backup e16/S2-rloo (Pod B)

e16/S2-rloo/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

e16/S2-rloo/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",
+    "gate_proj",
+    "in_proj_qkv",
+    "up_proj",
+    "down_proj",
+    "q_proj",
+    "in_proj_z",
+    "k_proj",
+    "v_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
+}

e16/S2-rloo/ckpt-16000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e9ea61c5b62a0507d995471b93167b366675eb0fa90d51d0d6e59d4c8a05e41
+size 160486456

e16/S2-rloo/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
+}

e16/S2-rloo/ckpt-16000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9cb9142b6d1f1cbc2f73616d0bf35a753721c1d51190f7c4a0f892272a985c3
+size 176641832

e16/S2-rloo/ckpt-16000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.0019170412560924888,
+  0.0014889377634972334,
+  -0.0025878134183585644,
+  -0.0012615364976227283,
+  -0.00025007399381138384,
+  0.0019558551721274853,
+  0.0013559526996687055,
+  -0.0023348911199718714,
+  -0.0004082925443071872,
+  -0.0002078723773593083,
+  -1.2957353646925185e-05,
+  0.0001676026004133746,
+  0.0005569425993598998,
+  -0.001412136247381568,
+  -0.0028663526754826307,
+  -0.005061559844762087,
+  -0.00593328895047307,
+  -0.004319756291806698,
+  -0.004922541789710522,
+  -0.0049067651852965355,
+  -0.003943523392081261,
+  -0.003319827374070883,
+  -0.005581587553024292,
+  -0.0034480984322726727
+ ],
+ "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"
+}

e16/S2-rloo/ckpt-16000/calibration.json

@@ -0,0 +1,17 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 3.999999975766852,
+  "choice": 1.0,
+  "score": 1.0
+ },
+ "fit": {
+  "noul": {
+   "n": 2000,
+   "T": 3.999999975766852,
+   "ce_before": 1.5079070329666138,
+   "ce_after": 0.6359632611274719
+  }
+ },
+ "source": "models/e16/S2-rloo/ckpt-16000"
+}

e16/S2-rloo/ckpt-16000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:955a60c8f3e6e0b477778ec8471dc40118eb17cd2bf4748a83b83c1f7b4f2387
+size 393464

e16/S2-rloo/ckpt-16000/judge_config.json

@@ -0,0 +1,160 @@
+{
+ "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/e16/S2-rloo/ckpt-16000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e15/F1-pool",
+  "data_manifest": {
+   "arm": "F1-pool",
+   "name": "F1-pool",
+   "train_sha256": "dad16f94a5659768b254a6c12538c01f9c64cd43d574a3435bcc70c9c1bac130",
+   "n_train": 64000
+  },
+  "init": "released",
+  "objective": "rloo",
+  "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": null,
+  "replay_frac": 0.0,
+  "replay_sha256": null,
+  "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": "bc7c08e836bd99bb2665b02903cd41404ea029ecfa9b68a9ec1a7cda96958ba4",
+  "event": "ckpt",
+  "tag": "ckpt-16000",
+  "rows_seen": 16000,
+  "replay_seen": 0,
+  "step": 500,
+  "tokens": 10544036,
+  "tok_per_s": 794.1320291177743,
+  "rows_per_s": 1.2050520749314817,
+  "val": {
+   "n": 2000,
+   "dropped_long": 0,
+   "noul_ce": 1.5079070329666138,
+   "noul_T": 3.999999975766852,
+   "noul_auc": 0.7524481645604246,
+   "noul_acc": 0.736,
+   "noul_f1": 0.8050221565731167,
+   "noul_nll": 1.1806853275026115,
+   "noul_ece": 0.12911880904436113,
+   "noul_brier": 0.20666864844414723,
+   "noul_sharp": 0.3424508273601532,
+   "entropy": 0.2810675799846649
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.236032962799072
+ }
+}

e16/S2-rloo/ckpt-4000/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

e16/S2-rloo/ckpt-4000/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",
+    "gate_proj",
+    "in_proj_qkv",
+    "up_proj",
+    "down_proj",
+    "q_proj",
+    "in_proj_z",
+    "k_proj",
+    "v_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
+}

e16/S2-rloo/ckpt-4000/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9429c4b1aaeb1e602c8bf241b0e156d2de743b7b6cc851294a45b8feca6b1cfe
+size 160486456

e16/S2-rloo/ckpt-4000/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
+}

e16/S2-rloo/ckpt-4000/adapter_vllm/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1807f3b7e8fcb510de2cf1aab869c509ec54497cc7523b9b9e5f3b311d31da47
+size 176641832

e16/S2-rloo/ckpt-4000/adapter_vllm/decision_head.json

@@ -0,0 +1,99 @@
+{
+ "bias": [
+  0.002288899151608348,
+  0.001431172713637352,
+  -0.0025878134183585644,
+  -0.0012615364976227283,
+  -0.00025007399381138384,
+  0.0019558551721274853,
+  0.0013559526996687055,
+  -0.0023348911199718714,
+  -0.0004082925443071872,
+  -0.0002078723773593083,
+  -1.2957353646925185e-05,
+  0.0001676026004133746,
+  0.0005569425993598998,
+  -0.001412136247381568,
+  -0.0028663526754826307,
+  -0.005061559844762087,
+  -0.00593328895047307,
+  -0.004319756291806698,
+  -0.004922541789710522,
+  -0.0049067651852965355,
+  -0.003943523392081261,
+  -0.003319827374070883,
+  -0.005581587553024292,
+  -0.0034480984322726727
+ ],
+ "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"
+}

e16/S2-rloo/ckpt-4000/calibration.json

@@ -0,0 +1,17 @@
+{
+ "version": 1,
+ "per_kind": {
+  "noul": 3.999999975766852,
+  "choice": 1.0,
+  "score": 1.0
+ },
+ "fit": {
+  "noul": {
+   "n": 2000,
+   "T": 3.999999975766852,
+   "ce_before": 1.2215243577957153,
+   "ce_after": 0.5708816051483154
+  }
+ },
+ "source": "models/e16/S2-rloo/ckpt-4000"
+}

e16/S2-rloo/ckpt-4000/head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:feccf69778eeffe4be05118147dd36f66c73640943596c8f73fc9242c11b464f
+size 393464

e16/S2-rloo/ckpt-4000/judge_config.json

@@ -0,0 +1,160 @@
+{
+ "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/e16/S2-rloo/ckpt-4000",
+ "softcap": null,
+ "e16": {
+  "data": "data/e15/F1-pool",
+  "data_manifest": {
+   "arm": "F1-pool",
+   "name": "F1-pool",
+   "train_sha256": "dad16f94a5659768b254a6c12538c01f9c64cd43d574a3435bcc70c9c1bac130",
+   "n_train": 64000
+  },
+  "init": "released",
+  "objective": "rloo",
+  "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": null,
+  "replay_frac": 0.0,
+  "replay_sha256": null,
+  "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": "bc7c08e836bd99bb2665b02903cd41404ea029ecfa9b68a9ec1a7cda96958ba4",
+  "event": "ckpt",
+  "tag": "ckpt-4000",
+  "rows_seen": 4000,
+  "replay_seen": 0,
+  "step": 125,
+  "tokens": 2631181,
+  "tok_per_s": 743.7598774535795,
+  "rows_per_s": 1.1306859940845242,
+  "val": {
+   "n": 2000,
+   "dropped_long": 0,
+   "noul_ce": 1.2215243577957153,
+   "noul_T": 3.999999975766852,
+   "noul_auc": 0.7675519247580864,
+   "noul_acc": 0.7355,
+   "noul_f1": 0.8115425721410758,
+   "noul_nll": 1.2159270420946378,
+   "noul_ece": 0.2217770409432269,
+   "noul_brier": 0.2385358699212228,
+   "noul_sharp": 0.4569796621799469,
+   "entropy": 0.11558810621500015
+  },
+  "dropped_long": 0,
+  "peak_gib": 18.23348903656006
+ }
+}

e16/S2-rloo/train_log.jsonl

@@ -0,0 +1,58 @@
+{"time": "14:02:23", "event": "start", "data": "data/e15/F1-pool", "data_manifest": {"arm": "F1-pool", "name": "F1-pool", "train_sha256": "dad16f94a5659768b254a6c12538c01f9c64cd43d574a3435bcc70c9c1bac130", "n_train": 64000}, "init": "released", "objective": "rloo", "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": null, "replay_frac": 0.0, "replay_sha256": null, "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": "bc7c08e836bd99bb2665b02903cd41404ea029ecfa9b68a9ec1a7cda96958ba4", "device": "cuda", "load_s": 78.6195502281189}
+{"time": "14:06:33", "event": "step", "step": 10, "rows_seen": 320, "lr": 7.333333333333333e-05, "tok_per_s": 848.4206331066098, "loss": 0.041438812389969826, "entropy": 0.35938069224357605, "sharp": 0.3437623120844364, "r_mean": 0.5256468219061693, "r_std": 0.1442729141563177, "adv_abs": 0.1284753413250049, "frac_lowstd": 0.027777778605620067}
+{"time": "14:10:39", "event": "step", "step": 20, "rows_seen": 640, "lr": 9.997508952966197e-05, "tok_per_s": 860.4783886024419, "loss": -0.024374883505515754, "entropy": 0.5490777641534805, "sharp": 0.2285260409116745, "r_mean": 0.5836443049567086, "r_std": 0.21817410843712942, "adv_abs": 0.18964819610118866, "frac_lowstd": 0.0}
+{"time": "14:14:41", "event": "step", "step": 30, "rows_seen": 960, "lr": 9.977596250541417e-05, "tok_per_s": 861.7982545865623, "loss": -0.013553223514463753, "entropy": 0.345874335616827, "sharp": 0.35876463539898396, "r_mean": 0.5813365876674652, "r_std": 0.11137886930789266, "adv_abs": 0.09503363738102573, "frac_lowstd": 0.08095238251345498}
+{"time": "14:18:47", "event": "step", "step": 40, "rows_seen": 1280, "lr": 9.937854366485982e-05, "tok_per_s": 865.683466535071, "loss": -0.004199856921331957, "entropy": 0.2116803330136463, "sharp": 0.4246623106300831, "r_mean": 0.22893459030560084, "r_std": 0.06734940023826701, "adv_abs": 0.05649631417223385, "frac_lowstd": 0.20476190745830536}
+{"time": "14:22:50", "event": "step", "step": 50, "rows_seen": 1600, "lr": 9.8784499920729e-05, "tok_per_s": 867.5823264301616, "loss": -0.013359824937651865, "entropy": 0.39283818379044533, "sharp": 0.3377805333584547, "r_mean": 0.3827137861933027, "r_std": 0.15597181980098998, "adv_abs": 0.14091989185128892, "frac_lowstd": 0.0}
+{"time": "14:26:52", "event": "step", "step": 60, "rows_seen": 1920, "lr": 9.79963228989294e-05, "tok_per_s": 868.325927139668, "loss": -0.035279660602100194, "entropy": 0.5153252314776182, "sharp": 0.25046114064753056, "r_mean": 0.5478154804025378, "r_std": 0.16538604774645396, "adv_abs": 0.14661468352590287, "frac_lowstd": 0.0}
+{"time": "14:30:52", "event": "step", "step": 70, "rows_seen": 2240, "lr": 9.701731848780134e-05, "tok_per_s": 868.1111419267623, "loss": -0.10882948664948344, "entropy": 0.587401308119297, "sharp": 0.18047377653419971, "r_mean": 0.44600629593644825, "r_std": 0.28085402292864664, "adv_abs": 0.24854700905936106, "frac_lowstd": 0.0}
+{"time": "14:34:55", "event": "step", "step": 80, "rows_seen": 2560, "lr": 9.585159297207315e-05, "tok_per_s": 868.7456635363826, "loss": 0.012059497297741473, "entropy": 0.4005155898630619, "sharp": 0.3006207291036844, "r_mean": 0.2404645127909524, "r_std": 0.24007371493748256, "adv_abs": 0.2066885307431221, "frac_lowstd": 0.028571428997176036}
+{"time": "14:38:59", "event": "step", "step": 90, "rows_seen": 2880, "lr": 9.450403580967612e-05, "tok_per_s": 868.3868272983132, "loss": -0.0038165750447660685, "entropy": 0.41817887127399445, "sharp": 0.321759270504117, "r_mean": 0.6030687093734741, "r_std": 0.1411557729755129, "adv_abs": 0.12439010185854775, "frac_lowstd": 0.023809524519102915}
+{"time": "14:42:59", "event": "step", "step": 100, "rows_seen": 3200, "lr": 9.298029912365855e-05, "tok_per_s": 869.5955947572941, "loss": -0.04541801943560131, "entropy": 0.3930474314838648, "sharp": 0.3333327081054449, "r_mean": 0.6234067891325269, "r_std": 0.1205363039459501, "adv_abs": 0.10534823153700147, "frac_lowstd": 0.11564626225403377}
+{"time": "14:44:45", "event": "val", "step": 100, "rows_seen": 3200, "n": 500, "dropped_long": 0, "noul_ce": 0.7124685049057007, "noul_T": 2.588144419134068, "noul_auc": 0.7464227945750903, "noul_acc": 0.706, "noul_f1": 0.7972413793103448, "noul_nll": 0.7124685025063809, "noul_ece": 0.15569849795103077, "noul_brier": 0.21813480559378093, "noul_sharp": 0.35806092619895935, "entropy": 0.3165280818939209}
+{"time": "14:48:45", "event": "step", "step": 110, "rows_seen": 3520, "lr": 9.128677399521752e-05, "tok_per_s": 838.4587975024943, "loss": -1.9818195141851902e-05, "entropy": 0.297555407974869, "sharp": 0.389674523845315, "r_mean": 0.6391956806182861, "r_std": 0.08577407005109958, "adv_abs": 0.07833598128386907, "frac_lowstd": 0.07619047803538186}
+{"time": "14:52:34", "event": "step", "step": 120, "rows_seen": 3840, "lr": 8.943056365728335e-05, "tok_per_s": 841.9619123776896, "loss": 0.0023734680279687836, "entropy": 0.02616528398758078, "sharp": 0.49341733381152153, "r_mean": 0.28208414784499575, "r_std": 0.011938862279403597, "adv_abs": 0.01020854864349222, "frac_lowstd": 0.8880952426365444}
+{"time": "15:01:23", "event": "ckpt", "tag": "ckpt-4000", "rows_seen": 4000, "replay_seen": 0, "step": 125, "tokens": 2631181, "tok_per_s": 743.7598774535795, "rows_per_s": 1.1306859940845242, "val": {"n": 2000, "dropped_long": 0, "noul_ce": 1.2215243577957153, "noul_T": 3.999999975766852, "noul_auc": 0.7675519247580864, "noul_acc": 0.7355, "noul_f1": 0.8115425721410758, "noul_nll": 1.2159270420946378, "noul_ece": 0.2217770409432269, "noul_brier": 0.2385358699212228, "noul_sharp": 0.4569796621799469, "entropy": 0.11558810621500015}, "dropped_long": 0, "peak_gib": 18.23348903656006}
+{"time": "15:03:24", "event": "step", "step": 130, "rows_seen": 4160, "lr": 8.741945370109202e-05, "tok_per_s": 748.3176293142499, "loss": 0.02652382419910282, "entropy": 0.11702041700482368, "sharp": 0.4637985359877348, "r_mean": -0.33958197331854273, "r_std": 0.10106892484639372, "adv_abs": 0.08185844256409577, "frac_lowstd": 0.26734694625650135}
+{"time": "15:07:22", "event": "step", "step": 140, "rows_seen": 4480, "lr": 8.526187942070997e-05, "tok_per_s": 756.4924489464803, "loss": -4.216732252260158e-07, "entropy": 0.0017725095433339666, "sharp": 0.49969928711652756, "r_mean": 0.5738158226013184, "r_std": 1.031888000088291e-05, "adv_abs": 9.571158518184575e-06, "frac_lowstd": 1.0}
+{"time": "15:11:21", "event": "step", "step": 150, "rows_seen": 4800, "lr": 8.296689043247915e-05, "tok_per_s": 764.117030910927, "loss": -4.310124118100944e-05, "entropy": 0.0015504464602113235, "sharp": 0.4997314177453518, "r_mean": 0.4866114194904055, "r_std": 0.0007602079631122852, "adv_abs": 0.0005663092046006568, "frac_lowstd": 0.976190481867109}
+{"time": "15:15:21", "event": "step", "step": 160, "rows_seen": 5120, "lr": 8.054411271778132e-05, "tok_per_s": 771.2204853501031, "loss": 1.1593284110489321e-07, "entropy": 5.788841535506704e-06, "sharp": 0.49999960511922836, "r_mean": 0.27500086029370624, "r_std": 7.439888906427919e-07, "adv_abs": 6.520795778897082e-07, "frac_lowstd": 1.0}
+{"time": "15:19:21", "event": "step", "step": 170, "rows_seen": 5440, "lr": 7.800370824832674e-05, "tok_per_s": 777.0945824607301, "loss": 1.9436788889493073e-07, "entropy": 1.2728918151128088e-05, "sharp": 0.49999901466071606, "r_mean": 0.17076036263080954, "r_std": 8.564047683583859e-06, "adv_abs": 7.991222862610162e-06, "frac_lowstd": 1.0}
+{"time": "15:23:11", "event": "step", "step": 180, "rows_seen": 5760, "lr": 7.535633236331324e-05, "tok_per_s": 782.3033865586416, "loss": -1.2110852094737723e-05, "entropy": 0.0014451713326915625, "sharp": 0.49975254386663437, "r_mean": 0.46103302389383316, "r_std": 0.0003418794247872133, "adv_abs": 0.0002785287925034889, "frac_lowstd": 0.9761904875437418}
+{"time": "15:27:04", "event": "step", "step": 190, "rows_seen": 6080, "lr": 7.26130890772308e-05, "tok_per_s": 787.5344688979986, "loss": 4.0386431632288833e-07, "entropy": 2.911732537356159e-06, "sharp": 0.4999997988343239, "r_mean": 0.32381031059083754, "r_std": 5.244163909472783e-07, "adv_abs": 4.712416595355405e-07, "frac_lowstd": 1.0}
+{"time": "15:31:11", "event": "step", "step": 200, "rows_seen": 6400, "lr": 6.978548450576568e-05, "tok_per_s": 792.1073083959546, "loss": 7.20719519620161e-08, "entropy": 2.5789326455699326e-06, "sharp": 0.49999983981251717, "r_mean": 0.5374154835229936, "r_std": 4.364753424722169e-07, "adv_abs": 3.727462935135658e-07, "frac_lowstd": 1.0}
+{"time": "15:32:59", "event": "val", "step": 200, "rows_seen": 6400, "n": 500, "dropped_long": 0, "noul_ce": 6.765253067016602, "noul_T": 3.999999389801139, "noul_auc": 0.6800867416768873, "noul_acc": 0.654, "noul_f1": 0.7223113964686998, "noul_nll": 3.1591656881656234, "noul_ece": 0.3381845265794412, "noul_brier": 0.34576921033477936, "noul_sharp": 0.4997777044773102, "entropy": 0.0009848192567005754}
+{"time": "15:36:58", "event": "step", "step": 210, "rows_seen": 6720, "lr": 6.688537860515266e-05, "tok_per_s": 781.0665753157797, "loss": 4.812768203965589e-07, "entropy": 2.221057011224034e-06, "sharp": 0.49999984726309776, "r_mean": 0.3714291639626026, "r_std": 4.980298159905163e-07, "adv_abs": 3.9892539303666907e-07, "frac_lowstd": 1.0}
+{"time": "15:40:49", "event": "step", "step": 220, "rows_seen": 7040, "lr": 6.392493542739706e-05, "tok_per_s": 785.3013905989139, "loss": 2.563021197055093e-06, "entropy": 1.8494287358742137e-05, "sharp": 0.49999842792749405, "r_mean": 0.29630509453515214, "r_std": 6.115514589547881e-06, "adv_abs": 5.361974298049669e-06, "frac_lowstd": 1.0}
+{"time": "15:44:46", "event": "step", "step": 230, "rows_seen": 7360, "lr": 6.091657210001309e-05, "tok_per_s": 789.8662714572279, "loss": 3.3425904266692896e-09, "entropy": 1.1936691517837908e-05, "sharp": 0.4999990537762642, "r_mean": 0.40680275431701113, "r_std": 1.3117384258163968e-08, "adv_abs": 1.0078919794003533e-08, "frac_lowstd": 1.0}
+{"time": "15:48:32", "event": "step", "step": 240, "rows_seen": 7680, "lr": 5.7872906744276015e-05, "tok_per_s": 793.7142664538043, "loss": -1.1313892881226195e-05, "entropy": 0.00039026095780928037, "sharp": 0.49994753301143646, "r_mean": 0.3062727791922433, "r_std": 6.358746743728716e-05, "adv_abs": 6.221470264985587e-05, "frac_lowstd": 0.976190481867109}
+{"time": "15:52:27", "event": "step", "step": 250, "rows_seen": 8000, "lr": 5.48067055504363e-05, "tok_per_s": 797.3956533089429, "loss": 5.620775129777655e-07, "entropy": 1.3571450058698692e-05, "sharp": 0.4999989718198776, "r_mean": 0.7023836033684867, "r_std": 1.7312777629285847e-06, "adv_abs": 1.5192089343989729e-06, "frac_lowstd": 1.0}
+{"time": "15:56:14", "event": "step", "step": 260, "rows_seen": 8320, "lr": 5.173082923188119e-05, "tok_per_s": 800.6967430579092, "loss": 1.4680231775976438e-05, "entropy": 0.00017474734318122454, "sharp": 0.49998070299625397, "r_mean": 0.44293783099523615, "r_std": 4.633854834829825e-05, "adv_abs": 4.1259150183999994e-05, "frac_lowstd": 0.976190481867109}
+{"time": "16:00:10", "event": "step", "step": 270, "rows_seen": 8640, "lr": 4.865817908283331e-05, "tok_per_s": 803.623305835375, "loss": 2.3891061937675317e-06, "entropy": 4.09833214571933e-05, "sharp": 0.49999646842479706, "r_mean": 0.29049361248811084, "r_std": 1.2978635330303709e-05, "adv_abs": 1.0903257693447207e-05, "frac_lowstd": 1.0}
+{"time": "16:04:06", "event": "step", "step": 280, "rows_seen": 8960, "lr": 4.560164286583922e-05, "tok_per_s": 806.6656045839693, "loss": 2.1505879033156816e-06, "entropy": 4.5353465949112015e-05, "sharp": 0.499996118247509, "r_mean": 0.20137228532920354, "r_std": 6.038571655153646e-06, "adv_abs": 5.2198092816979624e-06, "frac_lowstd": 1.0}
+{"time": "16:08:06", "event": "step", "step": 290, "rows_seen": 9280, "lr": 4.257404075601443e-05, "tok_per_s": 809.4904006908586, "loss": 0.0101719691301696, "entropy": 0.08484099153430225, "sharp": 0.4755537435412407, "r_mean": 0.5028826211180005, "r_std": 0.04199060403542327, "adv_abs": 0.03555102371943316, "frac_lowstd": 0.6772109014647347}
+{"time": "16:12:02", "event": "step", "step": 300, "rows_seen": 9600, "lr": 3.958807156877371e-05, "tok_per_s": 811.7714260326845, "loss": -0.010017310676630586, "entropy": 0.10340080829383272, "sharp": 0.47150031477212906, "r_mean": 0.39518492243119646, "r_std": 0.02440022587377046, "adv_abs": 0.021558061042534455, "frac_lowstd": 0.6680272200277874}
+{"time": "16:13:49", "event": "val", "step": 300, "rows_seen": 9600, "n": 500, "dropped_long": 0, "noul_ce": 5.107815265655518, "noul_T": 3.999999975766852, "noul_auc": 0.6497804795677136, "noul_acc": 0.68, "noul_f1": 0.7597597597597597, "noul_nll": 1.808052728241426, "noul_ece": 0.23847889530658722, "noul_brier": 0.26261958120450773, "noul_sharp": 0.4159337878227234, "entropy": 0.18219774961471558}
+{"time": "16:17:53", "event": "step", "step": 310, "rows_seen": 9920, "lr": 3.665625949658603e-05, "tok_per_s": 803.2710249517927, "loss": -0.007569627574783411, "entropy": 0.07659936918687622, "sharp": 0.47645895555615425, "r_mean": 0.026435966470411847, "r_std": 0.02814806130300391, "adv_abs": 0.025403001478740146, "frac_lowstd": 0.765986408506121}
+{"time": "16:21:49", "event": "step", "step": 320, "rows_seen": 10240, "lr": 3.379090157815887e-05, "tok_per_s": 805.6789636327768, "loss": -0.016026170749682933, "entropy": 0.16931198718198898, "sharp": 0.3965419754385948, "r_mean": 0.758103717650686, "r_std": 0.05220586562063545, "adv_abs": 0.04737243374776361, "frac_lowstd": 0.7863945705550057}
+{"time": "16:25:45", "event": "step", "step": 330, "rows_seen": 10560, "lr": 3.10040161203844e-05, "tok_per_s": 807.8345685978627, "loss": -0.017027361784130335, "entropy": 0.27022605086030466, "sharp": 0.3390402514487505, "r_mean": 0.1903770168622335, "r_std": 0.1266057783116897, "adv_abs": 0.11638866985837619, "frac_lowstd": 0.6073412795861562}
+{"time": "16:29:47", "event": "step", "step": 340, "rows_seen": 10880, "lr": 2.8307292289383776e-05, "tok_per_s": 810.1094387259834, "loss": 0.015920213772915304, "entropy": 0.15424613751181507, "sharp": 0.43892211467027664, "r_mean": 0.3835828974843025, "r_std": 0.04472087589757783, "adv_abs": 0.041113456445080895, "frac_lowstd": 0.7571428631033216}
+{"time": "16:33:45", "event": "step", "step": 350, "rows_seen": 11200, "lr": 2.571204108208194e-05, "tok_per_s": 812.2729421091227, "loss": -0.000223238836042583, "entropy": 0.18978856213373219, "sharp": 0.4031846523284912, "r_mean": 0.5291653592139482, "r_std": 0.05633466877043247, "adv_abs": 0.0476149981841445, "frac_lowstd": 0.7931547723710537}
+{"time": "16:37:39", "event": "step", "step": 360, "rows_seen": 11520, "lr": 2.322914788395595e-05, "tok_per_s": 814.1338536671688, "loss": -0.03515684534795582, "entropy": 0.25567242965146053, "sharp": 0.33540387637913227, "r_mean": 0.5728065414088113, "r_std": 0.11937233619391918, "adv_abs": 0.09927775524556637, "frac_lowstd": 0.5440476260014943}
+{"time": "16:41:33", "event": "step", "step": 370, "rows_seen": 11840, "lr": 2.0869026811945157e-05, "tok_per_s": 816.0723592622695, "loss": -0.04282050940673798, "entropy": 0.2759116512502162, "sharp": 0.3542269617319107, "r_mean": 0.5493344153676715, "r_std": 0.079061039856502, "adv_abs": 0.07046731827514512, "frac_lowstd": 0.6612245057310376}
+{"time": "16:45:30", "event": "step", "step": 380, "rows_seen": 12160, "lr": 1.8641577034025627e-05, "tok_per_s": 817.8491058020059, "loss": -0.017985909013077617, "entropy": 0.22087599349935605, "sharp": 0.3877222593873739, "r_mean": 0.412617452442646, "r_std": 0.08140920209033149, "adv_abs": 0.07382439715521676, "frac_lowstd": 0.6530612366540092}
+{"time": "16:49:27", "event": "step", "step": 390, "rows_seen": 12480, "lr": 1.6556141248659472e-05, "tok_per_s": 819.5500252366804, "loss": -0.05172500389016932, "entropy": 0.1787478996690111, "sharp": 0.40695178136229515, "r_mean": 0.693155859197889, "r_std": 0.050711712799966335, "adv_abs": 0.04353492179818984, "frac_lowstd": 0.7750000059604645}
+{"time": "16:53:26", "event": "step", "step": 400, "rows_seen": 12800, "lr": 1.4621466498270773e-05, "tok_per_s": 821.0560817583192, "loss": -0.027716718148440123, "entropy": 0.2309380701100302, "sharp": 0.3581426665186882, "r_mean": 0.7190770581364632, "r_std": 0.06966425757855177, "adv_abs": 0.06024910602718592, "frac_lowstd": 0.7520833369344473}
+{"time": "16:55:12", "event": "val", "step": 400, "rows_seen": 12800, "n": 500, "dropped_long": 0, "noul_ce": 2.9049057960510254, "noul_T": 3.9999996435739296, "noul_auc": 0.7026520201212251, "noul_acc": 0.71, "noul_f1": 0.8109517601043025, "noul_nll": 1.629098666259204, "noul_ece": 0.1890362129211426, "noul_brier": 0.2443055842543106, "noul_sharp": 0.3550748825073242, "entropy": 0.23144176602363586}
+{"time": "16:59:13", "event": "step", "step": 410, "rows_seen": 13120, "lr": 1.284566748110963e-05, "tok_per_s": 814.406409894994, "loss": -0.03614769794876338, "entropy": 0.16643379653557444, "sharp": 0.3998713120818138, "r_mean": 0.552335167924563, "r_std": 0.06953857963283856, "adv_abs": 0.06135301167766253, "frac_lowstd": 0.784127006928126}
+{"time": "17:03:08", "event": "step", "step": 420, "rows_seen": 13440, "lr": 1.1236192515387821e-05, "tok_per_s": 816.285432616545, "loss": 0.007989837788045406, "entropy": 0.25903365251146493, "sharp": 0.37447835970669985, "r_mean": 0.53512761422566, "r_std": 0.10292787317718778, "adv_abs": 0.08854631653853826, "frac_lowstd": 0.5598639505250114}
+{"time": "17:07:10", "event": "step", "step": 430, "rows_seen": 13760, "lr": 9.799792298443728e-06, "tok_per_s": 817.766170597754, "loss": -0.03441343091253657, "entropy": 0.1900832688893388, "sharp": 0.3822648860514164, "r_mean": 0.3533231882112367, "r_std": 0.09546010211176638, "adv_abs": 0.08237754531936455, "frac_lowstd": 0.5857143018926892}
+{"time": "17:11:02", "event": "step", "step": 440, "rows_seen": 14080, "lr": 8.542491591971554e-06, "tok_per_s": 819.2763991862423, "loss": 0.010488941334187984, "entropy": 0.21148950433139646, "sharp": 0.3791016414761543, "r_mean": 0.6031063782672087, "r_std": 0.07813772062460582, "adv_abs": 0.07014671713113785, "frac_lowstd": 0.6825396964947382}
+{"time": "17:15:02", "event": "step", "step": 450, "rows_seen": 14400, "lr": 7.469563952077155e-06, "tok_per_s": 820.6484615911579, "loss": -0.020251245417966857, "entropy": 0.2839726215951708, "sharp": 0.3468824680894613, "r_mean": 0.45927407698971884, "r_std": 0.09063564780600634, "adv_abs": 0.0833189386342253, "frac_lowstd": 0.5091836814369474}
+{"time": "17:18:59", "event": "step", "step": 460, "rows_seen": 14720, "lr": 6.585509610150972e-06, "tok_per_s": 822.0141432530062, "loss": -0.06382000492885709, "entropy": 0.1763061433685551, "sharp": 0.40359387546777725, "r_mean": 0.516355322940009, "r_std": 0.0662103822188718, "adv_abs": 0.05741172789462975, "frac_lowstd": 0.7149659991264343}
+{"time": "17:23:01", "event": "step", "step": 470, "rows_seen": 15040, "lr": 5.894036597333646e-06, "tok_per_s": 823.4066416450296, "loss": -0.017219372108229436, "entropy": 0.2730797070852873, "sharp": 0.3535568416118622, "r_mean": 0.6464930405574185, "r_std": 0.07881742822272438, "adv_abs": 0.07014063479644912, "frac_lowstd": 0.5612244989190783}
+{"time": "17:26:57", "event": "step", "step": 480, "rows_seen": 15360, "lr": 5.398045191744809e-06, "tok_per_s": 824.4867106048238, "loss": -0.0025347549880674336, "entropy": 0.3349777192155443, "sharp": 0.3215961568057537, "r_mean": 0.6051120460033417, "r_std": 0.1263539929091261, "adv_abs": 0.11071629769035714, "frac_lowstd": 0.5194444457689921}
+{"time": "17:30:56", "event": "step", "step": 490, "rows_seen": 15680, "lr": 5.099615753708697e-06, "tok_per_s": 825.7022178670427, "loss": -0.02235355350798507, "entropy": 0.19848686380890257, "sharp": 0.38790277391672134, "r_mean": 0.4845417338822569, "r_std": 0.05893349673332945, "adv_abs": 0.055278957308266694, "frac_lowstd": 0.7707483129841941}
+{"time": "17:34:56", "event": "step", "step": 500, "rows_seen": 16000, "lr": 5e-06, "tok_per_s": 826.8248367142727, "loss": -0.03278993806134167, "entropy": 0.2378368345678723, "sharp": 0.37316108494997025, "r_mean": 0.812812422712644, "r_std": 0.08386530583781375, "adv_abs": 0.07314288669446493, "frac_lowstd": 0.6476190586884817}
+{"time": "17:36:43", "event": "val", "step": 500, "rows_seen": 16000, "n": 500, "dropped_long": 0, "noul_ce": 1.5844616889953613, "noul_T": 3.999999975766852, "noul_auc": 0.7348246502781778, "noul_acc": 0.724, "noul_f1": 0.7952522255192879, "noul_nll": 1.208536469994171, "noul_ece": 0.1279940184354782, "noul_brier": 0.21053544312077907, "noul_sharp": 0.33431458473205566, "entropy": 0.2954961061477661}
+{"time": "17:43:44", "event": "ckpt", "tag": "ckpt-16000", "rows_seen": 16000, "replay_seen": 0, "step": 500, "tokens": 10544036, "tok_per_s": 794.1320291177743, "rows_per_s": 1.2050520749314817, "val": {"n": 2000, "dropped_long": 0, "noul_ce": 1.5079070329666138, "noul_T": 3.999999975766852, "noul_auc": 0.7524481645604246, "noul_acc": 0.736, "noul_f1": 0.8050221565731167, "noul_nll": 1.1806853275026115, "noul_ece": 0.12911880904436113, "noul_brier": 0.20666864844414723, "noul_sharp": 0.3424508273601532, "entropy": 0.2810675799846649}, "dropped_long": 0, "peak_gib": 18.236032962799072}