pod backup 2026-11-07: e45 (0-44)

e45/S1-E45-NOISE-ts0-syn/DONE

@@ -0,0 +1 @@
+2026-11-02 01:38:41

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter/README.md

@@ -0,0 +1,206 @@
+---
+base_model: /workspace/jev_kt/models/hf/JEV-9B
+library_name: peft
+tags:
+- base_model:adapter:/workspace/jev_kt/models/hf/JEV-9B
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.2

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter/adapter_config.json

@@ -0,0 +1,57 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": "/workspace/jev_kt/models/hf/JEV-9B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "up_proj",
+    "in_proj_qkv",
+    "q_proj",
+    "v_proj",
+    "k_proj",
+    "gate_proj",
+    "in_proj_z",
+    "down_proj",
+    "out_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f72f33f63052eb0a8ad9ee52fc05faf8aa81d26f2b890013df600846b4e1a973
+size 160486456

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter_vllm/adapter_config.json

@@ -0,0 +1,58 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": null,
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64.0,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "gate_proj",
+    "in_proj_qkv",
+    "in_proj_z",
+    "k_proj",
+    "lm_head",
+    "o_proj",
+    "out_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e87aab2f9812f0b5f3b0ecfbce346a8d783796d2a38a7c69538be53ff7f96ec
+size 176641832

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.0004553915932774544,
+  -0.0004556486092042178,
+  -0.002269732067361474,
+  -0.0013550732983276248,
+  -0.00046324654249474406,
+  0.0020020289812237024,
+  0.0013502774527296424,
+  -0.0022309948690235615,
+  -0.00030999755836091936,
+  -0.0003666585253085941,
+  0.00019026632071472704,
+  0.0005198323633521795,
+  0.0007564538391306996,
+  -0.003509402507916093,
+  -0.004493982531130314,
+  -0.0051369937136769295,
+  -0.00585640873759985,
+  -0.004349171184003353,
+  -0.004932026378810406,
+  -0.004964736290276051,
+  -0.003783627413213253,
+  -0.0033698950428515673,
+  -0.005658139009028673,
+  -0.0034731589257717133
+ ],
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "note": "decision logits = lm_head-LoRA logprobs of verbalizer ids + bias, then per-kind temperature"
+}

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/calibration.json

@@ -0,0 +1,29 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 0.8911908555798087,
+  "choice": 1.0220903707073858,
+  "score": 0.9871672059187314
+ },
+ "fit": {
+  "noul": {
+   "n": 1087,
+   "T": 0.8911908555798087,
+   "ce_before": 0.5531377196311951,
+   "ce_after": 0.5520544052124023
+  },
+  "choice": {
+   "n": 60,
+   "T": 1.0220903707073858,
+   "ce_before": 0.6708876490592957,
+   "ce_after": 0.6707836985588074
+  },
+  "score": {
+   "n": 53,
+   "T": 0.9871672059187314,
+   "ce_before": 1.0566658973693848,
+   "ce_after": 1.0566120147705078
+  }
+ },
+ "source": "models/e45/S1-E45-NOISE-ts0-syn/ckpt-16000"
+}

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fca975dc37f12b26965b8fdab9d80bd1ceb00277f42dbc4194e987ecf1f571d
+size 393464

e45/S1-E45-NOISE-ts0-syn/ckpt-16000/judge_config.json

@@ -0,0 +1,164 @@
+{
+ "base_model_path": "/root/models/Qwen3.5-9B",
+ "hidden_size": 4096,
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "kinds": [
+  "noul",
+  "choice",
+  "score"
+ ],
+ "model_name": "jev-judge-qwen35-9b",
+ "model_version": "0.8.0",
+ "lm_head_included": true,
+ "weights_mode": "unmerged",
+ "adapter_subfolder": "adapter",
+ "lora": {
+  "r": 16,
+  "lora_alpha": 32,
+  "lora_dropout": 0.05,
+  "target_modules": [
+   "down_proj",
+   "gate_proj",
+   "in_proj_qkv",
+   "in_proj_z",
+   "k_proj",
+   "o_proj",
+   "out_proj",
+   "q_proj",
+   "up_proj",
+   "v_proj"
+  ]
+ },
+ "trained_stage": "e16",
+ "source_checkpoint": "models/e45/S1-E45-NOISE-ts0-syn/ckpt-16000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e45/E45-NOISE",
+  "data_manifest": {
+   "arm": "E45-NOISE",
+   "name": null,
+   "train_sha256": "53ee5956d8073d8dad59f35016d62ec8cf548a2dec30ffbbe807ee3b8851bb89",
+   "n_train": 16000
+  },
+  "init": "released",
+  "objective": "s-log",
+  "sigma": 0.5,
+  "K": 8,
+  "reward": "brier",
+  "clip": 0.2,
+  "passes": 1,
+  "lr": 0.0001,
+  "lr_head": 0.0002,
+  "rows_per_step": 32,
+  "train_seed": 0,
+  "replay": "data/e16/replay/train.jsonl",
+  "replay_frac": 0.05,
+  "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e",
+  "max_len": 3072,
+  "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps",
+  "total_rows": 16000,
+  "lora": {
+   "r": 16,
+   "lora_alpha": 32,
+   "lora_dropout": 0.05
+  },
+  "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3",
+  "event": "ckpt",
+  "tag": "ckpt-16000",
+  "rows_seen": 16000,
+  "replay_seen": 842,
+  "step": 500,
+  "tokens": 9113478,
+  "tok_per_s": 775.9031187478307,
+  "rows_per_s": 1.4338938789175861,
+  "val": {
+   "n": 1200,
+   "dropped_long": 0,
+   "noul_ce": 0.5531377196311951,
+   "noul_T": 0.8911908555798087,
+   "choice_ce": 0.6708876490592957,
+   "choice_T": 1.0220903707073858,
+   "score_ce": 1.0566658973693848,
+   "score_T": 0.9871672059187314,
+   "noul_auc": 0.7480106666666667,
+   "noul_acc": 0.709,
+   "noul_f1": 0.781038374717833,
+   "noul_nll": 0.5666926582697779,
+   "noul_ece": 0.03606485928595067,
+   "noul_brier": 0.1934133365102207,
+   "noul_sharp": 0.18925884366035461,
+   "entropy": 0.5806081295013428
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.237416744232178
+ }
+}

e45/S1-E45-NOISE-ts0-syn/train_log.jsonl

@@ -0,0 +1,57 @@
+{"time": "22:21:45", "event": "start", "data": "data/e45/E45-NOISE", "data_manifest": {"arm": "E45-NOISE", "name": null, "train_sha256": "53ee5956d8073d8dad59f35016d62ec8cf548a2dec30ffbbe807ee3b8851bb89", "n_train": 16000}, "init": "released", "objective": "s-log", "sigma": 0.5, "K": 8, "reward": "brier", "clip": 0.2, "passes": 1, "lr": 0.0001, "lr_head": 0.0002, "rows_per_step": 32, "train_seed": 0, "replay": "data/e16/replay/train.jsonl", "replay_frac": 0.05, "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e", "max_len": 3072, "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps", "total_rows": 16000, "lora": {"r": 16, "lora_alpha": 32, "lora_dropout": 0.05}, "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3", "device": "cuda", "load_s": 75.91583919525146}
+{"time": "22:25:26", "event": "step", "step": 10, "rows_seen": 320, "lr": 7.333333333333333e-05, "tok_per_s": 813.6539683878292, "loss": 0.577488923445344, "entropy": 0.5371629167348146, "sharp": 0.2510472182184458}
+{"time": "22:29:10", "event": "step", "step": 20, "rows_seen": 640, "lr": 9.997508952966197e-05, "tok_per_s": 812.2799692748733, "loss": 0.6161677315831184, "entropy": 0.5316848270595074, "sharp": 0.24935242906212807}
+{"time": "22:32:55", "event": "step", "step": 30, "rows_seen": 960, "lr": 9.977596250541417e-05, "tok_per_s": 814.3714298479673, "loss": 0.6004001908004284, "entropy": 0.581790978088975, "sharp": 0.20492732897400856}
+{"time": "22:36:34", "event": "step", "step": 40, "rows_seen": 1280, "lr": 9.937854366485982e-05, "tok_per_s": 814.9059647678858, "loss": 0.5924058668315411, "entropy": 0.6551112532615662, "sharp": 0.11331460811197758}
+{"time": "22:40:20", "event": "step", "step": 50, "rows_seen": 1600, "lr": 9.8784499920729e-05, "tok_per_s": 816.5747557678841, "loss": 0.49044235423207283, "entropy": 0.48978979513049126, "sharp": 0.2550374250859022}
+{"time": "22:43:58", "event": "step", "step": 60, "rows_seen": 1920, "lr": 9.79963228989294e-05, "tok_per_s": 818.0567462637081, "loss": 0.5913941822946072, "entropy": 0.6587455235421658, "sharp": 0.1154701029881835}
+{"time": "22:47:43", "event": "step", "step": 70, "rows_seen": 2240, "lr": 9.701731848780134e-05, "tok_per_s": 818.0287887770498, "loss": 0.5806628204882145, "entropy": 0.5434654653072357, "sharp": 0.2351485677063465}
+{"time": "22:51:24", "event": "step", "step": 80, "rows_seen": 2560, "lr": 9.585159297207315e-05, "tok_per_s": 817.8545281189062, "loss": 0.6320503326132894, "entropy": 0.6085324361920357, "sharp": 0.17389390245079994}
+{"time": "22:55:12", "event": "step", "step": 90, "rows_seen": 2880, "lr": 9.450403580967612e-05, "tok_per_s": 819.0848385824808, "loss": 0.6345171425491571, "entropy": 0.5344767961651087, "sharp": 0.23385817371308804}
+{"time": "22:58:57", "event": "step", "step": 100, "rows_seen": 3200, "lr": 9.298029912365855e-05, "tok_per_s": 818.6446398429629, "loss": 0.5553547888994217, "entropy": 0.6228579804301262, "sharp": 0.16205373965203762}
+{"time": "23:00:27", "event": "val", "step": 100, "rows_seen": 3200, "n": 500, "dropped_long": 0, "noul_ce": 0.5866313576698303, "noul_T": 0.8289709968072323, "noul_auc": 0.7376753134968187, "noul_acc": 0.692, "noul_f1": 0.7475409836065574, "noul_nll": 0.590714715987444, "noul_ece": 0.0678528763651848, "noul_brier": 0.2025076320438328, "noul_sharp": 0.1656135618686676, "entropy": 0.6120926737785339}
+{"time": "23:04:06", "event": "step", "step": 110, "rows_seen": 3520, "lr": 9.128677399521752e-05, "tok_per_s": 790.351410354805, "loss": 0.6075951606035233, "entropy": 0.5805092714726925, "sharp": 0.19646665640175343}
+{"time": "23:08:00", "event": "step", "step": 120, "rows_seen": 3840, "lr": 8.943056365728335e-05, "tok_per_s": 793.5821388629806, "loss": 0.5906296093016863, "entropy": 0.5608868785202503, "sharp": 0.2055661454796791}
+{"time": "23:11:51", "event": "step", "step": 130, "rows_seen": 4160, "lr": 8.741945370109202e-05, "tok_per_s": 796.4268035701265, "loss": 0.5829800367355347, "entropy": 0.5096677727997303, "sharp": 0.2565692625939846}
+{"time": "23:15:35", "event": "step", "step": 140, "rows_seen": 4480, "lr": 8.526187942070997e-05, "tok_per_s": 797.7773881315642, "loss": 0.5973723791539669, "entropy": 0.6242774985730648, "sharp": 0.1549453902989626}
+{"time": "23:19:22", "event": "step", "step": 150, "rows_seen": 4800, "lr": 8.296689043247915e-05, "tok_per_s": 799.2394784968031, "loss": 0.66122030839324, "entropy": 0.6240631230175495, "sharp": 0.14638944528996944}
+{"time": "23:22:58", "event": "step", "step": 160, "rows_seen": 5120, "lr": 8.054411271778132e-05, "tok_per_s": 800.6844224711881, "loss": 0.6311543304473162, "entropy": 0.5718916598707438, "sharp": 0.19216259568929672}
+{"time": "23:26:43", "event": "step", "step": 170, "rows_seen": 5440, "lr": 7.800370824832674e-05, "tok_per_s": 802.2344113615454, "loss": 0.6034145280718803, "entropy": 0.611348208039999, "sharp": 0.16784686222672462}
+{"time": "23:30:25", "event": "step", "step": 180, "rows_seen": 5760, "lr": 7.535633236331324e-05, "tok_per_s": 803.2766086758899, "loss": 0.5944302603602409, "entropy": 0.6237442083656788, "sharp": 0.15582537651062012}
+{"time": "23:34:08", "event": "step", "step": 190, "rows_seen": 6080, "lr": 7.26130890772308e-05, "tok_per_s": 804.0315913305724, "loss": 0.6191864684224129, "entropy": 0.5512723177671432, "sharp": 0.217588746920228}
+{"time": "23:37:51", "event": "step", "step": 200, "rows_seen": 6400, "lr": 6.978548450576568e-05, "tok_per_s": 804.9704282536054, "loss": 0.5586341023445129, "entropy": 0.5816444866359234, "sharp": 0.21053589694201946}
+{"time": "23:39:21", "event": "val", "step": 200, "rows_seen": 6400, "n": 500, "dropped_long": 0, "noul_ce": 0.5714945197105408, "noul_T": 0.8396485132267894, "noul_auc": 0.7467722063762637, "noul_acc": 0.69, "noul_f1": 0.7683109118086696, "noul_nll": 0.5749939913153649, "noul_ece": 0.06094935488700867, "noul_brier": 0.19602544412931885, "noul_sharp": 0.18119440972805023, "entropy": 0.5985406041145325}
+{"time": "23:43:08", "event": "step", "step": 210, "rows_seen": 6720, "lr": 6.688537860515266e-05, "tok_per_s": 791.176478771693, "loss": 0.5447567291557789, "entropy": 0.5726124010980129, "sharp": 0.21290836110711098}
+{"time": "23:46:46", "event": "step", "step": 220, "rows_seen": 7040, "lr": 6.392493542739706e-05, "tok_per_s": 792.2680080098102, "loss": 0.5979290567338467, "entropy": 0.5627127699553967, "sharp": 0.21210376918315887}
+{"time": "23:50:22", "event": "step", "step": 230, "rows_seen": 7360, "lr": 6.091657210001309e-05, "tok_per_s": 793.4351311439757, "loss": 0.5559693370014429, "entropy": 0.5679895356297493, "sharp": 0.22135603800415993}
+{"time": "23:54:01", "event": "step", "step": 240, "rows_seen": 7680, "lr": 5.7872906744276015e-05, "tok_per_s": 794.4672083944674, "loss": 0.5577775835990906, "entropy": 0.5578128173947334, "sharp": 0.21624882519245148}
+{"time": "23:57:59", "event": "step", "step": 250, "rows_seen": 8000, "lr": 5.48067055504363e-05, "tok_per_s": 795.2338354195869, "loss": 0.5926106758415699, "entropy": 0.5828048177063465, "sharp": 0.19314078241586685}
+{"time": "00:01:27", "event": "step", "step": 260, "rows_seen": 8320, "lr": 5.173082923188119e-05, "tok_per_s": 796.0332942529573, "loss": 0.5824083816260099, "entropy": 0.5969515740871429, "sharp": 0.171417155303061}
+{"time": "00:05:10", "event": "step", "step": 270, "rows_seen": 8640, "lr": 4.865817908283331e-05, "tok_per_s": 796.8198104933426, "loss": 0.6419656090438366, "entropy": 0.6046277731657028, "sharp": 0.16751221287995577}
+{"time": "00:08:50", "event": "step", "step": 280, "rows_seen": 8960, "lr": 4.560164286583922e-05, "tok_per_s": 797.6053623369786, "loss": 0.6085697417147458, "entropy": 0.5505957510322332, "sharp": 0.22415123693645}
+{"time": "00:12:37", "event": "step", "step": 290, "rows_seen": 9280, "lr": 4.257404075601443e-05, "tok_per_s": 798.4635199283907, "loss": 0.6016140468418598, "entropy": 0.5419358648359776, "sharp": 0.2187774581834674}
+{"time": "00:16:21", "event": "step", "step": 300, "rows_seen": 9600, "lr": 3.958807156877371e-05, "tok_per_s": 799.4145723541988, "loss": 0.5394259113818407, "entropy": 0.5766057875007391, "sharp": 0.1751016043126583}
+{"time": "00:17:51", "event": "val", "step": 300, "rows_seen": 9600, "n": 500, "dropped_long": 0, "noul_ce": 0.5705423951148987, "noul_T": 0.7897022029325768, "noul_auc": 0.7468903478422304, "noul_acc": 0.69, "noul_f1": 0.7566718995290423, "noul_nll": 0.57470042078197, "noul_ece": 0.048341548919677735, "noul_brier": 0.1965242978512345, "noul_sharp": 0.16892224550247192, "entropy": 0.6030024886131287}
+{"time": "00:21:26", "event": "step", "step": 310, "rows_seen": 9920, "lr": 3.665625949658603e-05, "tok_per_s": 790.0398472234662, "loss": 0.5552126560360193, "entropy": 0.590318825095892, "sharp": 0.17391379363834858}
+{"time": "00:24:58", "event": "step", "step": 320, "rows_seen": 10240, "lr": 3.379090157815887e-05, "tok_per_s": 790.595842243345, "loss": 0.6531740333884954, "entropy": 0.5938975121825933, "sharp": 0.1760867815464735}
+{"time": "00:28:40", "event": "step", "step": 330, "rows_seen": 10560, "lr": 3.10040161203844e-05, "tok_per_s": 791.5601330906195, "loss": 0.5646311268210411, "entropy": 0.5547943003475666, "sharp": 0.2204857561737299}
+{"time": "00:32:21", "event": "step", "step": 340, "rows_seen": 10880, "lr": 2.8307292289383776e-05, "tok_per_s": 792.3643918584958, "loss": 0.6441864799708128, "entropy": 0.6056978944689035, "sharp": 0.1601609317585826}
+{"time": "00:35:56", "event": "step", "step": 350, "rows_seen": 11200, "lr": 2.571204108208194e-05, "tok_per_s": 792.9287138547262, "loss": 0.6113428547978401, "entropy": 0.6029693968594074, "sharp": 0.16428233589977026}
+{"time": "00:39:35", "event": "step", "step": 360, "rows_seen": 11520, "lr": 2.322914788395595e-05, "tok_per_s": 793.8067028556798, "loss": 0.6038871817290783, "entropy": 0.6176639050245285, "sharp": 0.16306819207966328}
+{"time": "00:43:12", "event": "step", "step": 370, "rows_seen": 11840, "lr": 2.0869026811945157e-05, "tok_per_s": 794.7055557632381, "loss": 0.6174044692888856, "entropy": 0.5644715474918485, "sharp": 0.21021682769060135}
+{"time": "00:47:01", "event": "step", "step": 380, "rows_seen": 12160, "lr": 1.8641577034025627e-05, "tok_per_s": 795.4030562943802, "loss": 0.5908838883042336, "entropy": 0.578949835151434, "sharp": 0.18708872236311436}
+{"time": "00:50:49", "event": "step", "step": 390, "rows_seen": 12480, "lr": 1.6556141248659472e-05, "tok_per_s": 796.3470878647198, "loss": 0.6300652399659157, "entropy": 0.5999844558537006, "sharp": 0.15865095611661673}
+{"time": "00:54:25", "event": "step", "step": 400, "rows_seen": 12800, "lr": 1.4621466498270773e-05, "tok_per_s": 796.6771601277901, "loss": 0.5785091854631901, "entropy": 0.5944705307483673, "sharp": 0.17086115293204784}
+{"time": "00:55:55", "event": "val", "step": 400, "rows_seen": 12800, "n": 500, "dropped_long": 0, "noul_ce": 0.5688287615776062, "noul_T": 0.9422926574527882, "noul_auc": 0.7455907917165954, "noul_acc": 0.702, "noul_f1": 0.7738998482549317, "noul_nll": 0.5742071931585669, "noul_ece": 0.03983503699302675, "noul_brier": 0.19655492965359214, "noul_sharp": 0.18843500316143036, "entropy": 0.5780614018440247}
+{"time": "00:59:36", "event": "step", "step": 410, "rows_seen": 13120, "lr": 1.284566748110963e-05, "tok_per_s": 789.4412731587039, "loss": 0.6104978807270527, "entropy": 0.5868110209703445, "sharp": 0.18706927634775639}
+{"time": "01:03:06", "event": "step", "step": 420, "rows_seen": 13440, "lr": 1.1236192515387821e-05, "tok_per_s": 789.9833813177559, "loss": 0.6568277468904853, "entropy": 0.6150003271177411, "sharp": 0.15597743541002274}
+{"time": "01:06:53", "event": "step", "step": 430, "rows_seen": 13760, "lr": 9.799792298443728e-06, "tok_per_s": 790.6631990684843, "loss": 0.5819298066198826, "entropy": 0.5761610679328442, "sharp": 0.1879867846146226}
+{"time": "01:10:27", "event": "step", "step": 440, "rows_seen": 14080, "lr": 8.542491591971554e-06, "tok_per_s": 791.3829024839538, "loss": 0.5931172147393227, "entropy": 0.5953016579151154, "sharp": 0.17613293882459402}
+{"time": "01:14:18", "event": "step", "step": 450, "rows_seen": 14400, "lr": 7.469563952077155e-06, "tok_per_s": 791.9936980330433, "loss": 0.5529641471803188, "entropy": 0.5828006211668253, "sharp": 0.19747456163167953}
+{"time": "01:17:49", "event": "step", "step": 460, "rows_seen": 14720, "lr": 6.585509610150972e-06, "tok_per_s": 792.5522837484663, "loss": 0.586802514269948, "entropy": 0.5967636238783598, "sharp": 0.18801777809858322}
+{"time": "01:21:22", "event": "step", "step": 470, "rows_seen": 15040, "lr": 5.894036597333646e-06, "tok_per_s": 793.1638712068302, "loss": 0.5895590409636497, "entropy": 0.5700822174549103, "sharp": 0.2148651648312807}
+{"time": "01:25:11", "event": "step", "step": 480, "rows_seen": 15360, "lr": 5.398045191744809e-06, "tok_per_s": 793.8580780830048, "loss": 0.5889317747205496, "entropy": 0.5603284165263176, "sharp": 0.2201318647712469}
+{"time": "01:28:58", "event": "step", "step": 490, "rows_seen": 15680, "lr": 5.099615753708697e-06, "tok_per_s": 794.4824516676765, "loss": 0.6158607490360737, "entropy": 0.5687417462468147, "sharp": 0.19970327615737915}
+{"time": "01:32:49", "event": "step", "step": 500, "rows_seen": 16000, "lr": 5e-06, "tok_per_s": 794.9774856811176, "loss": 0.5487720184028149, "entropy": 0.5344221629202366, "sharp": 0.23927496001124382}
+{"time": "01:34:19", "event": "val", "step": 500, "rows_seen": 16000, "n": 500, "dropped_long": 0, "noul_ce": 0.5672667026519775, "noul_T": 0.9176552540260776, "noul_auc": 0.7497763750822771, "noul_acc": 0.71, "noul_f1": 0.7730829420970265, "noul_nll": 0.5704260485731065, "noul_ece": 0.044484161734580994, "noul_brier": 0.1950136731050071, "noul_sharp": 0.18746744096279144, "entropy": 0.5805526375770569}
+{"time": "01:38:41", "event": "ckpt", "tag": "ckpt-16000", "rows_seen": 16000, "replay_seen": 842, "step": 500, "tokens": 9113478, "tok_per_s": 775.9031187478307, "rows_per_s": 1.4338938789175861, "val": {"n": 1200, "dropped_long": 0, "noul_ce": 0.5531377196311951, "noul_T": 0.8911908555798087, "choice_ce": 0.6708876490592957, "choice_T": 1.0220903707073858, "score_ce": 1.0566658973693848, "score_T": 0.9871672059187314, "noul_auc": 0.7480106666666667, "noul_acc": 0.709, "noul_f1": 0.781038374717833, "noul_nll": 0.5666926582697779, "noul_ece": 0.03606485928595067, "noul_brier": 0.1934133365102207, "noul_sharp": 0.18925884366035461, "entropy": 0.5806081295013428}, "dropped_long": 0, "peak_gib": 18.237416744232178}

e45/S1-E45-REL-ts0-syn/DONE

@@ -0,0 +1 @@
+2026-11-01 20:14:41

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter/README.md

@@ -0,0 +1,206 @@
+---
+base_model: /workspace/jev_kt/models/hf/JEV-9B
+library_name: peft
+tags:
+- base_model:adapter:/workspace/jev_kt/models/hf/JEV-9B
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.2

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter/adapter_config.json

@@ -0,0 +1,57 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": "/workspace/jev_kt/models/hf/JEV-9B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "down_proj",
+    "q_proj",
+    "o_proj",
+    "k_proj",
+    "in_proj_qkv",
+    "up_proj",
+    "in_proj_z",
+    "out_proj",
+    "gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f7b0ac1f4c0e3d168a2531a32e417927d52c0ebe854408e2476239968a5ec93
+size 160486456

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter_vllm/adapter_config.json

@@ -0,0 +1,58 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": null,
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64.0,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "gate_proj",
+    "in_proj_qkv",
+    "in_proj_z",
+    "k_proj",
+    "lm_head",
+    "o_proj",
+    "out_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1f2e506f8718798fa5beb112d2b40c79215ba7fd5142cdb6fa8d7857ccce73c
+size 176641832

e45/S1-E45-REL-ts0-syn/ckpt-16000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.000300432147923857,
+  -0.0003006891056429595,
+  -0.0028248338494449854,
+  -0.001711845281533897,
+  -0.0006357599631883204,
+  0.0017683561891317368,
+  0.0015166114317253232,
+  -0.0021193367429077625,
+  -0.0002721119672060013,
+  -0.00045633752597495914,
+  0.00012549264647532254,
+  0.00036712491419166327,
+  0.0004224546137265861,
+  -0.0037310936022549868,
+  -0.004655673634260893,
+  -0.005180039908736944,
+  -0.005864499602466822,
+  -0.004469653125852346,
+  -0.004952496849000454,
+  -0.005033113062381744,
+  -0.003895251778885722,
+  -0.003449047217145562,
+  -0.005729627329856157,
+  -0.0035173753276467323
+ ],
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "note": "decision logits = lm_head-LoRA logprobs of verbalizer ids + bias, then per-kind temperature"
+}

e45/S1-E45-REL-ts0-syn/ckpt-16000/calibration.json

@@ -0,0 +1,29 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 0.8986131205248705,
+  "choice": 0.9752874633489123,
+  "score": 0.9821101673825277
+ },
+ "fit": {
+  "noul": {
+   "n": 1087,
+   "T": 0.8986131205248705,
+   "ce_before": 0.5520605444908142,
+   "ce_after": 0.5511251091957092
+  },
+  "choice": {
+   "n": 60,
+   "T": 0.9752874633489123,
+   "ce_before": 0.6677042245864868,
+   "ce_after": 0.6675605773925781
+  },
+  "score": {
+   "n": 53,
+   "T": 0.9821101673825277,
+   "ce_before": 1.0668580532073975,
+   "ce_after": 1.0667593479156494
+  }
+ },
+ "source": "models/e45/S1-E45-REL-ts0-syn/ckpt-16000"
+}

e45/S1-E45-REL-ts0-syn/ckpt-16000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:586c28467b1ef73bd4e025e3e7d583501e9161c74110243696bca571f7d771e0
+size 393464

e45/S1-E45-REL-ts0-syn/ckpt-16000/judge_config.json

@@ -0,0 +1,164 @@
+{
+ "base_model_path": "/root/models/Qwen3.5-9B",
+ "hidden_size": 4096,
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "kinds": [
+  "noul",
+  "choice",
+  "score"
+ ],
+ "model_name": "jev-judge-qwen35-9b",
+ "model_version": "0.8.0",
+ "lm_head_included": true,
+ "weights_mode": "unmerged",
+ "adapter_subfolder": "adapter",
+ "lora": {
+  "r": 16,
+  "lora_alpha": 32,
+  "lora_dropout": 0.05,
+  "target_modules": [
+   "down_proj",
+   "gate_proj",
+   "in_proj_qkv",
+   "in_proj_z",
+   "k_proj",
+   "o_proj",
+   "out_proj",
+   "q_proj",
+   "up_proj",
+   "v_proj"
+  ]
+ },
+ "trained_stage": "e16",
+ "source_checkpoint": "models/e45/S1-E45-REL-ts0-syn/ckpt-16000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e45/E45-REL",
+  "data_manifest": {
+   "arm": "E45-REL",
+   "name": null,
+   "train_sha256": "9b28e9361099b0a8af4abb1ca5a7de90044f7952fc15ff52b8b48740865e2d11",
+   "n_train": 16000
+  },
+  "init": "released",
+  "objective": "s-log",
+  "sigma": 0.5,
+  "K": 8,
+  "reward": "brier",
+  "clip": 0.2,
+  "passes": 1,
+  "lr": 0.0001,
+  "lr_head": 0.0002,
+  "rows_per_step": 32,
+  "train_seed": 0,
+  "replay": "data/e16/replay/train.jsonl",
+  "replay_frac": 0.05,
+  "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e",
+  "max_len": 3072,
+  "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps",
+  "total_rows": 16000,
+  "lora": {
+   "r": 16,
+   "lora_alpha": 32,
+   "lora_dropout": 0.05
+  },
+  "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3",
+  "event": "ckpt",
+  "tag": "ckpt-16000",
+  "rows_seen": 16000,
+  "replay_seen": 842,
+  "step": 500,
+  "tokens": 9179126,
+  "tok_per_s": 778.0060240998838,
+  "rows_per_s": 1.4274972858277077,
+  "val": {
+   "n": 1200,
+   "dropped_long": 0,
+   "noul_ce": 0.5520605444908142,
+   "noul_T": 0.8986131205248705,
+   "choice_ce": 0.6677042245864868,
+   "choice_T": 0.9752874633489123,
+   "score_ce": 1.0668580532073975,
+   "score_T": 0.9821101673825277,
+   "noul_auc": 0.7464661333333333,
+   "noul_acc": 0.706,
+   "noul_f1": 0.7805970149253731,
+   "noul_nll": 0.5664093009401113,
+   "noul_ece": 0.026353577703237534,
+   "noul_brier": 0.1935194658173043,
+   "noul_sharp": 0.1905725747346878,
+   "entropy": 0.5781111121177673
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.237416744232178
+ }
+}

e45/S1-E45-REL-ts0-syn/train_log.jsonl

@@ -0,0 +1,57 @@
+{"time": "16:58:01", "event": "start", "data": "data/e45/E45-REL", "data_manifest": {"arm": "E45-REL", "name": null, "train_sha256": "9b28e9361099b0a8af4abb1ca5a7de90044f7952fc15ff52b8b48740865e2d11", "n_train": 16000}, "init": "released", "objective": "s-log", "sigma": 0.5, "K": 8, "reward": "brier", "clip": 0.2, "passes": 1, "lr": 0.0001, "lr_head": 0.0002, "rows_per_step": 32, "train_seed": 0, "replay": "data/e16/replay/train.jsonl", "replay_frac": 0.05, "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e", "max_len": 3072, "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps", "total_rows": 16000, "lora": {"r": 16, "lora_alpha": 32, "lora_dropout": 0.05}, "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3", "device": "cuda", "load_s": 87.95106148719788}
+{"time": "17:01:46", "event": "step", "step": 10, "rows_seen": 320, "lr": 7.333333333333333e-05, "tok_per_s": 807.2548328173016, "loss": 0.5859159016981721, "entropy": 0.5220357663929462, "sharp": 0.2621446754783392}
+{"time": "17:05:31", "event": "step", "step": 20, "rows_seen": 640, "lr": 9.997508952966197e-05, "tok_per_s": 810.3049407408063, "loss": 0.609489593654871, "entropy": 0.5048594251275063, "sharp": 0.27302069775760174}
+{"time": "17:09:18", "event": "step", "step": 30, "rows_seen": 960, "lr": 9.977596250541417e-05, "tok_per_s": 812.0906100819824, "loss": 0.6011982820928097, "entropy": 0.5863940715789795, "sharp": 0.2008225740864873}
+{"time": "17:12:59", "event": "step", "step": 40, "rows_seen": 1280, "lr": 9.937854366485982e-05, "tok_per_s": 813.511917834499, "loss": 0.5325711295008659, "entropy": 0.5941349901258945, "sharp": 0.19071562588214874}
+{"time": "17:16:47", "event": "step", "step": 50, "rows_seen": 1600, "lr": 9.8784499920729e-05, "tok_per_s": 814.9273939064152, "loss": 0.5196209773421288, "entropy": 0.5497458279132843, "sharp": 0.22399818152189255}
+{"time": "17:20:28", "event": "step", "step": 60, "rows_seen": 1920, "lr": 9.79963228989294e-05, "tok_per_s": 816.1146020410687, "loss": 0.5912675000727177, "entropy": 0.6419856585562229, "sharp": 0.13590564485639334}
+{"time": "17:24:14", "event": "step", "step": 70, "rows_seen": 2240, "lr": 9.701731848780134e-05, "tok_per_s": 816.2519339620111, "loss": 0.5597155205905437, "entropy": 0.602167870849371, "sharp": 0.17172034922987223}
+{"time": "17:27:57", "event": "step", "step": 80, "rows_seen": 2560, "lr": 9.585159297207315e-05, "tok_per_s": 816.5848901766232, "loss": 0.6373289618641138, "entropy": 0.5693444423377514, "sharp": 0.21241891384124756}
+{"time": "17:31:46", "event": "step", "step": 90, "rows_seen": 2880, "lr": 9.450403580967612e-05, "tok_per_s": 818.1055981128305, "loss": 0.6420655455440283, "entropy": 0.5909573175013065, "sharp": 0.18392074014991522}
+{"time": "17:35:31", "event": "step", "step": 100, "rows_seen": 3200, "lr": 9.298029912365855e-05, "tok_per_s": 818.0727220302199, "loss": 0.5440826080739498, "entropy": 0.6009539850056171, "sharp": 0.18309033941477537}
+{"time": "17:37:02", "event": "val", "step": 100, "rows_seen": 3200, "n": 500, "dropped_long": 0, "noul_ce": 0.5779311060905457, "noul_T": 0.880844320099016, "noul_auc": 0.7422321986126816, "noul_acc": 0.706, "noul_f1": 0.7655502392344498, "noul_nll": 0.5817958881109953, "noul_ece": 0.05630689781904222, "noul_brier": 0.1990440036433597, "noul_sharp": 0.17626458406448364, "entropy": 0.5965948700904846}
+{"time": "17:40:42", "event": "step", "step": 110, "rows_seen": 3520, "lr": 9.128677399521752e-05, "tok_per_s": 790.1116905729485, "loss": 0.6103495545685291, "entropy": 0.5782968066632748, "sharp": 0.1957132536917925}
+{"time": "17:44:37", "event": "step", "step": 120, "rows_seen": 3840, "lr": 8.943056365728335e-05, "tok_per_s": 793.2841019927674, "loss": 0.586980065330863, "entropy": 0.6020913925021887, "sharp": 0.16804809868335724}
+{"time": "17:48:31", "event": "step", "step": 130, "rows_seen": 4160, "lr": 8.741945370109202e-05, "tok_per_s": 796.016064841255, "loss": 0.5760547667741776, "entropy": 0.5098970495164394, "sharp": 0.2505270354449749}
+{"time": "17:52:16", "event": "step", "step": 140, "rows_seen": 4480, "lr": 8.526187942070997e-05, "tok_per_s": 797.5907036110415, "loss": 0.588726045563817, "entropy": 0.5915715508162975, "sharp": 0.18919290602207184}
+{"time": "17:56:04", "event": "step", "step": 150, "rows_seen": 4800, "lr": 8.296689043247915e-05, "tok_per_s": 799.081284836039, "loss": 0.6570097580552101, "entropy": 0.6380404345691204, "sharp": 0.14082889910787344}
+{"time": "17:59:43", "event": "step", "step": 160, "rows_seen": 5120, "lr": 8.054411271778132e-05, "tok_per_s": 800.4618488199133, "loss": 0.6297760903835297, "entropy": 0.5893573388457298, "sharp": 0.17969845049083233}
+{"time": "18:03:30", "event": "step", "step": 170, "rows_seen": 5440, "lr": 7.800370824832674e-05, "tok_per_s": 801.7603733005637, "loss": 0.5962249897420406, "entropy": 0.5910500138998032, "sharp": 0.1914851674810052}
+{"time": "18:07:13", "event": "step", "step": 180, "rows_seen": 5760, "lr": 7.535633236331324e-05, "tok_per_s": 803.0116347975378, "loss": 0.5863769315183163, "entropy": 0.617675743997097, "sharp": 0.15867283195257187}
+{"time": "18:10:58", "event": "step", "step": 190, "rows_seen": 6080, "lr": 7.26130890772308e-05, "tok_per_s": 803.684088102376, "loss": 0.6129954941570759, "entropy": 0.5540583729743958, "sharp": 0.2156384475529194}
+{"time": "18:14:42", "event": "step", "step": 200, "rows_seen": 6400, "lr": 6.978548450576568e-05, "tok_per_s": 804.754047443891, "loss": 0.5601796135306358, "entropy": 0.573649849742651, "sharp": 0.21057562343776226}
+{"time": "18:16:12", "event": "val", "step": 200, "rows_seen": 6400, "n": 500, "dropped_long": 0, "noul_ce": 0.5716017484664917, "noul_T": 0.892537978095737, "noul_auc": 0.7445106411706132, "noul_acc": 0.678, "noul_f1": 0.7670043415340086, "noul_nll": 0.5777599147334694, "noul_ece": 0.055663657426834115, "noul_brier": 0.19777273337476287, "noul_sharp": 0.18703073263168335, "entropy": 0.5899770259857178}
+{"time": "18:20:00", "event": "step", "step": 210, "rows_seen": 6720, "lr": 6.688537860515266e-05, "tok_per_s": 791.0806056266221, "loss": 0.5494696125388145, "entropy": 0.5707839652895927, "sharp": 0.20891448017209768}
+{"time": "18:23:39", "event": "step", "step": 220, "rows_seen": 7040, "lr": 6.392493542739706e-05, "tok_per_s": 792.4182634401858, "loss": 0.600819818675518, "entropy": 0.5623787865042686, "sharp": 0.2135008368641138}
+{"time": "18:27:15", "event": "step", "step": 230, "rows_seen": 7360, "lr": 6.091657210001309e-05, "tok_per_s": 793.5993626890156, "loss": 0.5575203914195299, "entropy": 0.564382541924715, "sharp": 0.22327942494302988}
+{"time": "18:30:56", "event": "step", "step": 240, "rows_seen": 7680, "lr": 5.7872906744276015e-05, "tok_per_s": 794.7269218728223, "loss": 0.564826387912035, "entropy": 0.5365338232368231, "sharp": 0.2335683573037386}
+{"time": "18:34:54", "event": "step", "step": 250, "rows_seen": 8000, "lr": 5.48067055504363e-05, "tok_per_s": 795.6100957928961, "loss": 0.5920643508434296, "entropy": 0.5942213386297226, "sharp": 0.18172888085246086}
+{"time": "18:38:23", "event": "step", "step": 260, "rows_seen": 8320, "lr": 5.173082923188119e-05, "tok_per_s": 796.5723098643724, "loss": 0.5873812437057495, "entropy": 0.5533143579959869, "sharp": 0.20813973061740398}
+{"time": "18:42:06", "event": "step", "step": 270, "rows_seen": 8640, "lr": 4.865817908283331e-05, "tok_per_s": 797.4336521718462, "loss": 0.6422621437814087, "entropy": 0.6008757147938013, "sharp": 0.1711443830281496}
+{"time": "18:45:48", "event": "step", "step": 280, "rows_seen": 8960, "lr": 4.560164286583922e-05, "tok_per_s": 798.2961361537441, "loss": 0.605023676995188, "entropy": 0.5507638463750482, "sharp": 0.22286107577383518}
+{"time": "18:49:35", "event": "step", "step": 290, "rows_seen": 9280, "lr": 4.257404075601443e-05, "tok_per_s": 799.2073680368666, "loss": 0.6119903288781643, "entropy": 0.5475618652999401, "sharp": 0.21826686710119247}
+{"time": "18:53:21", "event": "step", "step": 300, "rows_seen": 9600, "lr": 3.958807156877371e-05, "tok_per_s": 800.1413526371964, "loss": 0.5355819761753082, "entropy": 0.5809243451803923, "sharp": 0.16378754749894142}
+{"time": "18:54:51", "event": "val", "step": 300, "rows_seen": 9600, "n": 500, "dropped_long": 0, "noul_ce": 0.5724663138389587, "noul_T": 0.7306648195462849, "noul_auc": 0.745303876727819, "noul_acc": 0.694, "noul_f1": 0.7598116169544741, "noul_nll": 0.5765132632516324, "noul_ece": 0.05441170465946198, "noul_brier": 0.19739392816312815, "noul_sharp": 0.1575111746788025, "entropy": 0.6112444400787354}
+{"time": "18:58:28", "event": "step", "step": 310, "rows_seen": 9920, "lr": 3.665625949658603e-05, "tok_per_s": 790.7888034789369, "loss": 0.5504805911332369, "entropy": 0.5683636218309402, "sharp": 0.1921995673328638}
+{"time": "19:01:59", "event": "step", "step": 320, "rows_seen": 10240, "lr": 3.379090157815887e-05, "tok_per_s": 791.5357089546385, "loss": 0.6405128911137581, "entropy": 0.6108909472823143, "sharp": 0.16367854736745358}
+{"time": "19:05:40", "event": "step", "step": 330, "rows_seen": 10560, "lr": 3.10040161203844e-05, "tok_per_s": 792.7184778788117, "loss": 0.5662273876369, "entropy": 0.5532629452645779, "sharp": 0.22433722764253616}
+{"time": "19:09:22", "event": "step", "step": 340, "rows_seen": 10880, "lr": 2.8307292289383776e-05, "tok_per_s": 793.6153484783152, "loss": 0.6378334686160088, "entropy": 0.6033659428358078, "sharp": 0.16770764626562595}
+{"time": "19:12:55", "event": "step", "step": 350, "rows_seen": 11200, "lr": 2.571204108208194e-05, "tok_per_s": 794.3910734803789, "loss": 0.6115953102707863, "entropy": 0.5891885012388229, "sharp": 0.17483517061918974}
+{"time": "19:16:36", "event": "step", "step": 360, "rows_seen": 11520, "lr": 2.322914788395595e-05, "tok_per_s": 795.2063749867742, "loss": 0.605057317763567, "entropy": 0.6163841933012009, "sharp": 0.165567209944129}
+{"time": "19:20:15", "event": "step", "step": 370, "rows_seen": 11840, "lr": 2.0869026811945157e-05, "tok_per_s": 796.1282829075584, "loss": 0.6140534365549684, "entropy": 0.5582990748807788, "sharp": 0.21383531764149666}
+{"time": "19:24:05", "event": "step", "step": 380, "rows_seen": 12160, "lr": 1.8641577034025627e-05, "tok_per_s": 796.7956926983873, "loss": 0.5914619974792004, "entropy": 0.5759841613471508, "sharp": 0.18765570037066936}
+{"time": "19:27:55", "event": "step", "step": 390, "rows_seen": 12480, "lr": 1.6556141248659472e-05, "tok_per_s": 797.7675759292986, "loss": 0.6305775959044695, "entropy": 0.598133459687233, "sharp": 0.16215836256742477}
+{"time": "19:31:30", "event": "step", "step": 400, "rows_seen": 12800, "lr": 1.4621466498270773e-05, "tok_per_s": 798.2436605927263, "loss": 0.5794996656477451, "entropy": 0.5955066867172718, "sharp": 0.16796090174466372}
+{"time": "19:33:00", "event": "val", "step": 400, "rows_seen": 12800, "n": 500, "dropped_long": 0, "noul_ce": 0.5678415894508362, "noul_T": 0.9189865268481029, "noul_auc": 0.7460296028759008, "noul_acc": 0.704, "noul_f1": 0.7757575757575758, "noul_nll": 0.572774373812601, "noul_ece": 0.051562166929245, "noul_brier": 0.1962206297461335, "noul_sharp": 0.18654365837574005, "entropy": 0.5807515382766724}
+{"time": "19:36:39", "event": "step", "step": 410, "rows_seen": 13120, "lr": 1.284566748110963e-05, "tok_per_s": 791.2294968832349, "loss": 0.5959222391247749, "entropy": 0.5914434641599655, "sharp": 0.18311581946909428}
+{"time": "19:40:11", "event": "step", "step": 420, "rows_seen": 13440, "lr": 1.1236192515387821e-05, "tok_per_s": 791.7543976296596, "loss": 0.6583403283730149, "entropy": 0.6149237090721726, "sharp": 0.1567009538412094}
+{"time": "19:43:58", "event": "step", "step": 430, "rows_seen": 13760, "lr": 9.799792298443728e-06, "tok_per_s": 792.4833429908712, "loss": 0.5807589301839471, "entropy": 0.5776796638965607, "sharp": 0.18376952782273293}
+{"time": "19:47:34", "event": "step", "step": 440, "rows_seen": 14080, "lr": 8.542491591971554e-06, "tok_per_s": 793.2114579870419, "loss": 0.5974132008850574, "entropy": 0.5908592492341995, "sharp": 0.17492880299687386}
+{"time": "19:51:25", "event": "step", "step": 450, "rows_seen": 14400, "lr": 7.469563952077155e-06, "tok_per_s": 793.8329637937026, "loss": 0.5518267657607794, "entropy": 0.5826755538582802, "sharp": 0.19769825227558613}
+{"time": "19:54:57", "event": "step", "step": 460, "rows_seen": 14720, "lr": 6.585509610150972e-06, "tok_per_s": 794.4148421853414, "loss": 0.5820136964321136, "entropy": 0.5885909758508205, "sharp": 0.19301331602036953}
+{"time": "19:58:30", "event": "step", "step": 470, "rows_seen": 15040, "lr": 5.894036597333646e-06, "tok_per_s": 795.0769700161079, "loss": 0.5867369882762432, "entropy": 0.5726926028728485, "sharp": 0.21091781370341778}
+{"time": "20:02:21", "event": "step", "step": 480, "rows_seen": 15360, "lr": 5.398045191744809e-06, "tok_per_s": 795.7072370840369, "loss": 0.5849715080112219, "entropy": 0.5596798043698072, "sharp": 0.217598307877779}
+{"time": "20:06:09", "event": "step", "step": 490, "rows_seen": 15680, "lr": 5.099615753708697e-06, "tok_per_s": 796.336443990316, "loss": 0.6060212701559067, "entropy": 0.5702655911445618, "sharp": 0.20112621411681175}
+{"time": "20:09:58", "event": "step", "step": 500, "rows_seen": 16000, "lr": 5e-06, "tok_per_s": 797.0595319298499, "loss": 0.548908244818449, "entropy": 0.5237538851797581, "sharp": 0.24975336156785488}
+{"time": "20:11:28", "event": "val", "step": 500, "rows_seen": 16000, "n": 500, "dropped_long": 0, "noul_ce": 0.5660662651062012, "noul_T": 0.9170049092020095, "noul_auc": 0.7498607618436821, "noul_acc": 0.698, "noul_f1": 0.7666151468315301, "noul_nll": 0.5690367848649621, "noul_ece": 0.03417138195037843, "noul_brier": 0.19473311058885026, "noul_sharp": 0.18767905235290527, "entropy": 0.5794568061828613}
+{"time": "20:14:41", "event": "ckpt", "tag": "ckpt-16000", "rows_seen": 16000, "replay_seen": 842, "step": 500, "tokens": 9179126, "tok_per_s": 778.0060240998838, "rows_per_s": 1.4274972858277077, "val": {"n": 1200, "dropped_long": 0, "noul_ce": 0.5520605444908142, "noul_T": 0.8986131205248705, "choice_ce": 0.6677042245864868, "choice_T": 0.9752874633489123, "score_ce": 1.0668580532073975, "score_T": 0.9821101673825277, "noul_auc": 0.7464661333333333, "noul_acc": 0.706, "noul_f1": 0.7805970149253731, "noul_nll": 0.5664093009401113, "noul_ece": 0.026353577703237534, "noul_brier": 0.1935194658173043, "noul_sharp": 0.1905725747346878, "entropy": 0.5781111121177673}, "dropped_long": 0, "peak_gib": 18.237416744232178}

e45/S1-E45-SHR-ts0-syn/DONE

@@ -0,0 +1 @@
+2026-11-01 14:56:35

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter/README.md

@@ -0,0 +1,206 @@
+---
+base_model: /workspace/jev_kt/models/hf/JEV-9B
+library_name: peft
+tags:
+- base_model:adapter:/workspace/jev_kt/models/hf/JEV-9B
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.2

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter/adapter_config.json

@@ -0,0 +1,57 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": "/workspace/jev_kt/models/hf/JEV-9B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "q_proj",
+    "o_proj",
+    "down_proj",
+    "in_proj_z",
+    "gate_proj",
+    "out_proj",
+    "in_proj_qkv",
+    "k_proj",
+    "up_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:368bf46a30363310c7fab7fe435292781edfe7b76a0ba184633be79ab6ab390b
+size 160486456

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter_vllm/adapter_config.json

@@ -0,0 +1,58 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": null,
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64.0,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "gate_proj",
+    "in_proj_qkv",
+    "in_proj_z",
+    "k_proj",
+    "lm_head",
+    "o_proj",
+    "out_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09258e383f2160c81daf5a0ddd46da64237e2b2e7e617e787526007f00ecc306
+size 176641832

e45/S1-E45-SHR-ts0-syn/ckpt-16000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.0003529029490891844,
+  -0.0003531621186994016,
+  -0.00257667968980968,
+  -0.0013844354543834925,
+  -0.0005795973120257258,
+  0.0018084368202835321,
+  0.0012453077360987663,
+  -0.002353015122935176,
+  -0.00039574294351041317,
+  -0.0004118162614759058,
+  -4.896678728982806e-05,
+  0.0006681560771539807,
+  -0.0009575835429131985,
+  -0.0036021440755575895,
+  -0.004302427638322115,
+  -0.005113600287586451,
+  -0.005915188696235418,
+  -0.004388121422380209,
+  -0.004948095418512821,
+  -0.005016533192247152,
+  -0.003963640425354242,
+  -0.0034523631911724806,
+  -0.005643394310027361,
+  -0.0035082357935607433
+ ],
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "note": "decision logits = lm_head-LoRA logprobs of verbalizer ids + bias, then per-kind temperature"
+}

e45/S1-E45-SHR-ts0-syn/ckpt-16000/calibration.json

@@ -0,0 +1,29 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 0.9000755854451329,
+  "choice": 1.008406819192068,
+  "score": 0.9500376826602371
+ },
+ "fit": {
+  "noul": {
+   "n": 1087,
+   "T": 0.9000755854451329,
+   "ce_before": 0.5519391894340515,
+   "ce_after": 0.5510237812995911
+  },
+  "choice": {
+   "n": 60,
+   "T": 1.008406819192068,
+   "ce_before": 0.6665812730789185,
+   "ce_after": 0.6665658354759216
+  },
+  "score": {
+   "n": 53,
+   "T": 0.9500376826602371,
+   "ce_before": 1.0586308240890503,
+   "ce_after": 1.057839274406433
+  }
+ },
+ "source": "models/e45/S1-E45-SHR-ts0-syn/ckpt-16000"
+}

e45/S1-E45-SHR-ts0-syn/ckpt-16000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8a1c5e546d378f027a69a13f4e1aa7d02b9c409613ecb1ba9fb82ef2b62a061
+size 393464

e45/S1-E45-SHR-ts0-syn/ckpt-16000/judge_config.json

@@ -0,0 +1,164 @@
+{
+ "base_model_path": "/root/models/Qwen3.5-9B",
+ "hidden_size": 4096,
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "kinds": [
+  "noul",
+  "choice",
+  "score"
+ ],
+ "model_name": "jev-judge-qwen35-9b",
+ "model_version": "0.8.0",
+ "lm_head_included": true,
+ "weights_mode": "unmerged",
+ "adapter_subfolder": "adapter",
+ "lora": {
+  "r": 16,
+  "lora_alpha": 32,
+  "lora_dropout": 0.05,
+  "target_modules": [
+   "down_proj",
+   "gate_proj",
+   "in_proj_qkv",
+   "in_proj_z",
+   "k_proj",
+   "o_proj",
+   "out_proj",
+   "q_proj",
+   "up_proj",
+   "v_proj"
+  ]
+ },
+ "trained_stage": "e16",
+ "source_checkpoint": "models/e45/S1-E45-SHR-ts0-syn/ckpt-16000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e45/E45-SHR",
+  "data_manifest": {
+   "arm": "E45-SHR",
+   "name": null,
+   "train_sha256": "cfe452b00b8ac702c42580427823a8e4bf112271713c1fea190e1f3214bb6fbf",
+   "n_train": 16000
+  },
+  "init": "released",
+  "objective": "s-log",
+  "sigma": 0.5,
+  "K": 8,
+  "reward": "brier",
+  "clip": 0.2,
+  "passes": 1,
+  "lr": 0.0001,
+  "lr_head": 0.0002,
+  "rows_per_step": 32,
+  "train_seed": 0,
+  "replay": "data/e16/replay/train.jsonl",
+  "replay_frac": 0.05,
+  "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e",
+  "max_len": 3072,
+  "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps",
+  "total_rows": 16000,
+  "lora": {
+   "r": 16,
+   "lora_alpha": 32,
+   "lora_dropout": 0.05
+  },
+  "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3",
+  "event": "ckpt",
+  "tag": "ckpt-16000",
+  "rows_seen": 16000,
+  "replay_seen": 842,
+  "step": 500,
+  "tokens": 9113125,
+  "tok_per_s": 773.7348806480621,
+  "rows_per_s": 1.4299422923548444,
+  "val": {
+   "n": 1200,
+   "dropped_long": 0,
+   "noul_ce": 0.5519391894340515,
+   "noul_T": 0.9000755854451329,
+   "choice_ce": 0.6665812730789185,
+   "choice_T": 1.008406819192068,
+   "score_ce": 1.0586308240890503,
+   "score_T": 0.9500376826602371,
+   "noul_auc": 0.7466709333333332,
+   "noul_acc": 0.702,
+   "noul_f1": 0.7776119402985074,
+   "noul_nll": 0.5664786969861015,
+   "noul_ece": 0.03434245572984219,
+   "noul_brier": 0.19362598993408342,
+   "noul_sharp": 0.19133822619915009,
+   "entropy": 0.5782209038734436
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.237416744232178
+ }
+}

e45/S1-E45-SHR-ts0-syn/train_log.jsonl

@@ -0,0 +1,57 @@
+{"time": "11:40:16", "event": "start", "data": "data/e45/E45-SHR", "data_manifest": {"arm": "E45-SHR", "name": null, "train_sha256": "cfe452b00b8ac702c42580427823a8e4bf112271713c1fea190e1f3214bb6fbf", "n_train": 16000}, "init": "released", "objective": "s-log", "sigma": 0.5, "K": 8, "reward": "brier", "clip": 0.2, "passes": 1, "lr": 0.0001, "lr_head": 0.0002, "rows_per_step": 32, "train_seed": 0, "replay": "data/e16/replay/train.jsonl", "replay_frac": 0.05, "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e", "max_len": 3072, "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps", "total_rows": 16000, "lora": {"r": 16, "lora_alpha": 32, "lora_dropout": 0.05}, "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3", "device": "cuda", "load_s": 76.93250226974487}
+{"time": "11:43:58", "event": "step", "step": 10, "rows_seen": 320, "lr": 7.333333333333333e-05, "tok_per_s": 809.0945990137965, "loss": 0.5829713176935911, "entropy": 0.5173479039222002, "sharp": 0.26559402979910374}
+{"time": "11:47:42", "event": "step", "step": 20, "rows_seen": 640, "lr": 9.997508952966197e-05, "tok_per_s": 810.9087614989214, "loss": 0.6111677531152964, "entropy": 0.5181166604161263, "sharp": 0.2590177971869707}
+{"time": "11:51:27", "event": "step", "step": 30, "rows_seen": 960, "lr": 9.977596250541417e-05, "tok_per_s": 813.5634272749753, "loss": 0.6006867382675409, "entropy": 0.6142154410481453, "sharp": 0.1746037481352687}
+{"time": "11:55:04", "event": "step", "step": 40, "rows_seen": 1280, "lr": 9.937854366485982e-05, "tok_per_s": 815.2677472337699, "loss": 0.5579716823995113, "entropy": 0.6279277540743351, "sharp": 0.15950517542660236}
+{"time": "11:58:51", "event": "step", "step": 50, "rows_seen": 1600, "lr": 9.8784499920729e-05, "tok_per_s": 817.0839966369149, "loss": 0.5131936836987734, "entropy": 0.5638627298176289, "sharp": 0.21227932255715132}
+{"time": "12:02:29", "event": "step", "step": 60, "rows_seen": 1920, "lr": 9.79963228989294e-05, "tok_per_s": 817.7965695729316, "loss": 0.5988873094320297, "entropy": 0.6518698707222939, "sharp": 0.12110545672476292}
+{"time": "12:06:14", "event": "step", "step": 70, "rows_seen": 2240, "lr": 9.701731848780134e-05, "tok_per_s": 817.5753349578803, "loss": 0.5667485781013966, "entropy": 0.5785291120409966, "sharp": 0.1924605956301093}
+{"time": "12:09:56", "event": "step", "step": 80, "rows_seen": 2560, "lr": 9.585159297207315e-05, "tok_per_s": 817.4368123882094, "loss": 0.6338031617924571, "entropy": 0.5941549204289913, "sharp": 0.18656211905181408}
+{"time": "12:13:45", "event": "step", "step": 90, "rows_seen": 2880, "lr": 9.450403580967612e-05, "tok_per_s": 818.5478930596403, "loss": 0.6401414461433887, "entropy": 0.5672495067119598, "sharp": 0.2097081197425723}
+{"time": "12:17:29", "event": "step", "step": 100, "rows_seen": 3200, "lr": 9.298029912365855e-05, "tok_per_s": 818.1722926734245, "loss": 0.5522025488317013, "entropy": 0.6230933591723442, "sharp": 0.1596109215170145}
+{"time": "12:18:59", "event": "val", "step": 100, "rows_seen": 3200, "n": 500, "dropped_long": 0, "noul_ce": 0.5826035141944885, "noul_T": 0.7816198731038794, "noul_auc": 0.7423840947832103, "noul_acc": 0.688, "noul_f1": 0.7507987220447284, "noul_nll": 0.5855747209638357, "noul_ece": 0.05609731382131578, "noul_brier": 0.20021346276631757, "noul_sharp": 0.16129377484321594, "entropy": 0.6149895787239075}
+{"time": "12:22:40", "event": "step", "step": 110, "rows_seen": 3520, "lr": 9.128677399521752e-05, "tok_per_s": 789.6425450756706, "loss": 0.6145443432033062, "entropy": 0.5715678967535496, "sharp": 0.20603726338595152}
+{"time": "12:26:34", "event": "step", "step": 120, "rows_seen": 3840, "lr": 8.943056365728335e-05, "tok_per_s": 792.6436079760787, "loss": 0.590669309720397, "entropy": 0.5705469585955143, "sharp": 0.1920371688902378}
+{"time": "12:30:26", "event": "step", "step": 130, "rows_seen": 4160, "lr": 8.741945370109202e-05, "tok_per_s": 795.4552149661098, "loss": 0.5670341663062572, "entropy": 0.501390814781189, "sharp": 0.24504408054053783}
+{"time": "12:34:10", "event": "step", "step": 140, "rows_seen": 4480, "lr": 8.526187942070997e-05, "tok_per_s": 796.8098795780792, "loss": 0.5941428132355213, "entropy": 0.619270958006382, "sharp": 0.16091445740312338}
+{"time": "12:37:57", "event": "step", "step": 150, "rows_seen": 4800, "lr": 8.296689043247915e-05, "tok_per_s": 798.2728472901189, "loss": 0.6544649824500084, "entropy": 0.6332171596586704, "sharp": 0.13141057640314102}
+{"time": "12:41:35", "event": "step", "step": 160, "rows_seen": 5120, "lr": 8.054411271778132e-05, "tok_per_s": 799.4032637590615, "loss": 0.6290031746029854, "entropy": 0.5619909763336182, "sharp": 0.19808954559266567}
+{"time": "12:45:21", "event": "step", "step": 170, "rows_seen": 5440, "lr": 7.800370824832674e-05, "tok_per_s": 800.6626301898854, "loss": 0.5942814461886883, "entropy": 0.6056827045977116, "sharp": 0.17220504395663738}
+{"time": "12:49:05", "event": "step", "step": 180, "rows_seen": 5760, "lr": 7.535633236331324e-05, "tok_per_s": 801.5231848211555, "loss": 0.5921114534139633, "entropy": 0.6296906433999538, "sharp": 0.1460996400564909}
+{"time": "12:52:48", "event": "step", "step": 190, "rows_seen": 6080, "lr": 7.26130890772308e-05, "tok_per_s": 802.2583552546262, "loss": 0.6204224918037653, "entropy": 0.5443447083234787, "sharp": 0.22998007852584124}
+{"time": "12:56:32", "event": "step", "step": 200, "rows_seen": 6400, "lr": 6.978548450576568e-05, "tok_per_s": 803.2369595244545, "loss": 0.554803317412734, "entropy": 0.5809122696518898, "sharp": 0.2053777789697051}
+{"time": "12:58:01", "event": "val", "step": 200, "rows_seen": 6400, "n": 500, "dropped_long": 0, "noul_ce": 0.5703958868980408, "noul_T": 0.8501770174936891, "noul_auc": 0.740713236907394, "noul_acc": 0.698, "noul_f1": 0.7756315007429421, "noul_nll": 0.5791702248603106, "noul_ece": 0.0653765162229538, "noul_brier": 0.1982222910803369, "noul_sharp": 0.18270958960056305, "entropy": 0.5957298874855042}
+{"time": "13:01:48", "event": "step", "step": 210, "rows_seen": 6720, "lr": 6.688537860515266e-05, "tok_per_s": 789.6498548928365, "loss": 0.5476706922054291, "entropy": 0.5626404620707035, "sharp": 0.2199501683935523}
+{"time": "13:05:27", "event": "step", "step": 220, "rows_seen": 7040, "lr": 6.392493542739706e-05, "tok_per_s": 790.7037530935604, "loss": 0.5957701317965984, "entropy": 0.5678802914917469, "sharp": 0.21464306116104126}
+{"time": "13:09:04", "event": "step", "step": 230, "rows_seen": 7360, "lr": 6.091657210001309e-05, "tok_per_s": 791.7659169678012, "loss": 0.5514545273035765, "entropy": 0.5611999426037073, "sharp": 0.22715231776237488}
+{"time": "13:12:44", "event": "step", "step": 240, "rows_seen": 7680, "lr": 5.7872906744276015e-05, "tok_per_s": 792.8021717304364, "loss": 0.5595566537231207, "entropy": 0.5545004373416305, "sharp": 0.22167368419468403}
+{"time": "13:16:42", "event": "step", "step": 250, "rows_seen": 8000, "lr": 5.48067055504363e-05, "tok_per_s": 793.4898108471275, "loss": 0.5935519970953465, "entropy": 0.5794208496809006, "sharp": 0.1964077139273286}
+{"time": "13:20:11", "event": "step", "step": 260, "rows_seen": 8320, "lr": 5.173082923188119e-05, "tok_per_s": 794.2774212177541, "loss": 0.5756309572607279, "entropy": 0.5848041549324989, "sharp": 0.18015472125262022}
+{"time": "13:23:55", "event": "step", "step": 270, "rows_seen": 8640, "lr": 4.865817908283331e-05, "tok_per_s": 795.002937213986, "loss": 0.6355277097318321, "entropy": 0.6157853128388524, "sharp": 0.15520272217690945}
+{"time": "13:27:36", "event": "step", "step": 280, "rows_seen": 8960, "lr": 4.560164286583922e-05, "tok_per_s": 795.7466667938447, "loss": 0.6049750228412449, "entropy": 0.5498039564117789, "sharp": 0.22492669895291328}
+{"time": "13:31:23", "event": "step", "step": 290, "rows_seen": 9280, "lr": 4.257404075601443e-05, "tok_per_s": 796.6314361984133, "loss": 0.5982255786657333, "entropy": 0.541251253336668, "sharp": 0.22079676389694214}
+{"time": "13:35:09", "event": "step", "step": 300, "rows_seen": 9600, "lr": 3.958807156877371e-05, "tok_per_s": 797.4425287109306, "loss": 0.5330794285982847, "entropy": 0.5843064188957214, "sharp": 0.16263543535023928}
+{"time": "13:36:39", "event": "val", "step": 300, "rows_seen": 9600, "n": 500, "dropped_long": 0, "noul_ce": 0.5733832716941833, "noul_T": 0.7454526687884386, "noul_auc": 0.743278594454102, "noul_acc": 0.698, "noul_f1": 0.7636932707355243, "noul_nll": 0.578485303029418, "noul_ece": 0.06693672430515289, "noul_brier": 0.1981588911471401, "noul_sharp": 0.15993563830852509, "entropy": 0.6108042597770691}
+{"time": "13:40:16", "event": "step", "step": 310, "rows_seen": 9920, "lr": 3.665625949658603e-05, "tok_per_s": 787.9661988848106, "loss": 0.5537699200212955, "entropy": 0.5826780945062637, "sharp": 0.18228471465408802}
+{"time": "13:43:47", "event": "step", "step": 320, "rows_seen": 10240, "lr": 3.379090157815887e-05, "tok_per_s": 788.627180110971, "loss": 0.6462431238032877, "entropy": 0.6098096650093794, "sharp": 0.16495265625417233}
+{"time": "13:47:30", "event": "step", "step": 330, "rows_seen": 10560, "lr": 3.10040161203844e-05, "tok_per_s": 789.5018999997784, "loss": 0.5649902373552322, "entropy": 0.5519063659012318, "sharp": 0.22365903109312057}
+{"time": "13:51:12", "event": "step", "step": 340, "rows_seen": 10880, "lr": 2.8307292289383776e-05, "tok_per_s": 790.3652673961274, "loss": 0.6401955969631672, "entropy": 0.5960954874753952, "sharp": 0.17324916925281286}
+{"time": "13:54:46", "event": "step", "step": 350, "rows_seen": 11200, "lr": 2.571204108208194e-05, "tok_per_s": 791.0146264644422, "loss": 0.6109335124492645, "entropy": 0.5936173722147942, "sharp": 0.17193381674587727}
+{"time": "13:58:26", "event": "step", "step": 360, "rows_seen": 11520, "lr": 2.322914788395595e-05, "tok_per_s": 791.8545855301425, "loss": 0.6058627925813198, "entropy": 0.6152241826057434, "sharp": 0.16478468663990498}
+{"time": "14:02:04", "event": "step", "step": 370, "rows_seen": 11840, "lr": 2.0869026811945157e-05, "tok_per_s": 792.7552212187459, "loss": 0.614081984385848, "entropy": 0.5671398248523474, "sharp": 0.21154329925775528}
+{"time": "14:05:54", "event": "step", "step": 380, "rows_seen": 12160, "lr": 1.8641577034025627e-05, "tok_per_s": 793.3586737538985, "loss": 0.5948928464204073, "entropy": 0.5795652717351913, "sharp": 0.18553717993199825}
+{"time": "14:09:44", "event": "step", "step": 390, "rows_seen": 12480, "lr": 1.6556141248659472e-05, "tok_per_s": 794.2201699575102, "loss": 0.6317391050979495, "entropy": 0.598414683714509, "sharp": 0.16268220730125904}
+{"time": "14:13:20", "event": "step", "step": 400, "rows_seen": 12800, "lr": 1.4621466498270773e-05, "tok_per_s": 794.5683980253532, "loss": 0.5768286697566509, "entropy": 0.5930426940321922, "sharp": 0.17356200516223907}
+{"time": "14:14:50", "event": "val", "step": 400, "rows_seen": 12800, "n": 500, "dropped_long": 0, "noul_ce": 0.5673655271530151, "noul_T": 0.9480546968416432, "noul_auc": 0.7434642453291926, "noul_acc": 0.688, "noul_f1": 0.7657657657657657, "noul_nll": 0.5742326405495405, "noul_ece": 0.04802528607845308, "noul_brier": 0.1969024044161867, "noul_sharp": 0.19254156947135925, "entropy": 0.5759285688400269}
+{"time": "14:18:30", "event": "step", "step": 410, "rows_seen": 13120, "lr": 1.284566748110963e-05, "tok_per_s": 787.4709065010605, "loss": 0.5981672033667564, "entropy": 0.5809904076159, "sharp": 0.19123374857008457}
+{"time": "14:22:01", "event": "step", "step": 420, "rows_seen": 13440, "lr": 1.1236192515387821e-05, "tok_per_s": 787.979423968617, "loss": 0.6540783168748021, "entropy": 0.616896815598011, "sharp": 0.15546196326613426}
+{"time": "14:25:50", "event": "step", "step": 430, "rows_seen": 13760, "lr": 9.799792298443728e-06, "tok_per_s": 788.5191419822469, "loss": 0.5748474206775427, "entropy": 0.5760907102376223, "sharp": 0.18339864537119865}
+{"time": "14:29:26", "event": "step", "step": 440, "rows_seen": 14080, "lr": 8.542491591971554e-06, "tok_per_s": 789.2027216771044, "loss": 0.5930272117257118, "entropy": 0.5850048363208771, "sharp": 0.18051939085125923}
+{"time": "14:33:18", "event": "step", "step": 450, "rows_seen": 14400, "lr": 7.469563952077155e-06, "tok_per_s": 789.7465334118826, "loss": 0.5453109331429005, "entropy": 0.5842502880841494, "sharp": 0.19576453790068626}
+{"time": "14:36:50", "event": "step", "step": 460, "rows_seen": 14720, "lr": 6.585509610150972e-06, "tok_per_s": 790.2953562211849, "loss": 0.5787798147648573, "entropy": 0.5871452391147614, "sharp": 0.19477210752665997}
+{"time": "14:40:23", "event": "step", "step": 470, "rows_seen": 15040, "lr": 5.894036597333646e-06, "tok_per_s": 790.9100156576685, "loss": 0.5844223834574223, "entropy": 0.5697647444903851, "sharp": 0.2137035597115755}
+{"time": "14:44:13", "event": "step", "step": 480, "rows_seen": 15360, "lr": 5.398045191744809e-06, "tok_per_s": 791.5516893641025, "loss": 0.5826439559459686, "entropy": 0.5540536381304264, "sharp": 0.22162705846130848}
+{"time": "14:48:02", "event": "step", "step": 490, "rows_seen": 15680, "lr": 5.099615753708697e-06, "tok_per_s": 792.1403214157307, "loss": 0.6066453568637371, "entropy": 0.572606336325407, "sharp": 0.19730823300778866}
+{"time": "14:51:52", "event": "step", "step": 500, "rows_seen": 16000, "lr": 5e-06, "tok_per_s": 792.710006574078, "loss": 0.5467582754790783, "entropy": 0.5287552196532488, "sharp": 0.24401992745697498}
+{"time": "14:53:22", "event": "val", "step": 500, "rows_seen": 16000, "n": 500, "dropped_long": 0, "noul_ce": 0.5664657950401306, "noul_T": 0.9196491534478922, "noul_auc": 0.7473966684106598, "noul_acc": 0.702, "noul_f1": 0.7697063369397218, "noul_nll": 0.5709486426021904, "noul_ece": 0.04796820443868638, "noul_brier": 0.19553080552094526, "noul_sharp": 0.18865476548671722, "entropy": 0.5795356035232544}
+{"time": "14:56:35", "event": "ckpt", "tag": "ckpt-16000", "rows_seen": 16000, "replay_seen": 842, "step": 500, "tokens": 9113125, "tok_per_s": 773.7348806480621, "rows_per_s": 1.4299422923548444, "val": {"n": 1200, "dropped_long": 0, "noul_ce": 0.5519391894340515, "noul_T": 0.9000755854451329, "choice_ce": 0.6665812730789185, "choice_T": 1.008406819192068, "score_ce": 1.0586308240890503, "score_T": 0.9500376826602371, "noul_auc": 0.7466709333333332, "noul_acc": 0.702, "noul_f1": 0.7776119402985074, "noul_nll": 0.5664786969861015, "noul_ece": 0.03434245572984219, "noul_brier": 0.19362598993408342, "noul_sharp": 0.19133822619915009, "entropy": 0.5782209038734436}, "dropped_long": 0, "peak_gib": 18.237416744232178}

e45/S1-E45-SHR-ts1-syn/DONE

@@ -0,0 +1 @@
+2026-11-02 22:10:06

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter/README.md

@@ -0,0 +1,206 @@
+---
+base_model: /workspace/jev_kt/models/hf/JEV-9B
+library_name: peft
+tags:
+- base_model:adapter:/workspace/jev_kt/models/hf/JEV-9B
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.2

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter/adapter_config.json

@@ -0,0 +1,57 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": "/workspace/jev_kt/models/hf/JEV-9B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "k_proj",
+    "q_proj",
+    "o_proj",
+    "out_proj",
+    "v_proj",
+    "down_proj",
+    "in_proj_z",
+    "in_proj_qkv",
+    "gate_proj",
+    "up_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82c36a64467a886a65c667e5d15be0c16fd583466d653660c2d83707454dee76
+size 160486456

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter_vllm/adapter_config.json

@@ -0,0 +1,58 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5ForCausalLM",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": null,
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64.0,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "gate_proj",
+    "in_proj_qkv",
+    "in_proj_z",
+    "k_proj",
+    "lm_head",
+    "o_proj",
+    "out_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c86cb63a52fa53ecfb0dd2c468d9705ae70fac6fde6aa6cd9aae2e4e2843bfd7
+size 176641832

e45/S1-E45-SHR-ts1-syn/ckpt-16000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.0003252780297771096,
+  -0.00032553746132180095,
+  -0.002109350636601448,
+  -0.0014750799164175987,
+  -0.00017475166532676667,
+  0.0019059103215113282,
+  0.0013611947651952505,
+  -0.0021511693485081196,
+  -0.0006911706877872348,
+  -0.000669070752337575,
+  -0.00036951518268324435,
+  0.00026391210849396884,
+  -2.4863120415830053e-05,
+  -0.0037267962470650673,
+  -0.004440860822796822,
+  -0.005115076433867216,
+  -0.005950355902314186,
+  -0.0043793111108243465,
+  -0.004942425061017275,
+  -0.005014660302549601,
+  -0.0038826335221529007,
+  -0.0034174597822129726,
+  -0.005730448290705681,
+  -0.003495796350762248
+ ],
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "note": "decision logits = lm_head-LoRA logprobs of verbalizer ids + bias, then per-kind temperature"
+}

e45/S1-E45-SHR-ts1-syn/ckpt-16000/calibration.json

@@ -0,0 +1,29 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 0.9782350344526988,
+  "choice": 1.0346329807161856,
+  "score": 0.9920105000320746
+ },
+ "fit": {
+  "noul": {
+   "n": 1087,
+   "T": 0.9782350344526988,
+   "ce_before": 0.5506337285041809,
+   "ce_after": 0.5505898594856262
+  },
+  "choice": {
+   "n": 60,
+   "T": 1.0346329807161856,
+   "ce_before": 0.6639683246612549,
+   "ce_after": 0.6637080907821655
+  },
+  "score": {
+   "n": 53,
+   "T": 0.9920105000320746,
+   "ce_before": 1.0500822067260742,
+   "ce_after": 1.0500611066818237
+  }
+ },
+ "source": "models/e45/S1-E45-SHR-ts1-syn/ckpt-16000"
+}

e45/S1-E45-SHR-ts1-syn/ckpt-16000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bf243817030ce0054d6b9aa79b60f6d9ec60ea67c200f5bc8fe5e4f0cb58fd3
+size 393464

e45/S1-E45-SHR-ts1-syn/ckpt-16000/judge_config.json

@@ -0,0 +1,164 @@
+{
+ "base_model_path": "/root/models/Qwen3.5-9B",
+ "hidden_size": 4096,
+ "slots": {
+  "num_slots": 24,
+  "ranges": {
+   "noul": [
+    0,
+    2
+   ],
+   "score": [
+    2,
+    8
+   ],
+   "choice": [
+    8,
+    24
+   ]
+  },
+  "verbalizers": [
+   "false",
+   "true",
+   "0",
+   "1",
+   "2",
+   "3",
+   "4",
+   "5",
+   "A",
+   "B",
+   "C",
+   "D",
+   "E",
+   "F",
+   "G",
+   "H",
+   "I",
+   "J",
+   "K",
+   "L",
+   "M",
+   "N",
+   "O",
+   "P"
+  ],
+  "template_version": "bare-v1"
+ },
+ "verbalizer_ids": [
+  3721,
+  1802,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20,
+  32,
+  33,
+  34,
+  35,
+  36,
+  37,
+  38,
+  39,
+  40,
+  41,
+  42,
+  43,
+  44,
+  45,
+  46,
+  47
+ ],
+ "kinds": [
+  "noul",
+  "choice",
+  "score"
+ ],
+ "model_name": "jev-judge-qwen35-9b",
+ "model_version": "0.8.0",
+ "lm_head_included": true,
+ "weights_mode": "unmerged",
+ "adapter_subfolder": "adapter",
+ "lora": {
+  "r": 16,
+  "lora_alpha": 32,
+  "lora_dropout": 0.05,
+  "target_modules": [
+   "down_proj",
+   "gate_proj",
+   "in_proj_qkv",
+   "in_proj_z",
+   "k_proj",
+   "o_proj",
+   "out_proj",
+   "q_proj",
+   "up_proj",
+   "v_proj"
+  ]
+ },
+ "trained_stage": "e16",
+ "source_checkpoint": "models/e45/S1-E45-SHR-ts1-syn/ckpt-16000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e45/E45-SHR",
+  "data_manifest": {
+   "arm": "E45-SHR",
+   "name": null,
+   "train_sha256": "cfe452b00b8ac702c42580427823a8e4bf112271713c1fea190e1f3214bb6fbf",
+   "n_train": 16000
+  },
+  "init": "released",
+  "objective": "s-log",
+  "sigma": 0.5,
+  "K": 8,
+  "reward": "brier",
+  "clip": 0.2,
+  "passes": 1,
+  "lr": 0.0001,
+  "lr_head": 0.0002,
+  "rows_per_step": 32,
+  "train_seed": 1,
+  "replay": "data/e16/replay/train.jsonl",
+  "replay_frac": 0.05,
+  "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e",
+  "max_len": 3072,
+  "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps",
+  "total_rows": 16000,
+  "lora": {
+   "r": 16,
+   "lora_alpha": 32,
+   "lora_dropout": 0.05
+  },
+  "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3",
+  "event": "ckpt",
+  "tag": "ckpt-16000",
+  "rows_seen": 16000,
+  "replay_seen": 842,
+  "step": 500,
+  "tokens": 9113125,
+  "tok_per_s": 772.8512325148486,
+  "rows_per_s": 1.4283092192638802,
+  "val": {
+   "n": 1200,
+   "dropped_long": 0,
+   "noul_ce": 0.5506337285041809,
+   "noul_T": 0.9782350344526988,
+   "choice_ce": 0.6639683246612549,
+   "choice_T": 1.0346329807161856,
+   "score_ce": 1.0500822067260742,
+   "score_T": 0.9920105000320746,
+   "noul_auc": 0.7479509333333333,
+   "noul_acc": 0.708,
+   "noul_f1": 0.7830609212481426,
+   "noul_nll": 0.5656526157269254,
+   "noul_ece": 0.03333447651565077,
+   "noul_brier": 0.19313562662046876,
+   "noul_sharp": 0.20571254193782806,
+   "entropy": 0.5634832382202148
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.23796796798706
+ }
+}

e45/S1-E45-SHR-ts1-syn/train_log.jsonl

@@ -0,0 +1,57 @@
+{"time": "18:53:33", "event": "start", "data": "data/e45/E45-SHR", "data_manifest": {"arm": "E45-SHR", "name": null, "train_sha256": "cfe452b00b8ac702c42580427823a8e4bf112271713c1fea190e1f3214bb6fbf", "n_train": 16000}, "init": "released", "objective": "s-log", "sigma": 0.5, "K": 8, "reward": "brier", "clip": 0.2, "passes": 1, "lr": 0.0001, "lr_head": 0.0002, "rows_per_step": 32, "train_seed": 1, "replay": "data/e16/replay/train.jsonl", "replay_frac": 0.05, "replay_sha256": "9dd2df19851678750cf79d2fdb47581720930ef2f0c6e18e6fbe23ef2181847e", "max_len": 3072, "schedule": "warmup 0.03 + cosine to x0.05 over 500 steps", "total_rows": 16000, "lora": {"r": 16, "lora_alpha": 32, "lora_dropout": 0.05}, "trainer_sha256": "a0b9d7552588ffb068565686509162a4f122f6f03392e693fa20088493e2a4f3", "device": "cuda", "load_s": 130.59742903709412}
+{"time": "18:57:11", "event": "step", "step": 10, "rows_seen": 320, "lr": 7.333333333333333e-05, "tok_per_s": 788.9108375228361, "loss": 0.596461858600378, "entropy": 0.5812841914594173, "sharp": 0.19884271174669266}
+{"time": "19:00:59", "event": "step", "step": 20, "rows_seen": 640, "lr": 9.997508952966197e-05, "tok_per_s": 802.7024365809305, "loss": 0.5706334626302123, "entropy": 0.5672882199287415, "sharp": 0.2183554070070386}
+{"time": "19:04:44", "event": "step", "step": 30, "rows_seen": 960, "lr": 9.977596250541417e-05, "tok_per_s": 807.2050923133419, "loss": 0.6246401928365231, "entropy": 0.6394802555441856, "sharp": 0.12832688260823488}
+{"time": "19:08:37", "event": "step", "step": 40, "rows_seen": 1280, "lr": 9.937854366485982e-05, "tok_per_s": 809.5677245734037, "loss": 0.5795888714492321, "entropy": 0.6061298176646233, "sharp": 0.17190081719309092}
+{"time": "19:12:24", "event": "step", "step": 50, "rows_seen": 1600, "lr": 9.8784499920729e-05, "tok_per_s": 812.3697530212604, "loss": 0.6682319892570376, "entropy": 0.5732773691415787, "sharp": 0.21406844817101955}
+{"time": "19:16:11", "event": "step", "step": 60, "rows_seen": 1920, "lr": 9.79963228989294e-05, "tok_per_s": 814.0704774618922, "loss": 0.6453411839902401, "entropy": 0.5222227796912193, "sharp": 0.24823970347642899}
+{"time": "19:19:43", "event": "step", "step": 70, "rows_seen": 2240, "lr": 9.701731848780134e-05, "tok_per_s": 814.5828929521163, "loss": 0.5599720664322376, "entropy": 0.6156051717698574, "sharp": 0.17455459292978048}
+{"time": "19:23:18", "event": "step", "step": 80, "rows_seen": 2560, "lr": 9.585159297207315e-05, "tok_per_s": 812.7039118394008, "loss": 0.7072415202856064, "entropy": 0.6348001547157764, "sharp": 0.1514818212017417}
+{"time": "19:27:08", "event": "step", "step": 90, "rows_seen": 2880, "lr": 9.450403580967612e-05, "tok_per_s": 813.2073721201566, "loss": 0.5855435654520988, "entropy": 0.5801449231803417, "sharp": 0.1799279311671853}
+{"time": "19:30:52", "event": "step", "step": 100, "rows_seen": 3200, "lr": 9.298029912365855e-05, "tok_per_s": 813.2511944553574, "loss": 0.5600015837699175, "entropy": 0.5587457902729511, "sharp": 0.2002639714628458}
+{"time": "19:32:22", "event": "val", "step": 100, "rows_seen": 3200, "n": 500, "dropped_long": 0, "noul_ce": 0.5789715051651001, "noul_T": 1.270569736742844, "noul_auc": 0.7408482557256417, "noul_acc": 0.684, "noul_f1": 0.7620481927710844, "noul_nll": 0.5863206258695572, "noul_ece": 0.0585094006061554, "noul_brier": 0.2003971316914961, "noul_sharp": 0.23130908608436584, "entropy": 0.5287705063819885}
+{"time": "19:36:05", "event": "step", "step": 110, "rows_seen": 3520, "lr": 9.128677399521752e-05, "tok_per_s": 785.3646677559285, "loss": 0.5870157070457935, "entropy": 0.6402954310178757, "sharp": 0.13630897924304008}
+{"time": "19:39:48", "event": "step", "step": 120, "rows_seen": 3840, "lr": 8.943056365728335e-05, "tok_per_s": 787.7989599508645, "loss": 0.6023912541568279, "entropy": 0.5724473372101784, "sharp": 0.21456132642924786}
+{"time": "19:43:33", "event": "step", "step": 130, "rows_seen": 4160, "lr": 8.741945370109202e-05, "tok_per_s": 790.5045637632652, "loss": 0.5498399389907718, "entropy": 0.5855995900928974, "sharp": 0.19138554856181145}
+{"time": "19:46:59", "event": "step", "step": 140, "rows_seen": 4480, "lr": 8.526187942070997e-05, "tok_per_s": 791.4048884715161, "loss": 0.5716703496873379, "entropy": 0.5842510499060154, "sharp": 0.1964036449790001}
+{"time": "19:50:43", "event": "step", "step": 150, "rows_seen": 4800, "lr": 8.296689043247915e-05, "tok_per_s": 793.0075659503759, "loss": 0.6169856302440166, "entropy": 0.603382833302021, "sharp": 0.17428217362612486}
+{"time": "19:54:30", "event": "step", "step": 160, "rows_seen": 5120, "lr": 8.054411271778132e-05, "tok_per_s": 794.1409043839134, "loss": 0.620785515755415, "entropy": 0.5521061196923256, "sharp": 0.22222652658820152}
+{"time": "19:58:21", "event": "step", "step": 170, "rows_seen": 5440, "lr": 7.800370824832674e-05, "tok_per_s": 796.0099915955241, "loss": 0.5343157220631838, "entropy": 0.5559712052345276, "sharp": 0.22559254430234432}
+{"time": "20:01:59", "event": "step", "step": 180, "rows_seen": 5760, "lr": 7.535633236331324e-05, "tok_per_s": 797.4218597139933, "loss": 0.6017148923128843, "entropy": 0.6178525611758232, "sharp": 0.15638525690883398}
+{"time": "20:05:44", "event": "step", "step": 190, "rows_seen": 6080, "lr": 7.26130890772308e-05, "tok_per_s": 798.3423226516074, "loss": 0.6396661214530468, "entropy": 0.5542013235390186, "sharp": 0.22321678698062897}
+{"time": "20:09:28", "event": "step", "step": 200, "rows_seen": 6400, "lr": 6.978548450576568e-05, "tok_per_s": 799.8433404694879, "loss": 0.5909261740744114, "entropy": 0.6189185008406639, "sharp": 0.16114056948572397}
+{"time": "20:10:58", "event": "val", "step": 200, "rows_seen": 6400, "n": 500, "dropped_long": 0, "noul_ce": 0.5841028690338135, "noul_T": 0.9263722563664525, "noul_auc": 0.7409326424870466, "noul_acc": 0.696, "noul_f1": 0.7571884984025559, "noul_nll": 0.5925828211307526, "noul_ece": 0.07404012078046801, "noul_brier": 0.20303559332346238, "noul_sharp": 0.1821174919605255, "entropy": 0.5956742167472839}
+{"time": "20:14:37", "event": "step", "step": 210, "rows_seen": 6720, "lr": 6.688537860515266e-05, "tok_per_s": 785.8143154522013, "loss": 0.6587777808308601, "entropy": 0.577390443533659, "sharp": 0.19967974349856377}
+{"time": "20:18:16", "event": "step", "step": 220, "rows_seen": 7040, "lr": 6.392493542739706e-05, "tok_per_s": 786.8886662761385, "loss": 0.6146156340837479, "entropy": 0.6455931141972542, "sharp": 0.13014758564531803}
+{"time": "20:21:58", "event": "step", "step": 230, "rows_seen": 7360, "lr": 6.091657210001309e-05, "tok_per_s": 788.3677107304746, "loss": 0.5936343315988779, "entropy": 0.5602922551333904, "sharp": 0.2292973306030035}
+{"time": "20:25:39", "event": "step", "step": 240, "rows_seen": 7680, "lr": 5.7872906744276015e-05, "tok_per_s": 789.3654320677376, "loss": 0.6161071266978979, "entropy": 0.5455464068800211, "sharp": 0.23009579442441463}
+{"time": "20:29:24", "event": "step", "step": 250, "rows_seen": 8000, "lr": 5.48067055504363e-05, "tok_per_s": 790.4050123279826, "loss": 0.5670283250510693, "entropy": 0.5637671612203121, "sharp": 0.21714975591748953}
+{"time": "20:33:14", "event": "step", "step": 260, "rows_seen": 8320, "lr": 5.173082923188119e-05, "tok_per_s": 791.9481862506342, "loss": 0.555771404877305, "entropy": 0.5772262215614319, "sharp": 0.19098717998713255}
+{"time": "20:36:50", "event": "step", "step": 270, "rows_seen": 8640, "lr": 4.865817908283331e-05, "tok_per_s": 792.4034987826083, "loss": 0.56440619379282, "entropy": 0.5734024085104465, "sharp": 0.20211188308894634}
+{"time": "20:40:35", "event": "step", "step": 280, "rows_seen": 8960, "lr": 4.560164286583922e-05, "tok_per_s": 793.2962897028677, "loss": 0.6423080265522003, "entropy": 0.6078539825975895, "sharp": 0.16601537633687258}
+{"time": "20:44:13", "event": "step", "step": 290, "rows_seen": 9280, "lr": 4.257404075601443e-05, "tok_per_s": 794.0565327450719, "loss": 0.578652473166585, "entropy": 0.5809906534850597, "sharp": 0.18850169982761145}
+{"time": "20:47:52", "event": "step", "step": 300, "rows_seen": 9600, "lr": 3.958807156877371e-05, "tok_per_s": 794.7385407470196, "loss": 0.6389638297259808, "entropy": 0.5469372831285, "sharp": 0.2150721661746502}
+{"time": "20:49:22", "event": "val", "step": 300, "rows_seen": 9600, "n": 500, "dropped_long": 0, "noul_ce": 0.5647706389427185, "noul_T": 1.0925834789923423, "noul_auc": 0.7496919883208722, "noul_acc": 0.696, "noul_f1": 0.7744807121661721, "noul_nll": 0.5713443754622712, "noul_ece": 0.05046275341510773, "noul_brier": 0.19556456142851664, "noul_sharp": 0.21890029311180115, "entropy": 0.547981321811676}
+{"time": "20:53:07", "event": "step", "step": 310, "rows_seen": 9920, "lr": 3.665625949658603e-05, "tok_per_s": 785.2971380725365, "loss": 0.6155873164534569, "entropy": 0.6031679771840572, "sharp": 0.17789683863520622}
+{"time": "20:56:41", "event": "step", "step": 320, "rows_seen": 10240, "lr": 3.379090157815887e-05, "tok_per_s": 786.4162737859978, "loss": 0.5580587852746248, "entropy": 0.5367497839033604, "sharp": 0.21209662314504385}
+{"time": "21:00:22", "event": "step", "step": 330, "rows_seen": 10560, "lr": 3.10040161203844e-05, "tok_per_s": 787.4998839128795, "loss": 0.5286420732736588, "entropy": 0.5696161724627018, "sharp": 0.18582206312566996}
+{"time": "21:04:13", "event": "step", "step": 340, "rows_seen": 10880, "lr": 2.8307292289383776e-05, "tok_per_s": 788.5914734898441, "loss": 0.6324968971312046, "entropy": 0.5659818165004253, "sharp": 0.19420167431235313}
+{"time": "21:07:55", "event": "step", "step": 350, "rows_seen": 11200, "lr": 2.571204108208194e-05, "tok_per_s": 789.5637626197662, "loss": 0.5199874304234982, "entropy": 0.5573053434491158, "sharp": 0.20493842475116253}
+{"time": "21:11:35", "event": "step", "step": 360, "rows_seen": 11520, "lr": 2.322914788395595e-05, "tok_per_s": 790.5788337598569, "loss": 0.6683460548520088, "entropy": 0.566481851041317, "sharp": 0.2028658501803875}
+{"time": "21:15:06", "event": "step", "step": 370, "rows_seen": 11840, "lr": 2.0869026811945157e-05, "tok_per_s": 791.2726378841171, "loss": 0.5499088987708092, "entropy": 0.556165836751461, "sharp": 0.21614781767129898}
+{"time": "21:18:47", "event": "step", "step": 380, "rows_seen": 12160, "lr": 1.8641577034025627e-05, "tok_per_s": 792.0976341294904, "loss": 0.618893870152533, "entropy": 0.6156249158084393, "sharp": 0.16819471772760153}
+{"time": "21:22:23", "event": "step", "step": 390, "rows_seen": 12480, "lr": 1.6556141248659472e-05, "tok_per_s": 792.8694295109489, "loss": 0.6679055094718933, "entropy": 0.6091725677251816, "sharp": 0.16861839033663273}
+{"time": "21:26:05", "event": "step", "step": 400, "rows_seen": 12800, "lr": 1.4621466498270773e-05, "tok_per_s": 793.4598625335732, "loss": 0.5506848655641079, "entropy": 0.5711837820708752, "sharp": 0.20783033594489098}
+{"time": "21:27:35", "event": "val", "step": 400, "rows_seen": 12800, "n": 500, "dropped_long": 0, "noul_ce": 0.5641158223152161, "noul_T": 1.0072295059291543, "noul_auc": 0.7495569695026244, "noul_acc": 0.708, "noul_f1": 0.771875, "noul_nll": 0.5695498082172126, "noul_ece": 0.04122230046987534, "noul_brier": 0.1946594093253056, "noul_sharp": 0.20612965524196625, "entropy": 0.5629354119300842}
+{"time": "21:31:15", "event": "step", "step": 410, "rows_seen": 13120, "lr": 1.284566748110963e-05, "tok_per_s": 786.402885339447, "loss": 0.5875614024698734, "entropy": 0.5901710875332355, "sharp": 0.18518336489796638}
+{"time": "21:35:08", "event": "step", "step": 420, "rows_seen": 13440, "lr": 1.1236192515387821e-05, "tok_per_s": 787.0352618457144, "loss": 0.530891191214323, "entropy": 0.5711709372699261, "sharp": 0.20252786949276924}
+{"time": "21:38:53", "event": "step", "step": 430, "rows_seen": 13760, "lr": 9.799792298443728e-06, "tok_per_s": 787.794053254493, "loss": 0.5797514878213406, "entropy": 0.5830639153718948, "sharp": 0.1920479591935873}
+{"time": "21:42:48", "event": "step", "step": 440, "rows_seen": 14080, "lr": 8.542491591971554e-06, "tok_per_s": 788.6121816191202, "loss": 0.6729071475565434, "entropy": 0.5920260474085808, "sharp": 0.18566079623997211}
+{"time": "21:46:36", "event": "step", "step": 450, "rows_seen": 14400, "lr": 7.469563952077155e-06, "tok_per_s": 789.247102360759, "loss": 0.601127521134913, "entropy": 0.593987112864852, "sharp": 0.18038462288677692}
+{"time": "21:50:21", "event": "step", "step": 460, "rows_seen": 14720, "lr": 6.585509610150972e-06, "tok_per_s": 789.6758922735186, "loss": 0.5831257402896881, "entropy": 0.5738185830414295, "sharp": 0.19984791800379753}
+{"time": "21:54:10", "event": "step", "step": 470, "rows_seen": 15040, "lr": 5.894036597333646e-06, "tok_per_s": 790.0808091776232, "loss": 0.6089097522199154, "entropy": 0.5740026235580444, "sharp": 0.20401583053171635}
+{"time": "21:57:52", "event": "step", "step": 480, "rows_seen": 15360, "lr": 5.398045191744809e-06, "tok_per_s": 790.6181575917767, "loss": 0.5815652180463076, "entropy": 0.585292112082243, "sharp": 0.19261294417083263}
+{"time": "22:01:41", "event": "step", "step": 490, "rows_seen": 15680, "lr": 5.099615753708697e-06, "tok_per_s": 791.1440328337729, "loss": 0.5746835768222809, "entropy": 0.6032003685832024, "sharp": 0.1721150204539299}
+{"time": "22:05:23", "event": "step", "step": 500, "rows_seen": 16000, "lr": 5e-06, "tok_per_s": 791.7637853489691, "loss": 0.5333664081990719, "entropy": 0.5551153346896172, "sharp": 0.20426847971975803}
+{"time": "22:06:53", "event": "val", "step": 500, "rows_seen": 16000, "n": 500, "dropped_long": 0, "noul_ce": 0.5648979544639587, "noul_T": 1.0084058584423725, "noul_auc": 0.7487637339454187, "noul_acc": 0.708, "noul_f1": 0.7760736196319018, "noul_nll": 0.5700209506750107, "noul_ece": 0.05854451191425324, "noul_brier": 0.19479488186794344, "noul_sharp": 0.2042987048625946, "entropy": 0.5634372234344482}
+{"time": "22:10:06", "event": "ckpt", "tag": "ckpt-16000", "rows_seen": 16000, "replay_seen": 842, "step": 500, "tokens": 9113125, "tok_per_s": 772.8512325148486, "rows_per_s": 1.4283092192638802, "val": {"n": 1200, "dropped_long": 0, "noul_ce": 0.5506337285041809, "noul_T": 0.9782350344526988, "choice_ce": 0.6639683246612549, "choice_T": 1.0346329807161856, "score_ce": 1.0500822067260742, "score_T": 0.9920105000320746, "noul_auc": 0.7479509333333333, "noul_acc": 0.708, "noul_f1": 0.7830609212481426, "noul_nll": 0.5656526157269254, "noul_ece": 0.03333447651565077, "noul_brier": 0.19313562662046876, "noul_sharp": 0.20571254193782806, "entropy": 0.5634832382202148}, "dropped_long": 0, "peak_gib": 18.23796796798706}