Upload task output 1

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,209 @@
+---
+base_model: None
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:/cache/models/Qwen--Qwen2.5-3B-Instruct
+- grpo
+- lora
+- transformers
+- trl
+---
+
+# 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 0.18.1

adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "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,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "up_proj",
+    "down_proj",
+    "o_proj",
+    "q_proj",
+    "v_proj",
+    "k_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c195cfd6ddd7d1c70e4ad509b9543b57c8623d87a0eed271cc25b593fc0bd100
+size 239536272

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

loss.txt

@@ -0,0 +1 @@
+50,0.47466856505721805

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,207 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,1267 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.0005,
+  "eval_steps": 500,
+  "global_step": 50,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14431.0,
+      "completions/max_terminated_length": 13090.0,
+      "completions/mean_length": 11799.78125,
+      "completions/mean_terminated_length": 7162.27294921875,
+      "completions/min_length": 235.0,
+      "completions/min_terminated_length": 235.0,
+      "entropy": 0.13893846329301596,
+      "epoch": 1e-05,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.7887574434280396,
+      "kl": 0.0,
+      "learning_rate": 0.0,
+      "loss": -0.0596,
+      "num_tokens": 404630.0,
+      "reward": -0.7717239856719971,
+      "reward_std": 0.42467162013053894,
+      "rewards/rollout_reward_func/mean": -0.7717239856719971,
+      "rewards/rollout_reward_func/std": 0.5858266353607178,
+      "sampling/importance_sampling_ratio/max": 1.7099345922470093,
+      "sampling/importance_sampling_ratio/mean": 0.9991271495819092,
+      "sampling/importance_sampling_ratio/min": 0.4281526207923889,
+      "sampling/sampling_logp_difference/max": 0.8482755422592163,
+      "sampling/sampling_logp_difference/mean": 0.012976451776921749,
+      "step": 1,
+      "step_time": 130.9108755219977
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "entropy": 0.13893846329301596,
+      "epoch": 2e-05,
+      "grad_norm": 1.804929494857788,
+      "kl": 0.0,
+      "learning_rate": 2.8571428571428573e-06,
+      "loss": -0.0596,
+      "step": 2,
+      "step_time": 44.13144713700058
+    },
+    {
+      "clip_ratio/high_max": 0.0022670181933790445,
+      "clip_ratio/high_mean": 0.0011335090966895223,
+      "clip_ratio/low_mean": 0.0029946123831905425,
+      "clip_ratio/low_min": 0.0005040322430431843,
+      "clip_ratio/region_mean": 0.004128121479880065,
+      "completions/clipped_ratio": 0.6875,
+      "completions/max_length": 14439.0,
+      "completions/max_terminated_length": 11449.0,
+      "completions/mean_length": 11617.5,
+      "completions/mean_terminated_length": 5924.5,
+      "completions/min_length": 323.0,
+      "completions/min_terminated_length": 323.0,
+      "entropy": 0.12676118151284754,
+      "epoch": 3e-05,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.4913136959075928,
+      "kl": 0.0009995185955631314,
+      "learning_rate": 5.7142857142857145e-06,
+      "loss": 0.0137,
+      "num_tokens": 803817.0,
+      "reward": -0.8155950307846069,
+      "reward_std": 0.4161786139011383,
+      "rewards/rollout_reward_func/mean": -0.8155950307846069,
+      "rewards/rollout_reward_func/std": 0.4885525703430176,
+      "sampling/importance_sampling_ratio/max": 2.0640246868133545,
+      "sampling/importance_sampling_ratio/mean": 0.9992372393608093,
+      "sampling/importance_sampling_ratio/min": 0.41666513681411743,
+      "sampling/sampling_logp_difference/max": 0.8754724264144897,
+      "sampling/sampling_logp_difference/mean": 0.011403773911297321,
+      "step": 3,
+      "step_time": 126.63972502899833
+    },
+    {
+      "clip_ratio/high_max": 0.005151324032340199,
+      "clip_ratio/high_mean": 0.0025756620161700994,
+      "clip_ratio/low_mean": 0.001626400655368343,
+      "clip_ratio/low_min": 0.0005040322430431843,
+      "clip_ratio/region_mean": 0.004202062671538442,
+      "entropy": 0.12616439966950566,
+      "epoch": 4e-05,
+      "grad_norm": 1.4003232717514038,
+      "kl": 0.001313611121076974,
+      "learning_rate": 8.571428571428573e-06,
+      "loss": 0.0134,
+      "step": 4,
+      "step_time": 44.62073808600053
+    },
+    {
+      "clip_ratio/high_max": 0.0016215514624491334,
+      "clip_ratio/high_mean": 0.0008107757312245667,
+      "clip_ratio/low_mean": 0.003547743574017659,
+      "clip_ratio/low_min": 0.0010995370685122907,
+      "clip_ratio/region_mean": 0.0043585193052422255,
+      "completions/clipped_ratio": 0.8125,
+      "completions/max_length": 14535.0,
+      "completions/max_terminated_length": 11037.0,
+      "completions/mean_length": 12972.78125,
+      "completions/mean_terminated_length": 7317.5,
+      "completions/min_length": 4349.0,
+      "completions/min_terminated_length": 4349.0,
+      "entropy": 0.1461638123728335,
+      "epoch": 5e-05,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 2.107853889465332,
+      "kl": 0.001256127497072157,
+      "learning_rate": 1.1428571428571429e-05,
+      "loss": 0.1588,
+      "num_tokens": 1245837.0,
+      "reward": -0.6968419551849365,
+      "reward_std": 0.4676758050918579,
+      "rewards/rollout_reward_func/mean": -0.6968419551849365,
+      "rewards/rollout_reward_func/std": 0.5711839199066162,
+      "sampling/importance_sampling_ratio/max": 3.0,
+      "sampling/importance_sampling_ratio/mean": 1.001021385192871,
+      "sampling/importance_sampling_ratio/min": 0.5821155905723572,
+      "sampling/sampling_logp_difference/max": 1.1031980514526367,
+      "sampling/sampling_logp_difference/mean": 0.01345288660377264,
+      "step": 5,
+      "step_time": 131.89395336500274
+    },
+    {
+      "clip_ratio/high_max": 0.0021937466226518154,
+      "clip_ratio/high_mean": 0.0010968733113259077,
+      "clip_ratio/low_mean": 0.0032513330806978047,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.004348206392023712,
+      "entropy": 0.14492363901808858,
+      "epoch": 6e-05,
+      "grad_norm": 1.6609207391738892,
+      "kl": 0.001644153301640472,
+      "learning_rate": 1.4285714285714285e-05,
+      "loss": 0.1578,
+      "step": 6,
+      "step_time": 46.381975520000196
+    },
+    {
+      "clip_ratio/high_max": 0.0031376838451251388,
+      "clip_ratio/high_mean": 0.0015688419225625694,
+      "clip_ratio/low_mean": 0.0033995709964074194,
+      "clip_ratio/low_min": 0.0015243901871144772,
+      "clip_ratio/region_mean": 0.004968412918969989,
+      "completions/clipped_ratio": 0.5625,
+      "completions/max_length": 14446.0,
+      "completions/max_terminated_length": 12650.0,
+      "completions/mean_length": 11378.28125,
+      "completions/mean_terminated_length": 7822.64306640625,
+      "completions/min_length": 1771.0,
+      "completions/min_terminated_length": 1771.0,
+      "entropy": 0.0922874582465738,
+      "epoch": 7e-05,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.8714547157287598,
+      "kl": 0.0019052933557759388,
+      "learning_rate": 1.7142857142857145e-05,
+      "loss": 0.3544,
+      "num_tokens": 1637608.0,
+      "reward": -0.4466933012008667,
+      "reward_std": 0.8536279201507568,
+      "rewards/rollout_reward_func/mean": -0.4466933012008667,
+      "rewards/rollout_reward_func/std": 0.8547493815422058,
+      "sampling/importance_sampling_ratio/max": 2.0912468433380127,
+      "sampling/importance_sampling_ratio/mean": 0.9991166591644287,
+      "sampling/importance_sampling_ratio/min": 0.48319441080093384,
+      "sampling/sampling_logp_difference/max": 0.7377604246139526,
+      "sampling/sampling_logp_difference/mean": 0.01108577847480774,
+      "step": 7,
+      "step_time": 126.40769893900142
+    },
+    {
+      "clip_ratio/high_max": 0.004281101922970265,
+      "clip_ratio/high_mean": 0.002941832965007052,
+      "clip_ratio/low_mean": 0.004934519121889025,
+      "clip_ratio/low_min": 0.001998567197006196,
+      "clip_ratio/region_mean": 0.007876352086896077,
+      "entropy": 0.08806704939343035,
+      "epoch": 8e-05,
+      "grad_norm": 1.634592890739441,
+      "kl": 0.003069787417189218,
+      "learning_rate": 2e-05,
+      "loss": 0.347,
+      "step": 8,
+      "step_time": 44.41796817000068
+    },
+    {
+      "clip_ratio/high_max": 0.0016580428346060216,
+      "clip_ratio/high_mean": 0.0008290214173030108,
+      "clip_ratio/low_mean": 0.00299244187772274,
+      "clip_ratio/low_min": 0.0006009615608491004,
+      "clip_ratio/region_mean": 0.003821463295025751,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14526.0,
+      "completions/max_terminated_length": 13961.0,
+      "completions/mean_length": 11837.40625,
+      "completions/mean_terminated_length": 7239.818359375,
+      "completions/min_length": 967.0,
+      "completions/min_terminated_length": 967.0,
+      "entropy": 0.10492791910655797,
+      "epoch": 9e-05,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.7168275117874146,
+      "kl": 0.004868662523222156,
+      "learning_rate": 2.2857142857142858e-05,
+      "loss": -0.0376,
+      "num_tokens": 2043692.0,
+      "reward": -0.6215145587921143,
+      "reward_std": 0.6429688930511475,
+      "rewards/rollout_reward_func/mean": -0.6215145587921143,
+      "rewards/rollout_reward_func/std": 0.7194304466247559,
+      "sampling/importance_sampling_ratio/max": 1.6615937948226929,
+      "sampling/importance_sampling_ratio/mean": 1.0002217292785645,
+      "sampling/importance_sampling_ratio/min": 0.47998860478401184,
+      "sampling/sampling_logp_difference/max": 0.7339929938316345,
+      "sampling/sampling_logp_difference/mean": 0.010960595682263374,
+      "step": 9,
+      "step_time": 124.44129867599804
+    },
+    {
+      "clip_ratio/high_max": 0.005017199146095663,
+      "clip_ratio/high_mean": 0.0025085995730478317,
+      "clip_ratio/low_mean": 0.00814974692184478,
+      "clip_ratio/low_min": 0.002315932302735746,
+      "clip_ratio/region_mean": 0.010658346494892612,
+      "entropy": 0.09843971265945584,
+      "epoch": 0.0001,
+      "grad_norm": 1.3552314043045044,
+      "kl": 0.01046636993123684,
+      "learning_rate": 2.5714285714285714e-05,
+      "loss": -0.0446,
+      "step": 10,
+      "step_time": 45.56444605700199
+    },
+    {
+      "clip_ratio/high_max": 0.0035993335768580437,
+      "clip_ratio/high_mean": 0.0017996667884290218,
+      "clip_ratio/low_mean": 0.003951315797166899,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.005750982585595921,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14459.0,
+      "completions/max_terminated_length": 13816.0,
+      "completions/mean_length": 11661.625,
+      "completions/mean_terminated_length": 6824.5458984375,
+      "completions/min_length": 339.0,
+      "completions/min_terminated_length": 339.0,
+      "entropy": 0.06328993663191795,
+      "epoch": 0.00011,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.7052122354507446,
+      "kl": 0.020232660688634496,
+      "learning_rate": 2.857142857142857e-05,
+      "loss": -0.0762,
+      "num_tokens": 2444128.0,
+      "reward": -0.7770951390266418,
+      "reward_std": 0.4045039117336273,
+      "rewards/rollout_reward_func/mean": -0.7770951390266418,
+      "rewards/rollout_reward_func/std": 0.4794606566429138,
+      "sampling/importance_sampling_ratio/max": 2.010491371154785,
+      "sampling/importance_sampling_ratio/mean": 0.9987582564353943,
+      "sampling/importance_sampling_ratio/min": 0.159025177359581,
+      "sampling/sampling_logp_difference/max": 1.8386926651000977,
+      "sampling/sampling_logp_difference/mean": 0.008627503179013729,
+      "step": 11,
+      "step_time": 121.11756841299757
+    },
+    {
+      "clip_ratio/high_max": 0.005686733347829431,
+      "clip_ratio/high_mean": 0.0028433666739147156,
+      "clip_ratio/low_mean": 0.012177724187495187,
+      "clip_ratio/low_min": 0.006776289606932551,
+      "clip_ratio/region_mean": 0.015021090890513733,
+      "entropy": 0.06035556551069021,
+      "epoch": 0.00012,
+      "grad_norm": 1.6369805335998535,
+      "kl": 0.06257348190410994,
+      "learning_rate": 3.142857142857143e-05,
+      "loss": -0.0747,
+      "step": 12,
+      "step_time": 44.37280418500268
+    },
+    {
+      "clip_ratio/high_max": 0.006585154333151877,
+      "clip_ratio/high_mean": 0.003569125838112086,
+      "clip_ratio/low_mean": 0.00462483472074382,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.008193960617063567,
+      "completions/clipped_ratio": 0.8125,
+      "completions/max_length": 14486.0,
+      "completions/max_terminated_length": 13435.0,
+      "completions/mean_length": 12772.375,
+      "completions/mean_terminated_length": 6521.1669921875,
+      "completions/min_length": 815.0,
+      "completions/min_terminated_length": 815.0,
+      "entropy": 0.09321935242041945,
+      "epoch": 0.00013,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.849172592163086,
+      "kl": 0.03665067162364721,
+      "learning_rate": 3.428571428571429e-05,
+      "loss": -0.1546,
+      "num_tokens": 2880635.0,
+      "reward": -0.927730917930603,
+      "reward_std": 0.10113543272018433,
+      "rewards/rollout_reward_func/mean": -0.927730917930603,
+      "rewards/rollout_reward_func/std": 0.10532300174236298,
+      "sampling/importance_sampling_ratio/max": 1.936368465423584,
+      "sampling/importance_sampling_ratio/mean": 0.9986484050750732,
+      "sampling/importance_sampling_ratio/min": 0.18729014694690704,
+      "sampling/sampling_logp_difference/max": 1.6750962734222412,
+      "sampling/sampling_logp_difference/mean": 0.01844533160328865,
+      "step": 13,
+      "step_time": 130.46155179900052
+    },
+    {
+      "clip_ratio/high_max": 0.007653064152691513,
+      "clip_ratio/high_mean": 0.0043634974281303585,
+      "clip_ratio/low_mean": 0.014142459142021835,
+      "clip_ratio/low_min": 0.0029499504598788917,
+      "clip_ratio/region_mean": 0.018505956570152193,
+      "entropy": 0.09064974845387042,
+      "epoch": 0.00014,
+      "grad_norm": 8.18207836151123,
+      "kl": 0.2905037718010135,
+      "learning_rate": 3.7142857142857143e-05,
+      "loss": -0.1459,
+      "step": 14,
+      "step_time": 44.158161802002724
+    },
+    {
+      "clip_ratio/high_max": 0.0027650215197354555,
+      "clip_ratio/high_mean": 0.0013825107598677278,
+      "clip_ratio/low_mean": 0.004234670544974506,
+      "clip_ratio/low_min": 0.0012649551499634981,
+      "clip_ratio/region_mean": 0.005617181275738403,
+      "completions/clipped_ratio": 0.8125,
+      "completions/max_length": 14400.0,
+      "completions/max_terminated_length": 9012.0,
+      "completions/mean_length": 12419.84375,
+      "completions/mean_terminated_length": 4916.33349609375,
+      "completions/min_length": 1128.0,
+      "completions/min_terminated_length": 1128.0,
+      "entropy": 0.07450041663832963,
+      "epoch": 0.00015,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 2.2030179500579834,
+      "kl": 0.07086092534882482,
+      "learning_rate": 4e-05,
+      "loss": 0.1292,
+      "num_tokens": 3305559.0,
+      "reward": -0.7803157567977905,
+      "reward_std": 0.42063942551612854,
+      "rewards/rollout_reward_func/mean": -0.7803157567977905,
+      "rewards/rollout_reward_func/std": 0.4826337397098541,
+      "sampling/importance_sampling_ratio/max": 2.0322628021240234,
+      "sampling/importance_sampling_ratio/mean": 0.9985103607177734,
+      "sampling/importance_sampling_ratio/min": 0.24719201028347015,
+      "sampling/sampling_logp_difference/max": 1.397589921951294,
+      "sampling/sampling_logp_difference/mean": 0.015935054048895836,
+      "step": 15,
+      "step_time": 129.98408898099842
+    },
+    {
+      "clip_ratio/high_max": 0.008143757819198072,
+      "clip_ratio/high_mean": 0.004071878909599036,
+      "clip_ratio/low_mean": 0.014440719707636163,
+      "clip_ratio/low_min": 0.0063073321944102645,
+      "clip_ratio/region_mean": 0.01851259867544286,
+      "entropy": 0.07352668652310967,
+      "epoch": 0.00016,
+      "grad_norm": 1.1360507011413574,
+      "kl": 0.08157501685491297,
+      "learning_rate": 4.2857142857142856e-05,
+      "loss": 0.1258,
+      "step": 16,
+      "step_time": 44.95600875300079
+    },
+    {
+      "clip_ratio/high_max": 0.000589622650295496,
+      "clip_ratio/high_mean": 0.000294811325147748,
+      "clip_ratio/low_mean": 0.004054940363857895,
+      "clip_ratio/low_min": 0.0010419560712762177,
+      "clip_ratio/region_mean": 0.004349751689005643,
+      "completions/clipped_ratio": 0.5625,
+      "completions/max_length": 14400.0,
+      "completions/max_terminated_length": 13862.0,
+      "completions/mean_length": 11858.40625,
+      "completions/mean_terminated_length": 8890.857421875,
+      "completions/min_length": 2159.0,
+      "completions/min_terminated_length": 2159.0,
+      "entropy": 0.07483385875821114,
+      "epoch": 0.00017,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.7257509231567383,
+      "kl": 0.05066429602447897,
+      "learning_rate": 4.5714285714285716e-05,
+      "loss": 0.334,
+      "num_tokens": 3712600.0,
+      "reward": -0.43021440505981445,
+      "reward_std": 0.8870621919631958,
+      "rewards/rollout_reward_func/mean": -0.43021440505981445,
+      "rewards/rollout_reward_func/std": 0.8449047803878784,
+      "sampling/importance_sampling_ratio/max": 2.4000132083892822,
+      "sampling/importance_sampling_ratio/mean": 1.002233624458313,
+      "sampling/importance_sampling_ratio/min": 0.1551634818315506,
+      "sampling/sampling_logp_difference/max": 1.8632760047912598,
+      "sampling/sampling_logp_difference/mean": 0.01753704622387886,
+      "step": 17,
+      "step_time": 132.86553408099644
+    },
+    {
+      "clip_ratio/high_max": 0.006706475804094225,
+      "clip_ratio/high_mean": 0.003931941610062495,
+      "clip_ratio/low_mean": 0.0171583968331106,
+      "clip_ratio/low_min": 0.010398818471003324,
+      "clip_ratio/region_mean": 0.021090338443173096,
+      "entropy": 0.06936636054888368,
+      "epoch": 0.00018,
+      "grad_norm": 1.771362543106079,
+      "kl": 0.10733210120815784,
+      "learning_rate": 4.8571428571428576e-05,
+      "loss": 0.3285,
+      "step": 18,
+      "step_time": 44.16366357000152
+    },
+    {
+      "clip_ratio/high_max": 0.0006443298771046102,
+      "clip_ratio/high_mean": 0.0003221649385523051,
+      "clip_ratio/low_mean": 0.0072666899068281054,
+      "clip_ratio/low_min": 0.0009964113123714924,
+      "clip_ratio/region_mean": 0.00758885481627658,
+      "completions/clipped_ratio": 0.625,
+      "completions/max_length": 14492.0,
+      "completions/max_terminated_length": 10697.0,
+      "completions/mean_length": 11052.375,
+      "completions/mean_terminated_length": 5732.83349609375,
+      "completions/min_length": 1322.0,
+      "completions/min_terminated_length": 1322.0,
+      "entropy": 0.06488334969617426,
+      "epoch": 0.00019,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.4038617610931396,
+      "kl": 0.08611430588643998,
+      "learning_rate": 5.142857142857143e-05,
+      "loss": 0.5711,
+      "num_tokens": 4093752.0,
+      "reward": -0.4095195233821869,
+      "reward_std": 0.8482590317726135,
+      "rewards/rollout_reward_func/mean": -0.4095195233821869,
+      "rewards/rollout_reward_func/std": 0.8331766724586487,
+      "sampling/importance_sampling_ratio/max": 2.3997995853424072,
+      "sampling/importance_sampling_ratio/mean": 0.9988876581192017,
+      "sampling/importance_sampling_ratio/min": 0.19376768171787262,
+      "sampling/sampling_logp_difference/max": 1.6410953998565674,
+      "sampling/sampling_logp_difference/mean": 0.016171647235751152,
+      "step": 19,
+      "step_time": 128.7513532999983
+    },
+    {
+      "clip_ratio/high_max": 0.009648803621530533,
+      "clip_ratio/high_mean": 0.004824401810765266,
+      "clip_ratio/low_mean": 0.014935461309505627,
+      "clip_ratio/low_min": 0.006239572481717914,
+      "clip_ratio/region_mean": 0.019759863178478554,
+      "entropy": 0.06792194582521915,
+      "epoch": 0.0002,
+      "grad_norm": 1.5674505233764648,
+      "kl": 0.07957532908767462,
+      "learning_rate": 5.428571428571428e-05,
+      "loss": 0.5691,
+      "step": 20,
+      "step_time": 44.17661613099699
+    },
+    {
+      "clip_ratio/high_max": 0.0010775862028822303,
+      "clip_ratio/high_mean": 0.0005387931014411151,
+      "clip_ratio/low_mean": 0.002510265534510836,
+      "clip_ratio/low_min": 0.0006067961221560836,
+      "clip_ratio/region_mean": 0.003049058635951951,
+      "completions/clipped_ratio": 0.53125,
+      "completions/max_length": 14388.0,
+      "completions/max_terminated_length": 13834.0,
+      "completions/mean_length": 11314.09375,
+      "completions/mean_terminated_length": 8011.1337890625,
+      "completions/min_length": 1359.0,
+      "completions/min_terminated_length": 1359.0,
+      "entropy": 0.06659986882004887,
+      "epoch": 0.00021,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.3160443305969238,
+      "kl": 0.04374655510764569,
+      "learning_rate": 5.714285714285714e-05,
+      "loss": 0.2579,
+      "num_tokens": 4483294.0,
+      "reward": -0.7569143772125244,
+      "reward_std": 0.38748466968536377,
+      "rewards/rollout_reward_func/mean": -0.7569143772125244,
+      "rewards/rollout_reward_func/std": 0.47438445687294006,
+      "sampling/importance_sampling_ratio/max": 2.5243661403656006,
+      "sampling/importance_sampling_ratio/mean": 1.0004966259002686,
+      "sampling/importance_sampling_ratio/min": 0.44545236229896545,
+      "sampling/sampling_logp_difference/max": 0.9259899854660034,
+      "sampling/sampling_logp_difference/mean": 0.011256169527769089,
+      "step": 21,
+      "step_time": 125.32715417199688
+    },
+    {
+      "clip_ratio/high_max": 0.007965195982251316,
+      "clip_ratio/high_mean": 0.004264129471266642,
+      "clip_ratio/low_mean": 0.01166028477018699,
+      "clip_ratio/low_min": 0.003762419626582414,
+      "clip_ratio/region_mean": 0.015924414212349802,
+      "entropy": 0.06479066435713321,
+      "epoch": 0.00022,
+      "grad_norm": 0.8557514548301697,
+      "kl": 0.0534607790177688,
+      "learning_rate": 6e-05,
+      "loss": 0.2505,
+      "step": 22,
+      "step_time": 44.035087997997834
+    },
+    {
+      "clip_ratio/high_max": 0.0011444397969171405,
+      "clip_ratio/high_mean": 0.0005722198984585702,
+      "clip_ratio/low_mean": 0.002795889973640442,
+      "clip_ratio/low_min": 0.0005040322430431843,
+      "clip_ratio/region_mean": 0.0033681098429951817,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14447.0,
+      "completions/max_terminated_length": 13549.0,
+      "completions/mean_length": 11808.34375,
+      "completions/mean_terminated_length": 7141.36376953125,
+      "completions/min_length": 784.0,
+      "completions/min_terminated_length": 784.0,
+      "entropy": 0.07080131443217397,
+      "epoch": 0.00023,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.2227104902267456,
+      "kl": 0.10710712906438857,
+      "learning_rate": 6.285714285714286e-05,
+      "loss": 0.357,
+      "num_tokens": 4888221.0,
+      "reward": -0.49963146448135376,
+      "reward_std": 0.8173227310180664,
+      "rewards/rollout_reward_func/mean": -0.49963146448135376,
+      "rewards/rollout_reward_func/std": 0.8128207921981812,
+      "sampling/importance_sampling_ratio/max": 1.9062927961349487,
+      "sampling/importance_sampling_ratio/mean": 1.0000483989715576,
+      "sampling/importance_sampling_ratio/min": 0.30003389716148376,
+      "sampling/sampling_logp_difference/max": 1.203859806060791,
+      "sampling/sampling_logp_difference/mean": 0.011820203624665737,
+      "step": 23,
+      "step_time": 134.11656286700236
+    },
+    {
+      "clip_ratio/high_max": 0.0063895476050674915,
+      "clip_ratio/high_mean": 0.0031947738025337458,
+      "clip_ratio/low_mean": 0.00737490743631497,
+      "clip_ratio/low_min": 0.0014960957923904061,
+      "clip_ratio/region_mean": 0.010569681267952546,
+      "entropy": 0.07403647690080106,
+      "epoch": 0.00024,
+      "grad_norm": 0.9927102327346802,
+      "kl": 0.07292677101213485,
+      "learning_rate": 6.571428571428571e-05,
+      "loss": 0.3453,
+      "step": 24,
+      "step_time": 44.88870704800138
+    },
+    {
+      "clip_ratio/high_max": 0.002575038291979581,
+      "clip_ratio/high_mean": 0.0012875191459897906,
+      "clip_ratio/low_mean": 0.0031292334315367043,
+      "clip_ratio/low_min": 0.0010516082984395325,
+      "clip_ratio/region_mean": 0.004416752519318834,
+      "completions/clipped_ratio": 0.6875,
+      "completions/max_length": 14441.0,
+      "completions/max_terminated_length": 13590.0,
+      "completions/mean_length": 11539.0625,
+      "completions/mean_terminated_length": 5616.10009765625,
+      "completions/min_length": 699.0,
+      "completions/min_terminated_length": 699.0,
+      "entropy": 0.09310276620090008,
+      "epoch": 0.00025,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.282138466835022,
+      "kl": 0.07473902264609933,
+      "learning_rate": 6.857142857142858e-05,
+      "loss": -0.0119,
+      "num_tokens": 5284116.0,
+      "reward": -0.7307593822479248,
+      "reward_std": 0.4555254876613617,
+      "rewards/rollout_reward_func/mean": -0.7307593822479248,
+      "rewards/rollout_reward_func/std": 0.5767623782157898,
+      "sampling/importance_sampling_ratio/max": 3.0,
+      "sampling/importance_sampling_ratio/mean": 1.0000531673431396,
+      "sampling/importance_sampling_ratio/min": 0.18337896466255188,
+      "sampling/sampling_logp_difference/max": 1.6962004899978638,
+      "sampling/sampling_logp_difference/mean": 0.014519629999995232,
+      "step": 25,
+      "step_time": 128.11358543400092
+    },
+    {
+      "clip_ratio/high_max": 0.006235250388272107,
+      "clip_ratio/high_mean": 0.0041256895638071,
+      "clip_ratio/low_mean": 0.019248695025453344,
+      "clip_ratio/low_min": 0.006461259792558849,
+      "clip_ratio/region_mean": 0.023374384531052783,
+      "entropy": 0.09371388074941933,
+      "epoch": 0.00026,
+      "grad_norm": 1.1407207250595093,
+      "kl": 0.10249935300089419,
+      "learning_rate": 7.142857142857143e-05,
+      "loss": -0.0179,
+      "step": 26,
+      "step_time": 44.78832636999505
+    },
+    {
+      "clip_ratio/high_max": 0.001961233268957585,
+      "clip_ratio/high_mean": 0.0009806166344787925,
+      "clip_ratio/low_mean": 0.0014239389565773308,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0024045555910561234,
+      "completions/clipped_ratio": 0.5,
+      "completions/max_length": 14387.0,
+      "completions/max_terminated_length": 13921.0,
+      "completions/mean_length": 10618.1875,
+      "completions/mean_terminated_length": 7001.0,
+      "completions/min_length": 1065.0,
+      "completions/min_terminated_length": 1065.0,
+      "entropy": 0.08063321514055133,
+      "epoch": 0.00027,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 3.7377917766571045,
+      "kl": 0.46857869252562523,
+      "learning_rate": 7.428571428571429e-05,
+      "loss": 0.0274,
+      "num_tokens": 5650556.0,
+      "reward": -0.4988600015640259,
+      "reward_std": 0.6857542991638184,
+      "rewards/rollout_reward_func/mean": -0.4988600015640259,
+      "rewards/rollout_reward_func/std": 0.8108665347099304,
+      "sampling/importance_sampling_ratio/max": 1.5844895839691162,
+      "sampling/importance_sampling_ratio/mean": 0.9988585710525513,
+      "sampling/importance_sampling_ratio/min": 0.14879053831100464,
+      "sampling/sampling_logp_difference/max": 1.9052157402038574,
+      "sampling/sampling_logp_difference/mean": 0.010426906868815422,
+      "step": 27,
+      "step_time": 117.60450818400022
+    },
+    {
+      "clip_ratio/high_max": 0.011021792015526444,
+      "clip_ratio/high_mean": 0.006221123243449256,
+      "clip_ratio/low_mean": 0.007159021100960672,
+      "clip_ratio/low_min": 0.0011492071789689362,
+      "clip_ratio/region_mean": 0.01338014446082525,
+      "entropy": 0.08160684490576386,
+      "epoch": 0.00028,
+      "grad_norm": 1.8652974367141724,
+      "kl": 0.0501517970697023,
+      "learning_rate": 7.714285714285715e-05,
+      "loss": 0.0131,
+      "step": 28,
+      "step_time": 43.69357772600051
+    },
+    {
+      "clip_ratio/high_max": 0.005480502266436815,
+      "clip_ratio/high_mean": 0.0027402511332184076,
+      "clip_ratio/low_mean": 0.0009808092145249248,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0037210604059509933,
+      "completions/clipped_ratio": 0.5,
+      "completions/max_length": 14524.0,
+      "completions/max_terminated_length": 13789.0,
+      "completions/mean_length": 10754.5,
+      "completions/mean_terminated_length": 7253.3125,
+      "completions/min_length": 1385.0,
+      "completions/min_terminated_length": 1385.0,
+      "entropy": 0.08766714809462428,
+      "epoch": 0.00029,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.9512283205986023,
+      "kl": 0.09233958297409117,
+      "learning_rate": 8e-05,
+      "loss": 0.3863,
+      "num_tokens": 6021181.0,
+      "reward": -0.43986791372299194,
+      "reward_std": 0.8378865718841553,
+      "rewards/rollout_reward_func/mean": -0.43986791372299194,
+      "rewards/rollout_reward_func/std": 0.8483731150627136,
+      "sampling/importance_sampling_ratio/max": 1.6999077796936035,
+      "sampling/importance_sampling_ratio/mean": 0.9990843534469604,
+      "sampling/importance_sampling_ratio/min": 0.4758984446525574,
+      "sampling/sampling_logp_difference/max": 0.7425508499145508,
+      "sampling/sampling_logp_difference/mean": 0.010583655908703804,
+      "step": 29,
+      "step_time": 122.50815704600063
+    },
+    {
+      "clip_ratio/high_max": 0.011935517424717546,
+      "clip_ratio/high_mean": 0.006662203115411103,
+      "clip_ratio/low_mean": 0.008739105775021017,
+      "clip_ratio/low_min": 0.004148350621107966,
+      "clip_ratio/region_mean": 0.01540130889043212,
+      "entropy": 0.09500867640599608,
+      "epoch": 0.0003,
+      "grad_norm": 0.7339725494384766,
+      "kl": 0.0838697346043773,
+      "learning_rate": 8.285714285714287e-05,
+      "loss": 0.3791,
+      "step": 30,
+      "step_time": 44.16028004800319
+    },
+    {
+      "clip_ratio/high_max": 0.002299328800290823,
+      "clip_ratio/high_mean": 0.0011496644001454115,
+      "clip_ratio/low_mean": 0.0013620644458569586,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.00251172884600237,
+      "completions/clipped_ratio": 0.71875,
+      "completions/max_length": 14550.0,
+      "completions/max_terminated_length": 13983.0,
+      "completions/mean_length": 12700.71875,
+      "completions/mean_terminated_length": 8789.888671875,
+      "completions/min_length": 3262.0,
+      "completions/min_terminated_length": 3262.0,
+      "entropy": 0.11977054877206683,
+      "epoch": 0.00031,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.086565613746643,
+      "kl": 0.05921647744253278,
+      "learning_rate": 8.571428571428571e-05,
+      "loss": 0.0947,
+      "num_tokens": 6454382.0,
+      "reward": -0.7849480509757996,
+      "reward_std": 0.38423284888267517,
+      "rewards/rollout_reward_func/mean": -0.7849480509757996,
+      "rewards/rollout_reward_func/std": 0.48005345463752747,
+      "sampling/importance_sampling_ratio/max": 1.8024439811706543,
+      "sampling/importance_sampling_ratio/mean": 1.000292181968689,
+      "sampling/importance_sampling_ratio/min": 0.5424980521202087,
+      "sampling/sampling_logp_difference/max": 0.6115708351135254,
+      "sampling/sampling_logp_difference/mean": 0.01138882152736187,
+      "step": 31,
+      "step_time": 127.06237872199927
+    },
+    {
+      "clip_ratio/high_max": 0.012531265849247575,
+      "clip_ratio/high_mean": 0.006265632924623787,
+      "clip_ratio/low_mean": 0.00913364501320757,
+      "clip_ratio/low_min": 0.0016989448922686279,
+      "clip_ratio/region_mean": 0.015399277937831357,
+      "entropy": 0.1310350843705237,
+      "epoch": 0.00032,
+      "grad_norm": 0.8760910034179688,
+      "kl": 0.06428600556682795,
+      "learning_rate": 8.857142857142857e-05,
+      "loss": 0.0848,
+      "step": 32,
+      "step_time": 45.15639938999993
+    },
+    {
+      "clip_ratio/high_max": 0.001622461248189211,
+      "clip_ratio/high_mean": 0.0008112306240946054,
+      "clip_ratio/low_mean": 0.0012673057353822514,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.002078536359476857,
+      "completions/clipped_ratio": 0.5,
+      "completions/max_length": 14485.0,
+      "completions/max_terminated_length": 13438.0,
+      "completions/mean_length": 11201.46875,
+      "completions/mean_terminated_length": 8163.75,
+      "completions/min_length": 1137.0,
+      "completions/min_terminated_length": 1137.0,
+      "entropy": 0.11091703665442765,
+      "epoch": 0.00033,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.763496994972229,
+      "kl": 0.08535079064313322,
+      "learning_rate": 9.142857142857143e-05,
+      "loss": 0.187,
+      "num_tokens": 6839240.0,
+      "reward": -0.606515109539032,
+      "reward_std": 0.644769549369812,
+      "rewards/rollout_reward_func/mean": -0.606515109539032,
+      "rewards/rollout_reward_func/std": 0.7157695293426514,
+      "sampling/importance_sampling_ratio/max": 1.9776629209518433,
+      "sampling/importance_sampling_ratio/mean": 1.0004677772521973,
+      "sampling/importance_sampling_ratio/min": 0.6363774538040161,
+      "sampling/sampling_logp_difference/max": 0.6819157600402832,
+      "sampling/sampling_logp_difference/mean": 0.010047465562820435,
+      "step": 33,
+      "step_time": 122.12401338600102
+    },
+    {
+      "clip_ratio/high_max": 0.003852536785416305,
+      "clip_ratio/high_mean": 0.0024514784454368055,
+      "clip_ratio/low_mean": 0.010322274029022083,
+      "clip_ratio/low_min": 0.0054294930887408555,
+      "clip_ratio/region_mean": 0.012773752474458888,
+      "entropy": 0.11235238378867507,
+      "epoch": 0.00034,
+      "grad_norm": 0.9141661524772644,
+      "kl": 0.08446309305145405,
+      "learning_rate": 9.428571428571429e-05,
+      "loss": 0.1825,
+      "step": 34,
+      "step_time": 43.25774885099963
+    },
+    {
+      "clip_ratio/high_max": 0.002067690948024392,
+      "clip_ratio/high_mean": 0.001033845474012196,
+      "clip_ratio/low_mean": 0.0013292695221025497,
+      "clip_ratio/low_min": 0.0005208333604969084,
+      "clip_ratio/region_mean": 0.0023631149961147457,
+      "completions/clipped_ratio": 0.625,
+      "completions/max_length": 14504.0,
+      "completions/max_terminated_length": 13750.0,
+      "completions/mean_length": 12662.125,
+      "completions/mean_terminated_length": 10072.6669921875,
+      "completions/min_length": 2136.0,
+      "completions/min_terminated_length": 2136.0,
+      "entropy": 0.13832575432024896,
+      "epoch": 0.00035,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.8794277906417847,
+      "kl": 0.08255716518033296,
+      "learning_rate": 9.714285714285715e-05,
+      "loss": 0.0863,
+      "num_tokens": 7271283.0,
+      "reward": -0.5410826206207275,
+      "reward_std": 0.7768348455429077,
+      "rewards/rollout_reward_func/mean": -0.5410826206207275,
+      "rewards/rollout_reward_func/std": 0.7597488760948181,
+      "sampling/importance_sampling_ratio/max": 1.5629594326019287,
+      "sampling/importance_sampling_ratio/mean": 1.0001167058944702,
+      "sampling/importance_sampling_ratio/min": 0.5588953495025635,
+      "sampling/sampling_logp_difference/max": 0.5817930698394775,
+      "sampling/sampling_logp_difference/mean": 0.009778056293725967,
+      "step": 35,
+      "step_time": 130.91799312700277
+    },
+    {
+      "clip_ratio/high_max": 0.011173836537636817,
+      "clip_ratio/high_mean": 0.005586918268818408,
+      "clip_ratio/low_mean": 0.00869341439101845,
+      "clip_ratio/low_min": 0.004238906840328127,
+      "clip_ratio/region_mean": 0.01428033280535601,
+      "entropy": 0.15124066965654492,
+      "epoch": 0.00036,
+      "grad_norm": 0.7353250980377197,
+      "kl": 0.08487725246232003,
+      "learning_rate": 0.0001,
+      "loss": 0.0769,
+      "step": 36,
+      "step_time": 45.46909211699858
+    },
+    {
+      "clip_ratio/high_max": 0.0015989625826478004,
+      "clip_ratio/high_mean": 0.0007994812913239002,
+      "clip_ratio/low_mean": 0.0020003379322588444,
+      "clip_ratio/low_min": 0.0005165289039723575,
+      "clip_ratio/region_mean": 0.0027998192235827446,
+      "completions/clipped_ratio": 0.625,
+      "completions/max_length": 14481.0,
+      "completions/max_terminated_length": 13367.0,
+      "completions/mean_length": 11635.15625,
+      "completions/mean_terminated_length": 7297.1669921875,
+      "completions/min_length": 439.0,
+      "completions/min_terminated_length": 439.0,
+      "entropy": 0.13707713410258293,
+      "epoch": 0.00037,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.8795878887176514,
+      "kl": 0.09341590362600982,
+      "learning_rate": 9.918053502751771e-05,
+      "loss": 0.1423,
+      "num_tokens": 7670326.0,
+      "reward": -0.6904894113540649,
+      "reward_std": 0.6018991470336914,
+      "rewards/rollout_reward_func/mean": -0.6904894113540649,
+      "rewards/rollout_reward_func/std": 0.6566587686538696,
+      "sampling/importance_sampling_ratio/max": 1.4667878150939941,
+      "sampling/importance_sampling_ratio/mean": 0.999151349067688,
+      "sampling/importance_sampling_ratio/min": 0.3191526234149933,
+      "sampling/sampling_logp_difference/max": 1.1420859098434448,
+      "sampling/sampling_logp_difference/mean": 0.01049021165817976,
+      "step": 37,
+      "step_time": 128.31693721999363
+    },
+    {
+      "clip_ratio/high_max": 0.012058351479936391,
+      "clip_ratio/high_mean": 0.007017562194960192,
+      "clip_ratio/low_mean": 0.010125949047505856,
+      "clip_ratio/low_min": 0.002969218941871077,
+      "clip_ratio/region_mean": 0.01714351132977754,
+      "entropy": 0.13681618124246597,
+      "epoch": 0.00038,
+      "grad_norm": 0.6959795951843262,
+      "kl": 0.09944733651354909,
+      "learning_rate": 9.675795466159755e-05,
+      "loss": 0.13,
+      "step": 38,
+      "step_time": 45.151127613000426
+    },
+    {
+      "clip_ratio/high_max": 0.001359933929052204,
+      "clip_ratio/high_mean": 0.000679966964526102,
+      "clip_ratio/low_mean": 0.001585537102073431,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0022655040957033634,
+      "completions/clipped_ratio": 0.59375,
+      "completions/max_length": 14393.0,
+      "completions/max_terminated_length": 13179.0,
+      "completions/mean_length": 11368.53125,
+      "completions/mean_terminated_length": 7292.30810546875,
+      "completions/min_length": 1023.0,
+      "completions/min_terminated_length": 1023.0,
+      "entropy": 0.11355275707319379,
+      "epoch": 0.00039,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.8372141122817993,
+      "kl": 0.12875043111853302,
+      "learning_rate": 9.283813728906054e-05,
+      "loss": -0.0014,
+      "num_tokens": 8061296.0,
+      "reward": -0.6022458076477051,
+      "reward_std": 0.6515352129936218,
+      "rewards/rollout_reward_func/mean": -0.6022458076477051,
+      "rewards/rollout_reward_func/std": 0.7095517516136169,
+      "sampling/importance_sampling_ratio/max": 1.636765480041504,
+      "sampling/importance_sampling_ratio/mean": 0.9989644289016724,
+      "sampling/importance_sampling_ratio/min": 0.47850245237350464,
+      "sampling/sampling_logp_difference/max": 0.7370939254760742,
+      "sampling/sampling_logp_difference/mean": 0.010086959227919579,
+      "step": 39,
+      "step_time": 126.83774206699854
+    },
+    {
+      "clip_ratio/high_max": 0.008263030322268605,
+      "clip_ratio/high_mean": 0.004131515161134303,
+      "clip_ratio/low_mean": 0.014172348921420053,
+      "clip_ratio/low_min": 0.0036224325885996222,
+      "clip_ratio/region_mean": 0.018303864169865847,
+      "entropy": 0.10010521637741476,
+      "epoch": 0.0004,
+      "grad_norm": 0.7000178098678589,
+      "kl": 0.14327447954565287,
+      "learning_rate": 8.759239773845718e-05,
+      "loss": -0.0115,
+      "step": 40,
+      "step_time": 44.916688461000376
+    },
+    {
+      "clip_ratio/high_max": 0.001846958533860743,
+      "clip_ratio/high_mean": 0.0009234792669303715,
+      "clip_ratio/low_mean": 0.002370692673139274,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0032941719400696456,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14506.0,
+      "completions/max_terminated_length": 13335.0,
+      "completions/mean_length": 11941.9375,
+      "completions/mean_terminated_length": 7638.63671875,
+      "completions/min_length": 1452.0,
+      "completions/min_terminated_length": 1452.0,
+      "entropy": 0.11166324256919324,
+      "epoch": 0.00041,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.0560439825057983,
+      "kl": 0.10882639326155186,
+      "learning_rate": 8.125000000000001e-05,
+      "loss": 0.5418,
+      "num_tokens": 8470213.0,
+      "reward": -0.4550766348838806,
+      "reward_std": 0.7941820621490479,
+      "rewards/rollout_reward_func/mean": -0.4550766348838806,
+      "rewards/rollout_reward_func/std": 0.7887347340583801,
+      "sampling/importance_sampling_ratio/max": 1.8251365423202515,
+      "sampling/importance_sampling_ratio/mean": 1.0004825592041016,
+      "sampling/importance_sampling_ratio/min": 0.5376880168914795,
+      "sampling/sampling_logp_difference/max": 0.6204767227172852,
+      "sampling/sampling_logp_difference/mean": 0.011826084926724434,
+      "step": 41,
+      "step_time": 126.00419519500247
+    },
+    {
+      "clip_ratio/high_max": 0.007851478818338364,
+      "clip_ratio/high_mean": 0.003925739409169182,
+      "clip_ratio/low_mean": 0.014391002681804821,
+      "clip_ratio/low_min": 0.005810233065858483,
+      "clip_ratio/region_mean": 0.018316742061870173,
+      "entropy": 0.10689377831295133,
+      "epoch": 0.00042,
+      "grad_norm": 1.3130351305007935,
+      "kl": 0.1170152414124459,
+      "learning_rate": 7.408813728906053e-05,
+      "loss": 0.5353,
+      "step": 42,
+      "step_time": 45.83305146299972
+    },
+    {
+      "clip_ratio/high_max": 0.002037170110270381,
+      "clip_ratio/high_mean": 0.0010185850551351905,
+      "clip_ratio/low_mean": 0.0016175404307432473,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0026361254567746073,
+      "completions/clipped_ratio": 0.53125,
+      "completions/max_length": 14370.0,
+      "completions/max_terminated_length": 13214.0,
+      "completions/mean_length": 11265.375,
+      "completions/mean_terminated_length": 7982.400390625,
+      "completions/min_length": 782.0,
+      "completions/min_terminated_length": 782.0,
+      "entropy": 0.09073930280283093,
+      "epoch": 0.00043,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.7020624876022339,
+      "kl": 0.13664298865478486,
+      "learning_rate": 6.6419817372537e-05,
+      "loss": 0.1856,
+      "num_tokens": 8857416.0,
+      "reward": -0.37660786509513855,
+      "reward_std": 0.8688997030258179,
+      "rewards/rollout_reward_func/mean": -0.37660786509513855,
+      "rewards/rollout_reward_func/std": 0.8781781196594238,
+      "sampling/importance_sampling_ratio/max": 1.9817417860031128,
+      "sampling/importance_sampling_ratio/mean": 1.0001003742218018,
+      "sampling/importance_sampling_ratio/min": 0.4231337010860443,
+      "sampling/sampling_logp_difference/max": 0.8600671291351318,
+      "sampling/sampling_logp_difference/mean": 0.009302067570388317,
+      "step": 43,
+      "step_time": 121.0720823820011
+    },
+    {
+      "clip_ratio/high_max": 0.007119149784557521,
+      "clip_ratio/high_mean": 0.0035595748922787607,
+      "clip_ratio/low_mean": 0.01176984136691317,
+      "clip_ratio/low_min": 0.002642020583152771,
+      "clip_ratio/region_mean": 0.01532941620098427,
+      "entropy": 0.08315122197382152,
+      "epoch": 0.00044,
+      "grad_norm": 5.769225597381592,
+      "kl": 0.6573639016132802,
+      "learning_rate": 5.858018262746301e-05,
+      "loss": 0.1936,
+      "step": 44,
+      "step_time": 44.69906600800459
+    },
+    {
+      "clip_ratio/high_max": 0.0023281591129489243,
+      "clip_ratio/high_mean": 0.0011640795564744622,
+      "clip_ratio/low_mean": 0.002193144930060953,
+      "clip_ratio/low_min": 0.0005530973430722952,
+      "clip_ratio/region_mean": 0.0033572244574315846,
+      "completions/clipped_ratio": 0.53125,
+      "completions/max_length": 14504.0,
+      "completions/max_terminated_length": 13747.0,
+      "completions/mean_length": 11228.96875,
+      "completions/mean_terminated_length": 7874.400390625,
+      "completions/min_length": 1631.0,
+      "completions/min_terminated_length": 1631.0,
+      "entropy": 0.08939703018404543,
+      "epoch": 0.00045,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.0448198318481445,
+      "kl": 0.11512639001011848,
+      "learning_rate": 5.0911862710939485e-05,
+      "loss": 0.0475,
+      "num_tokens": 9243782.0,
+      "reward": -0.49534285068511963,
+      "reward_std": 0.7042239904403687,
+      "rewards/rollout_reward_func/mean": -0.49534285068511963,
+      "rewards/rollout_reward_func/std": 0.8103145360946655,
+      "sampling/importance_sampling_ratio/max": 1.6891045570373535,
+      "sampling/importance_sampling_ratio/mean": 1.0011885166168213,
+      "sampling/importance_sampling_ratio/min": 0.5372036695480347,
+      "sampling/sampling_logp_difference/max": 0.6213779449462891,
+      "sampling/sampling_logp_difference/mean": 0.0091001121327281,
+      "step": 45,
+      "step_time": 120.37874379399727
+    },
+    {
+      "clip_ratio/high_max": 0.008119428472127765,
+      "clip_ratio/high_mean": 0.004059714236063883,
+      "clip_ratio/low_mean": 0.008602545305620879,
+      "clip_ratio/low_min": 0.0037590391584672034,
+      "clip_ratio/region_mean": 0.01266225945437327,
+      "entropy": 0.08823490096256137,
+      "epoch": 0.00046,
+      "grad_norm": 0.8318017721176147,
+      "kl": 0.120926967356354,
+      "learning_rate": 4.375000000000001e-05,
+      "loss": 0.0362,
+      "step": 46,
+      "step_time": 44.49746013800177
+    },
+    {
+      "clip_ratio/high_max": 0.003532699483912438,
+      "clip_ratio/high_mean": 0.001766349741956219,
+      "clip_ratio/low_mean": 0.0016052687860792503,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0033716185280354694,
+      "completions/clipped_ratio": 0.65625,
+      "completions/max_length": 14315.0,
+      "completions/max_terminated_length": 13481.0,
+      "completions/mean_length": 11687.5625,
+      "completions/mean_terminated_length": 7038.7275390625,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.07309431000612676,
+      "epoch": 0.00047,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 0.8768713474273682,
+      "kl": 0.08262627921067178,
+      "learning_rate": 3.740760226154283e-05,
+      "loss": 0.3828,
+      "num_tokens": 9644940.0,
+      "reward": -0.4843946695327759,
+      "reward_std": 0.8082762360572815,
+      "rewards/rollout_reward_func/mean": -0.4843946695327759,
+      "rewards/rollout_reward_func/std": 0.8041583895683289,
+      "sampling/importance_sampling_ratio/max": 1.4661790132522583,
+      "sampling/importance_sampling_ratio/mean": 1.000061273574829,
+      "sampling/importance_sampling_ratio/min": 0.3773066997528076,
+      "sampling/sampling_logp_difference/max": 0.9746968746185303,
+      "sampling/sampling_logp_difference/mean": 0.007087397389113903,
+      "step": 47,
+      "step_time": 121.44068199299727
+    },
+    {
+      "clip_ratio/high_max": 0.01053907914320007,
+      "clip_ratio/high_mean": 0.005612946202745661,
+      "clip_ratio/low_mean": 0.00564750301418826,
+      "clip_ratio/low_min": 0.0020543700957205147,
+      "clip_ratio/region_mean": 0.01126044918783009,
+      "entropy": 0.07405747356824577,
+      "epoch": 0.00048,
+      "grad_norm": 0.811284601688385,
+      "kl": 0.08346198999788612,
+      "learning_rate": 3.2161862710939476e-05,
+      "loss": 0.374,
+      "step": 48,
+      "step_time": 44.960870664997856
+    },
+    {
+      "clip_ratio/high_max": 0.002270299184601754,
+      "clip_ratio/high_mean": 0.001135149592300877,
+      "clip_ratio/low_mean": 0.003153837809804827,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.004288987343898043,
+      "completions/clipped_ratio": 0.625,
+      "completions/max_length": 14530.0,
+      "completions/max_terminated_length": 13232.0,
+      "completions/mean_length": 11548.625,
+      "completions/mean_terminated_length": 7039.5,
+      "completions/min_length": 1709.0,
+      "completions/min_terminated_length": 1709.0,
+      "entropy": 0.08768759481608868,
+      "epoch": 0.00049,
+      "frac_reward_zero_std": 0.0,
+      "grad_norm": 1.6150716543197632,
+      "kl": 0.3609654267784208,
+      "learning_rate": 2.8242045338402468e-05,
+      "loss": 0.4855,
+      "num_tokens": 10041134.0,
+      "reward": -0.5441310405731201,
+      "reward_std": 0.6545704007148743,
+      "rewards/rollout_reward_func/mean": -0.5441310405731201,
+      "rewards/rollout_reward_func/std": 0.7639203667640686,
+      "sampling/importance_sampling_ratio/max": 1.7008589506149292,
+      "sampling/importance_sampling_ratio/mean": 1.0008265972137451,
+      "sampling/importance_sampling_ratio/min": 0.45911118388175964,
+      "sampling/sampling_logp_difference/max": 0.7784628868103027,
+      "sampling/sampling_logp_difference/mean": 0.010290013626217842,
+      "step": 49,
+      "step_time": 124.03581126300378
+    },
+    {
+      "clip_ratio/high_max": 0.004471153952181339,
+      "clip_ratio/high_mean": 0.0022355769760906696,
+      "clip_ratio/low_mean": 0.007084499462507665,
+      "clip_ratio/low_min": 0.0021499572903849185,
+      "clip_ratio/region_mean": 0.009320076409494504,
+      "entropy": 0.08996679773554206,
+      "epoch": 0.0005,
+      "grad_norm": 0.8601014018058777,
+      "kl": 0.24557857040781528,
+      "learning_rate": 2.5819464972482288e-05,
+      "loss": 0.4747,
+      "step": 50,
+      "step_time": 45.77899998500652
+    },
+    {
+      "epoch": 0.0005,
+      "eval_clip_ratio/high_max": 0.0,
+      "eval_clip_ratio/high_mean": 0.0,
+      "eval_clip_ratio/low_mean": 0.0,
+      "eval_clip_ratio/low_min": 0.0,
+      "eval_clip_ratio/region_mean": 0.0,
+      "eval_completions/clipped_ratio": 0.60875,
+      "eval_completions/max_length": 13600.67375,
+      "eval_completions/max_terminated_length": 5307.3125,
+      "eval_completions/mean_length": 11798.84625,
+      "eval_completions/mean_terminated_length": 4963.2975,
+      "eval_completions/min_length": 9997.01875,
+      "eval_completions/min_terminated_length": 4619.2825,
+      "eval_entropy": 0.09010161134880036,
+      "eval_frac_reward_zero_std": 1.0,
+      "eval_kl": 0.13719033321132884,
+      "eval_loss": 0.00026262112078256905,
+      "eval_num_tokens": 10041134.0,
+      "eval_reward": -0.47466856505721805,
+      "eval_reward_std": 0.0,
+      "eval_rewards/rollout_reward_func/mean": -0.47466856505721805,
+      "eval_rewards/rollout_reward_func/std": 0.509479589665134,
+      "eval_runtime": 4806.419,
+      "eval_samples_per_second": 0.042,
+      "eval_sampling/importance_sampling_ratio/max": 1.2884997601807118,
+      "eval_sampling/importance_sampling_ratio/mean": 1.000060096681118,
+      "eval_sampling/importance_sampling_ratio/min": 0.7314270212873817,
+      "eval_sampling/sampling_logp_difference/max": 0.38731383727863433,
+      "eval_sampling/sampling_logp_difference/mean": 0.00951688139612088,
+      "eval_steps_per_second": 0.021,
+      "step": 50
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 50,
+  "num_input_tokens_seen": 10041134,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8adb550be405a8f1f15b5e74bdcc3049b647f9135632241d7f42aaad574cb8f5
+size 8145