660622, "leak": 0.4621390737593174, "replacement": 0.32790372148156166}
|
| 247 |
+
{"step": 247, "steps": 250, "lr": 5.032436662306811e-06, "grad_norm": 3.4916539192199707, "loss_scale": 4.0, "retries": 0, "seconds": 243.7, "max_memory_gb": 24.74, "pair": 0.27882567793130875, "sign": 0.026062700781039894, "keep": 0.40578716462414377, "leak": 0.588326059281826, "replacement": 0.6418321058154106}
|
| 248 |
+
{"step": 248, "steps": 250, "lr": 5.018247541139684e-06, "grad_norm": 1.0092524290084839, "loss_scale": 4.0, "retries": 0, "seconds": 244.7, "max_memory_gb": 24.74, "pair": 0.2234644591808319, "sign": 0.03656419966137037, "keep": 0.04224802361932234, "leak": 0.6100469517987221, "replacement": 0.4153823032975197}
|
| 249 |
+
{"step": 249, "steps": 250, "lr": 5.00811062740508e-06, "grad_norm": 1.568901538848877, "loss_scale": 4.0, "retries": 0, "seconds": 245.7, "max_memory_gb": 24.74, "pair": 0.25128657929599285, "sign": 0.08098955918103456, "keep": 0.04905144526128424, "leak": 0.8649901300668716, "replacement": 0.5325546264648438}
|
| 250 |
+
{"step": 250, "steps": 250, "lr": 5.002027748223778e-06, "grad_norm": 2.3615927696228027, "loss_scale": 4.0, "retries": 0, "seconds": 246.7, "max_memory_gb": 24.74, "pair": 0.48619531095027924, "sign": 0.042292216792702675, "keep": 0.12492530727104167, "leak": 0.1866348385810852, "replacement": 0.6164627745747566}
|
runs/single_internvl/replacement/L07_s13/adapter/README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: /data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:/data/OPSD/assets/models/InternVL3_5-8B-HF
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- 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. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
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).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.20.0
|
runs/single_internvl/replacement/L07_s13/adapter/adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/data/OPSD/assets/models/InternVL3_5-8B-HF",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 16,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": null,
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"model.language_model.layers.7.mlp.down_proj",
|
| 35 |
+
"model.language_model.layers.7.mlp.gate_proj",
|
| 36 |
+
"model.language_model.layers.7.self_attn.v_proj",
|
| 37 |
+
"model.language_model.layers.7.self_attn.k_proj",
|
| 38 |
+
"model.language_model.layers.7.self_attn.o_proj",
|
| 39 |
+
"model.language_model.layers.7.self_attn.q_proj",
|
| 40 |
+
"model.language_model.layers.7.mlp.up_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|