Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- README.md +207 -0
- adapter_config.json +42 -0
- adapter_model.safetensors +3 -0
- added_tokens.json +28 -0
- chat_template.jinja +61 -0
- merges.txt +0 -0
- special_tokens_map.json +31 -0
- tokenizer.json +3 -0
- tokenizer_config.json +239 -0
- trainer_state.json +2034 -0
- training_args.bin +3 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3-4B-Instruct-2507
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- base_model:adapter:Qwen/Qwen3-4B-Instruct-2507
|
| 7 |
+
- lora
|
| 8 |
+
- transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Model ID
|
| 12 |
+
|
| 13 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Model Details
|
| 18 |
+
|
| 19 |
+
### Model Description
|
| 20 |
+
|
| 21 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
- **Developed by:** [More Information Needed]
|
| 26 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 27 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 28 |
+
- **Model type:** [More Information Needed]
|
| 29 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 30 |
+
- **License:** [More Information Needed]
|
| 31 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 32 |
+
|
| 33 |
+
### Model Sources [optional]
|
| 34 |
+
|
| 35 |
+
<!-- Provide the basic links for the model. -->
|
| 36 |
+
|
| 37 |
+
- **Repository:** [More Information Needed]
|
| 38 |
+
- **Paper [optional]:** [More Information Needed]
|
| 39 |
+
- **Demo [optional]:** [More Information Needed]
|
| 40 |
+
|
| 41 |
+
## Uses
|
| 42 |
+
|
| 43 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 44 |
+
|
| 45 |
+
### Direct Use
|
| 46 |
+
|
| 47 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 48 |
+
|
| 49 |
+
[More Information Needed]
|
| 50 |
+
|
| 51 |
+
### Downstream Use [optional]
|
| 52 |
+
|
| 53 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 54 |
+
|
| 55 |
+
[More Information Needed]
|
| 56 |
+
|
| 57 |
+
### Out-of-Scope Use
|
| 58 |
+
|
| 59 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 60 |
+
|
| 61 |
+
[More Information Needed]
|
| 62 |
+
|
| 63 |
+
## Bias, Risks, and Limitations
|
| 64 |
+
|
| 65 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 66 |
+
|
| 67 |
+
[More Information Needed]
|
| 68 |
+
|
| 69 |
+
### Recommendations
|
| 70 |
+
|
| 71 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 72 |
+
|
| 73 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 74 |
+
|
| 75 |
+
## How to Get Started with the Model
|
| 76 |
+
|
| 77 |
+
Use the code below to get started with the model.
|
| 78 |
+
|
| 79 |
+
[More Information Needed]
|
| 80 |
+
|
| 81 |
+
## Training Details
|
| 82 |
+
|
| 83 |
+
### Training Data
|
| 84 |
+
|
| 85 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 86 |
+
|
| 87 |
+
[More Information Needed]
|
| 88 |
+
|
| 89 |
+
### Training Procedure
|
| 90 |
+
|
| 91 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 92 |
+
|
| 93 |
+
#### Preprocessing [optional]
|
| 94 |
+
|
| 95 |
+
[More Information Needed]
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
#### Training Hyperparameters
|
| 99 |
+
|
| 100 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 101 |
+
|
| 102 |
+
#### Speeds, Sizes, Times [optional]
|
| 103 |
+
|
| 104 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 105 |
+
|
| 106 |
+
[More Information Needed]
|
| 107 |
+
|
| 108 |
+
## Evaluation
|
| 109 |
+
|
| 110 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 111 |
+
|
| 112 |
+
### Testing Data, Factors & Metrics
|
| 113 |
+
|
| 114 |
+
#### Testing Data
|
| 115 |
+
|
| 116 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 117 |
+
|
| 118 |
+
[More Information Needed]
|
| 119 |
+
|
| 120 |
+
#### Factors
|
| 121 |
+
|
| 122 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 123 |
+
|
| 124 |
+
[More Information Needed]
|
| 125 |
+
|
| 126 |
+
#### Metrics
|
| 127 |
+
|
| 128 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 129 |
+
|
| 130 |
+
[More Information Needed]
|
| 131 |
+
|
| 132 |
+
### Results
|
| 133 |
+
|
| 134 |
+
[More Information Needed]
|
| 135 |
+
|
| 136 |
+
#### Summary
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
|
| 140 |
+
## Model Examination [optional]
|
| 141 |
+
|
| 142 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 143 |
+
|
| 144 |
+
[More Information Needed]
|
| 145 |
+
|
| 146 |
+
## Environmental Impact
|
| 147 |
+
|
| 148 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 149 |
+
|
| 150 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 151 |
+
|
| 152 |
+
- **Hardware Type:** [More Information Needed]
|
| 153 |
+
- **Hours used:** [More Information Needed]
|
| 154 |
+
- **Cloud Provider:** [More Information Needed]
|
| 155 |
+
- **Compute Region:** [More Information Needed]
|
| 156 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 157 |
+
|
| 158 |
+
## Technical Specifications [optional]
|
| 159 |
+
|
| 160 |
+
### Model Architecture and Objective
|
| 161 |
+
|
| 162 |
+
[More Information Needed]
|
| 163 |
+
|
| 164 |
+
### Compute Infrastructure
|
| 165 |
+
|
| 166 |
+
[More Information Needed]
|
| 167 |
+
|
| 168 |
+
#### Hardware
|
| 169 |
+
|
| 170 |
+
[More Information Needed]
|
| 171 |
+
|
| 172 |
+
#### Software
|
| 173 |
+
|
| 174 |
+
[More Information Needed]
|
| 175 |
+
|
| 176 |
+
## Citation [optional]
|
| 177 |
+
|
| 178 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 179 |
+
|
| 180 |
+
**BibTeX:**
|
| 181 |
+
|
| 182 |
+
[More Information Needed]
|
| 183 |
+
|
| 184 |
+
**APA:**
|
| 185 |
+
|
| 186 |
+
[More Information Needed]
|
| 187 |
+
|
| 188 |
+
## Glossary [optional]
|
| 189 |
+
|
| 190 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 191 |
+
|
| 192 |
+
[More Information Needed]
|
| 193 |
+
|
| 194 |
+
## More Information [optional]
|
| 195 |
+
|
| 196 |
+
[More Information Needed]
|
| 197 |
+
|
| 198 |
+
## Model Card Authors [optional]
|
| 199 |
+
|
| 200 |
+
[More Information Needed]
|
| 201 |
+
|
| 202 |
+
## Model Card Contact
|
| 203 |
+
|
| 204 |
+
[More Information Needed]
|
| 205 |
+
### Framework versions
|
| 206 |
+
|
| 207 |
+
- PEFT 0.17.1
|
adapter_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "Qwen/Qwen3-4B-Instruct-2507",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.0,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"qalora_group_size": 16,
|
| 24 |
+
"r": 1,
|
| 25 |
+
"rank_pattern": {},
|
| 26 |
+
"revision": null,
|
| 27 |
+
"target_modules": [
|
| 28 |
+
"q_proj",
|
| 29 |
+
"down_proj",
|
| 30 |
+
"v_proj",
|
| 31 |
+
"up_proj",
|
| 32 |
+
"k_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"gate_proj"
|
| 35 |
+
],
|
| 36 |
+
"target_parameters": null,
|
| 37 |
+
"task_type": "CAUSAL_LM",
|
| 38 |
+
"trainable_token_indices": null,
|
| 39 |
+
"use_dora": false,
|
| 40 |
+
"use_qalora": false,
|
| 41 |
+
"use_rslora": false
|
| 42 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:febee3feb5f0c8e065e204cb15c426e336c7b73d04ff4408b69bd5c3fc5ea5ee
|
| 3 |
+
size 4194640
|
added_tokens.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</think>": 151668,
|
| 3 |
+
"</tool_call>": 151658,
|
| 4 |
+
"</tool_response>": 151666,
|
| 5 |
+
"<think>": 151667,
|
| 6 |
+
"<tool_call>": 151657,
|
| 7 |
+
"<tool_response>": 151665,
|
| 8 |
+
"<|box_end|>": 151649,
|
| 9 |
+
"<|box_start|>": 151648,
|
| 10 |
+
"<|endoftext|>": 151643,
|
| 11 |
+
"<|file_sep|>": 151664,
|
| 12 |
+
"<|fim_middle|>": 151660,
|
| 13 |
+
"<|fim_pad|>": 151662,
|
| 14 |
+
"<|fim_prefix|>": 151659,
|
| 15 |
+
"<|fim_suffix|>": 151661,
|
| 16 |
+
"<|im_end|>": 151645,
|
| 17 |
+
"<|im_start|>": 151644,
|
| 18 |
+
"<|image_pad|>": 151655,
|
| 19 |
+
"<|object_ref_end|>": 151647,
|
| 20 |
+
"<|object_ref_start|>": 151646,
|
| 21 |
+
"<|quad_end|>": 151651,
|
| 22 |
+
"<|quad_start|>": 151650,
|
| 23 |
+
"<|repo_name|>": 151663,
|
| 24 |
+
"<|video_pad|>": 151656,
|
| 25 |
+
"<|vision_end|>": 151653,
|
| 26 |
+
"<|vision_pad|>": 151654,
|
| 27 |
+
"<|vision_start|>": 151652
|
| 28 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# 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>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\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" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- for message in messages %}
|
| 18 |
+
{%- if message.content is string %}
|
| 19 |
+
{%- set content = message.content %}
|
| 20 |
+
{%- else %}
|
| 21 |
+
{%- set content = '' %}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 24 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 25 |
+
{%- elif message.role == "assistant" %}
|
| 26 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 27 |
+
{%- if message.tool_calls %}
|
| 28 |
+
{%- for tool_call in message.tool_calls %}
|
| 29 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 30 |
+
{{- '\n' }}
|
| 31 |
+
{%- endif %}
|
| 32 |
+
{%- if tool_call.function %}
|
| 33 |
+
{%- set tool_call = tool_call.function %}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 36 |
+
{{- tool_call.name }}
|
| 37 |
+
{{- '", "arguments": ' }}
|
| 38 |
+
{%- if tool_call.arguments is string %}
|
| 39 |
+
{{- tool_call.arguments }}
|
| 40 |
+
{%- else %}
|
| 41 |
+
{{- tool_call.arguments | tojson }}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{{- '}\n</tool_call>' }}
|
| 44 |
+
{%- endfor %}
|
| 45 |
+
{%- endif %}
|
| 46 |
+
{{- '<|im_end|>\n' }}
|
| 47 |
+
{%- elif message.role == "tool" %}
|
| 48 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 49 |
+
{{- '<|im_start|>user' }}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{{- '\n<tool_response>\n' }}
|
| 52 |
+
{{- content }}
|
| 53 |
+
{{- '\n</tool_response>' }}
|
| 54 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 55 |
+
{{- '<|im_end|>\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- endif %}
|
| 58 |
+
{%- endfor %}
|
| 59 |
+
{%- if add_generation_prompt %}
|
| 60 |
+
{{- '<|im_start|>assistant\n' }}
|
| 61 |
+
{%- endif %}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|im_start|>",
|
| 4 |
+
"<|im_end|>",
|
| 5 |
+
"<|object_ref_start|>",
|
| 6 |
+
"<|object_ref_end|>",
|
| 7 |
+
"<|box_start|>",
|
| 8 |
+
"<|box_end|>",
|
| 9 |
+
"<|quad_start|>",
|
| 10 |
+
"<|quad_end|>",
|
| 11 |
+
"<|vision_start|>",
|
| 12 |
+
"<|vision_end|>",
|
| 13 |
+
"<|vision_pad|>",
|
| 14 |
+
"<|image_pad|>",
|
| 15 |
+
"<|video_pad|>"
|
| 16 |
+
],
|
| 17 |
+
"eos_token": {
|
| 18 |
+
"content": "<|im_end|>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
},
|
| 24 |
+
"pad_token": {
|
| 25 |
+
"content": "<|endoftext|>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": false,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
}
|
| 31 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
|
| 3 |
+
size 11422654
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,239 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
}
|
| 213 |
+
},
|
| 214 |
+
"additional_special_tokens": [
|
| 215 |
+
"<|im_start|>",
|
| 216 |
+
"<|im_end|>",
|
| 217 |
+
"<|object_ref_start|>",
|
| 218 |
+
"<|object_ref_end|>",
|
| 219 |
+
"<|box_start|>",
|
| 220 |
+
"<|box_end|>",
|
| 221 |
+
"<|quad_start|>",
|
| 222 |
+
"<|quad_end|>",
|
| 223 |
+
"<|vision_start|>",
|
| 224 |
+
"<|vision_end|>",
|
| 225 |
+
"<|vision_pad|>",
|
| 226 |
+
"<|image_pad|>",
|
| 227 |
+
"<|video_pad|>"
|
| 228 |
+
],
|
| 229 |
+
"bos_token": null,
|
| 230 |
+
"clean_up_tokenization_spaces": false,
|
| 231 |
+
"eos_token": "<|im_end|>",
|
| 232 |
+
"errors": "replace",
|
| 233 |
+
"extra_special_tokens": {},
|
| 234 |
+
"model_max_length": 1010000,
|
| 235 |
+
"pad_token": "<|endoftext|>",
|
| 236 |
+
"split_special_tokens": false,
|
| 237 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 238 |
+
"unk_token": null
|
| 239 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2034 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.2,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 100,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"advantage/absmean": 0.0,
|
| 14 |
+
"entropy": 0.49213120341300964,
|
| 15 |
+
"epoch": 0.002,
|
| 16 |
+
"grad_norm": 0.0,
|
| 17 |
+
"importance_ratio": 0.9995924234390259,
|
| 18 |
+
"learning_rate": 0.0,
|
| 19 |
+
"loss": 0.0,
|
| 20 |
+
"mismatch_kl": 0.0013128521386533976,
|
| 21 |
+
"reward": 0.009999999776482582,
|
| 22 |
+
"reward/refusal_reward_func": 0.009999999776482582,
|
| 23 |
+
"reward/std": 0.0,
|
| 24 |
+
"step": 1,
|
| 25 |
+
"timing/generation_ms": 3254.1263923048973,
|
| 26 |
+
"timing/scoring_ms": 25275.689974427223,
|
| 27 |
+
"timing/total_ms": 28529.81636673212,
|
| 28 |
+
"tokens/completion": 196.125,
|
| 29 |
+
"tokens/masked_fraction": 0.0,
|
| 30 |
+
"wall_clock/generate_s": 37.95693778991699
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"advantage/absmean": 0.17499999701976776,
|
| 34 |
+
"entropy": 0.6319499611854553,
|
| 35 |
+
"epoch": 0.004,
|
| 36 |
+
"grad_norm": 0.21545597111492612,
|
| 37 |
+
"importance_ratio": 0.9992015957832336,
|
| 38 |
+
"learning_rate": 1e-05,
|
| 39 |
+
"loss": -0.0015,
|
| 40 |
+
"mismatch_kl": 0.0010142761748284101,
|
| 41 |
+
"reward": 0.7100000381469727,
|
| 42 |
+
"reward/refusal_reward_func": 0.7100000381469727,
|
| 43 |
+
"reward/std": 0.26457512378692627,
|
| 44 |
+
"step": 2,
|
| 45 |
+
"timing/generation_ms": 5644.344195723534,
|
| 46 |
+
"timing/scoring_ms": 31337.847255170345,
|
| 47 |
+
"timing/total_ms": 36982.19145089388,
|
| 48 |
+
"tokens/completion": 642.9375,
|
| 49 |
+
"tokens/masked_fraction": 0.0,
|
| 50 |
+
"wall_clock/generate_s": 92.69584393501282
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"advantage/absmean": 0.08203125,
|
| 54 |
+
"entropy": 0.4765586853027344,
|
| 55 |
+
"epoch": 0.006,
|
| 56 |
+
"grad_norm": 0.2892202194064109,
|
| 57 |
+
"importance_ratio": 1.000748872756958,
|
| 58 |
+
"learning_rate": 2e-05,
|
| 59 |
+
"loss": 0.0023,
|
| 60 |
+
"mismatch_kl": 0.001345986733213067,
|
| 61 |
+
"reward": 0.06593750417232513,
|
| 62 |
+
"reward/refusal_reward_func": 0.06593750417232513,
|
| 63 |
+
"reward/std": 0.15905697643756866,
|
| 64 |
+
"step": 3,
|
| 65 |
+
"timing/generation_ms": 2269.3659961223602,
|
| 66 |
+
"timing/scoring_ms": 26262.165516614914,
|
| 67 |
+
"timing/total_ms": 28531.531512737274,
|
| 68 |
+
"tokens/completion": 240.15625,
|
| 69 |
+
"tokens/masked_fraction": 0.0,
|
| 70 |
+
"wall_clock/generate_s": 143.94128561019897
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"advantage/absmean": 0.0237890612334013,
|
| 74 |
+
"entropy": 0.5391862988471985,
|
| 75 |
+
"epoch": 0.008,
|
| 76 |
+
"grad_norm": 0.03657768868171977,
|
| 77 |
+
"importance_ratio": 1.0003156661987305,
|
| 78 |
+
"learning_rate": 3e-05,
|
| 79 |
+
"loss": 0.0003,
|
| 80 |
+
"mismatch_kl": 0.0014281703624874353,
|
| 81 |
+
"reward": 0.02812499925494194,
|
| 82 |
+
"reward/refusal_reward_func": 0.02812499925494194,
|
| 83 |
+
"reward/std": 0.028986798599362373,
|
| 84 |
+
"step": 4,
|
| 85 |
+
"timing/generation_ms": 3297.7539896965027,
|
| 86 |
+
"timing/scoring_ms": 18404.439702630043,
|
| 87 |
+
"timing/total_ms": 21702.193692326546,
|
| 88 |
+
"tokens/completion": 328.96875,
|
| 89 |
+
"tokens/masked_fraction": 0.0,
|
| 90 |
+
"wall_clock/generate_s": 30.997375011444092
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"advantage/absmean": 0.37681642174720764,
|
| 94 |
+
"entropy": 0.6411616206169128,
|
| 95 |
+
"epoch": 0.01,
|
| 96 |
+
"grad_norm": 0.3743397229309058,
|
| 97 |
+
"importance_ratio": 1.0011004209518433,
|
| 98 |
+
"learning_rate": 4e-05,
|
| 99 |
+
"loss": -0.0173,
|
| 100 |
+
"mismatch_kl": 0.00182775326538831,
|
| 101 |
+
"reward": 0.37406250834465027,
|
| 102 |
+
"reward/refusal_reward_func": 0.37406250834465027,
|
| 103 |
+
"reward/std": 0.3808777630329132,
|
| 104 |
+
"step": 5,
|
| 105 |
+
"timing/generation_ms": 4316.511310636997,
|
| 106 |
+
"timing/scoring_ms": 31669.483192265034,
|
| 107 |
+
"timing/total_ms": 35985.99450290203,
|
| 108 |
+
"tokens/completion": 485.3125,
|
| 109 |
+
"tokens/masked_fraction": 0.0,
|
| 110 |
+
"wall_clock/generate_s": 72.83462309837341
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"advantage/absmean": 0.007929688319563866,
|
| 114 |
+
"entropy": 0.5038114786148071,
|
| 115 |
+
"epoch": 0.012,
|
| 116 |
+
"grad_norm": 0.02765669861934211,
|
| 117 |
+
"importance_ratio": 0.999994158744812,
|
| 118 |
+
"learning_rate": 5e-05,
|
| 119 |
+
"loss": -0.0013,
|
| 120 |
+
"mismatch_kl": 0.0035878715571016073,
|
| 121 |
+
"reward": 0.014374999329447746,
|
| 122 |
+
"reward/refusal_reward_func": 0.014374999329447746,
|
| 123 |
+
"reward/std": 0.015398356132209301,
|
| 124 |
+
"step": 6,
|
| 125 |
+
"timing/generation_ms": 3028.248645365238,
|
| 126 |
+
"timing/scoring_ms": 25548.00620675087,
|
| 127 |
+
"timing/total_ms": 28576.254852116108,
|
| 128 |
+
"tokens/completion": 334.125,
|
| 129 |
+
"tokens/masked_fraction": 0.0,
|
| 130 |
+
"wall_clock/generate_s": 36.018136739730835
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"advantage/absmean": 0.0,
|
| 134 |
+
"entropy": 0.29107579588890076,
|
| 135 |
+
"epoch": 0.014,
|
| 136 |
+
"grad_norm": 0.0,
|
| 137 |
+
"importance_ratio": 1.0008047819137573,
|
| 138 |
+
"learning_rate": 6e-05,
|
| 139 |
+
"loss": 0.0,
|
| 140 |
+
"mismatch_kl": 0.003742748638615012,
|
| 141 |
+
"reward": 0.009999999776482582,
|
| 142 |
+
"reward/refusal_reward_func": 0.009999999776482582,
|
| 143 |
+
"reward/std": 0.0,
|
| 144 |
+
"step": 7,
|
| 145 |
+
"timing/generation_ms": 1139.9633809924126,
|
| 146 |
+
"timing/scoring_ms": 19651.204399764538,
|
| 147 |
+
"timing/total_ms": 20791.16778075695,
|
| 148 |
+
"tokens/completion": 100.375,
|
| 149 |
+
"tokens/masked_fraction": 0.0,
|
| 150 |
+
"wall_clock/generate_s": 23.77466917037964
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"advantage/absmean": 0.006562499795109034,
|
| 154 |
+
"entropy": 0.3522435426712036,
|
| 155 |
+
"epoch": 0.016,
|
| 156 |
+
"grad_norm": 0.013866793274081895,
|
| 157 |
+
"importance_ratio": 1.0039043426513672,
|
| 158 |
+
"learning_rate": 7e-05,
|
| 159 |
+
"loss": 0.0029,
|
| 160 |
+
"mismatch_kl": 0.022596202790737152,
|
| 161 |
+
"reward": 0.013749999925494194,
|
| 162 |
+
"reward/refusal_reward_func": 0.013749999925494194,
|
| 163 |
+
"reward/std": 0.009921567514538765,
|
| 164 |
+
"step": 8,
|
| 165 |
+
"timing/generation_ms": 784.0555533766747,
|
| 166 |
+
"timing/scoring_ms": 19302.494660019875,
|
| 167 |
+
"timing/total_ms": 20086.55021339655,
|
| 168 |
+
"tokens/completion": 56.875,
|
| 169 |
+
"tokens/masked_fraction": 0.0,
|
| 170 |
+
"wall_clock/generate_s": 24.927189350128174
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"advantage/absmean": 0.17390625178813934,
|
| 174 |
+
"entropy": 0.8274978995323181,
|
| 175 |
+
"epoch": 0.018,
|
| 176 |
+
"grad_norm": 0.11705006346082461,
|
| 177 |
+
"importance_ratio": 1.0001081228256226,
|
| 178 |
+
"learning_rate": 8e-05,
|
| 179 |
+
"loss": 0.0021,
|
| 180 |
+
"mismatch_kl": 0.0024926774203777313,
|
| 181 |
+
"reward": 0.6775000095367432,
|
| 182 |
+
"reward/refusal_reward_func": 0.6775000095367432,
|
| 183 |
+
"reward/std": 0.24893523752689362,
|
| 184 |
+
"step": 9,
|
| 185 |
+
"timing/generation_ms": 9839.449286460876,
|
| 186 |
+
"timing/scoring_ms": 40201.816976070404,
|
| 187 |
+
"timing/total_ms": 50041.26626253128,
|
| 188 |
+
"tokens/completion": 1103.59375,
|
| 189 |
+
"tokens/masked_fraction": 0.0,
|
| 190 |
+
"wall_clock/generate_s": 94.76600408554077
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"advantage/absmean": 0.0035156249068677425,
|
| 194 |
+
"entropy": 0.24158422648906708,
|
| 195 |
+
"epoch": 0.02,
|
| 196 |
+
"grad_norm": 0.005773329550661251,
|
| 197 |
+
"importance_ratio": 0.9983059763908386,
|
| 198 |
+
"learning_rate": 9e-05,
|
| 199 |
+
"loss": 0.0009,
|
| 200 |
+
"mismatch_kl": 0.05829961970448494,
|
| 201 |
+
"reward": 0.011874999850988388,
|
| 202 |
+
"reward/refusal_reward_func": 0.011874999850988388,
|
| 203 |
+
"reward/std": 0.007261843420565128,
|
| 204 |
+
"step": 10,
|
| 205 |
+
"timing/generation_ms": 553.7310987710953,
|
| 206 |
+
"timing/scoring_ms": 18607.969902455807,
|
| 207 |
+
"timing/total_ms": 19161.701001226902,
|
| 208 |
+
"tokens/completion": 26.25,
|
| 209 |
+
"tokens/masked_fraction": 0.0,
|
| 210 |
+
"wall_clock/generate_s": 23.25968861579895
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"advantage/absmean": 0.027421876788139343,
|
| 214 |
+
"entropy": 0.5188111066818237,
|
| 215 |
+
"epoch": 0.022,
|
| 216 |
+
"grad_norm": 0.030197590581975912,
|
| 217 |
+
"importance_ratio": 0.9986535310745239,
|
| 218 |
+
"learning_rate": 0.0001,
|
| 219 |
+
"loss": -0.0035,
|
| 220 |
+
"mismatch_kl": 0.005903073586523533,
|
| 221 |
+
"reward": 0.026874996721744537,
|
| 222 |
+
"reward/refusal_reward_func": 0.026874996721744537,
|
| 223 |
+
"reward/std": 0.046599194407463074,
|
| 224 |
+
"step": 11,
|
| 225 |
+
"timing/generation_ms": 2835.02546697855,
|
| 226 |
+
"timing/scoring_ms": 24443.98508220911,
|
| 227 |
+
"timing/total_ms": 27279.01054918766,
|
| 228 |
+
"tokens/completion": 308.53125,
|
| 229 |
+
"tokens/masked_fraction": 0.0,
|
| 230 |
+
"wall_clock/generate_s": 145.36076593399048
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"advantage/absmean": 0.12355469167232513,
|
| 234 |
+
"entropy": 0.44281068444252014,
|
| 235 |
+
"epoch": 0.024,
|
| 236 |
+
"grad_norm": 0.42446378552193,
|
| 237 |
+
"importance_ratio": 0.9987350106239319,
|
| 238 |
+
"learning_rate": 0.0001,
|
| 239 |
+
"loss": -0.0423,
|
| 240 |
+
"mismatch_kl": 0.006952627561986446,
|
| 241 |
+
"reward": 0.09437499195337296,
|
| 242 |
+
"reward/refusal_reward_func": 0.09437499195337296,
|
| 243 |
+
"reward/std": 0.21499907970428467,
|
| 244 |
+
"step": 12,
|
| 245 |
+
"timing/generation_ms": 2469.3235754966736,
|
| 246 |
+
"timing/scoring_ms": 35024.46338534355,
|
| 247 |
+
"timing/total_ms": 37493.786960840225,
|
| 248 |
+
"tokens/completion": 273.28125,
|
| 249 |
+
"tokens/masked_fraction": 0.0,
|
| 250 |
+
"wall_clock/generate_s": 154.79029417037964
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"advantage/absmean": 0.09375,
|
| 254 |
+
"entropy": 0.5496101379394531,
|
| 255 |
+
"epoch": 0.026,
|
| 256 |
+
"grad_norm": 0.1257011193040082,
|
| 257 |
+
"importance_ratio": 0.9992591738700867,
|
| 258 |
+
"learning_rate": 0.0001,
|
| 259 |
+
"loss": -0.0035,
|
| 260 |
+
"mismatch_kl": 0.0034747051540762186,
|
| 261 |
+
"reward": 0.7599999904632568,
|
| 262 |
+
"reward/refusal_reward_func": 0.7599999904632568,
|
| 263 |
+
"reward/std": 0.19364915788173676,
|
| 264 |
+
"step": 13,
|
| 265 |
+
"timing/generation_ms": 8158.617563545704,
|
| 266 |
+
"timing/scoring_ms": 36215.20960330963,
|
| 267 |
+
"timing/total_ms": 44373.827166855335,
|
| 268 |
+
"tokens/completion": 930.6875,
|
| 269 |
+
"tokens/masked_fraction": 0.0,
|
| 270 |
+
"wall_clock/generate_s": 122.23734498023987
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"advantage/absmean": 0.1844140589237213,
|
| 274 |
+
"entropy": 0.6212608814239502,
|
| 275 |
+
"epoch": 0.028,
|
| 276 |
+
"grad_norm": 0.26809699141136345,
|
| 277 |
+
"importance_ratio": 1.0004856586456299,
|
| 278 |
+
"learning_rate": 0.0001,
|
| 279 |
+
"loss": -0.0125,
|
| 280 |
+
"mismatch_kl": 0.0054090130142867565,
|
| 281 |
+
"reward": 0.1978124976158142,
|
| 282 |
+
"reward/refusal_reward_func": 0.1978124976158142,
|
| 283 |
+
"reward/std": 0.21153803169727325,
|
| 284 |
+
"step": 14,
|
| 285 |
+
"timing/generation_ms": 3771.018899977207,
|
| 286 |
+
"timing/scoring_ms": 32557.9876229167,
|
| 287 |
+
"timing/total_ms": 36329.006522893906,
|
| 288 |
+
"tokens/completion": 423.1875,
|
| 289 |
+
"tokens/masked_fraction": 0.0,
|
| 290 |
+
"wall_clock/generate_s": 76.63339400291443
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"advantage/absmean": 0.09375,
|
| 294 |
+
"entropy": 0.5119910836219788,
|
| 295 |
+
"epoch": 0.03,
|
| 296 |
+
"grad_norm": 0.2293179923709544,
|
| 297 |
+
"importance_ratio": 1.0004829168319702,
|
| 298 |
+
"learning_rate": 0.0001,
|
| 299 |
+
"loss": 0.0032,
|
| 300 |
+
"mismatch_kl": 0.0062421830371022224,
|
| 301 |
+
"reward": 0.05999999865889549,
|
| 302 |
+
"reward/refusal_reward_func": 0.05999999865889549,
|
| 303 |
+
"reward/std": 0.19364915788173676,
|
| 304 |
+
"step": 15,
|
| 305 |
+
"timing/generation_ms": 3333.340108394623,
|
| 306 |
+
"timing/scoring_ms": 27093.5076251626,
|
| 307 |
+
"timing/total_ms": 30426.847733557224,
|
| 308 |
+
"tokens/completion": 374.25,
|
| 309 |
+
"tokens/masked_fraction": 0.0,
|
| 310 |
+
"wall_clock/generate_s": 143.25073266029358
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"advantage/absmean": 0.023906249552965164,
|
| 314 |
+
"entropy": 0.7846541404724121,
|
| 315 |
+
"epoch": 0.032,
|
| 316 |
+
"grad_norm": 0.03020597167190001,
|
| 317 |
+
"importance_ratio": 0.9995452165603638,
|
| 318 |
+
"learning_rate": 0.0001,
|
| 319 |
+
"loss": 0.0001,
|
| 320 |
+
"mismatch_kl": 0.006196495145559311,
|
| 321 |
+
"reward": 0.04218750074505806,
|
| 322 |
+
"reward/refusal_reward_func": 0.04218750074505806,
|
| 323 |
+
"reward/std": 0.030489176511764526,
|
| 324 |
+
"step": 16,
|
| 325 |
+
"timing/generation_ms": 6338.90475332737,
|
| 326 |
+
"timing/scoring_ms": 35081.99892938137,
|
| 327 |
+
"timing/total_ms": 41420.90368270874,
|
| 328 |
+
"tokens/completion": 725.5625,
|
| 329 |
+
"tokens/masked_fraction": 0.0,
|
| 330 |
+
"wall_clock/generate_s": 89.63312363624573
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"advantage/absmean": 0.2109375,
|
| 334 |
+
"entropy": 0.4431617856025696,
|
| 335 |
+
"epoch": 0.034,
|
| 336 |
+
"grad_norm": 0.11997986504420295,
|
| 337 |
+
"importance_ratio": 1.0006847381591797,
|
| 338 |
+
"learning_rate": 0.0001,
|
| 339 |
+
"loss": -0.0079,
|
| 340 |
+
"mismatch_kl": 0.0046984292566776276,
|
| 341 |
+
"reward": 0.6850000023841858,
|
| 342 |
+
"reward/refusal_reward_func": 0.6850000023841858,
|
| 343 |
+
"reward/std": 0.2904737591743469,
|
| 344 |
+
"step": 17,
|
| 345 |
+
"timing/generation_ms": 4871.280819177628,
|
| 346 |
+
"timing/scoring_ms": 27726.198948919773,
|
| 347 |
+
"timing/total_ms": 32597.4797680974,
|
| 348 |
+
"tokens/completion": 541.3125,
|
| 349 |
+
"tokens/masked_fraction": 0.0,
|
| 350 |
+
"wall_clock/generate_s": 72.54996109008789
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"advantage/absmean": 0.020624998956918716,
|
| 354 |
+
"entropy": 0.005482906475663185,
|
| 355 |
+
"epoch": 0.036,
|
| 356 |
+
"grad_norm": 2.4822412849126587e-05,
|
| 357 |
+
"importance_ratio": 0.9996475577354431,
|
| 358 |
+
"learning_rate": 0.0001,
|
| 359 |
+
"loss": 0.0,
|
| 360 |
+
"mismatch_kl": 3.828452292964357e-07,
|
| 361 |
+
"reward": 0.023749999701976776,
|
| 362 |
+
"reward/refusal_reward_func": 0.023749999701976776,
|
| 363 |
+
"reward/std": 0.026896795257925987,
|
| 364 |
+
"step": 18,
|
| 365 |
+
"timing/generation_ms": 466.4832055568695,
|
| 366 |
+
"timing/scoring_ms": 18893.90940964222,
|
| 367 |
+
"timing/total_ms": 19360.39261519909,
|
| 368 |
+
"tokens/completion": 13.0,
|
| 369 |
+
"tokens/masked_fraction": 0.0,
|
| 370 |
+
"wall_clock/generate_s": 24.94274139404297
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"advantage/absmean": 0.06175781413912773,
|
| 374 |
+
"entropy": 0.5270810723304749,
|
| 375 |
+
"epoch": 0.038,
|
| 376 |
+
"grad_norm": 0.19852274538781398,
|
| 377 |
+
"importance_ratio": 1.0000227689743042,
|
| 378 |
+
"learning_rate": 0.0001,
|
| 379 |
+
"loss": 0.0024,
|
| 380 |
+
"mismatch_kl": 0.005923233926296234,
|
| 381 |
+
"reward": 0.06562499701976776,
|
| 382 |
+
"reward/refusal_reward_func": 0.06562499701976776,
|
| 383 |
+
"reward/std": 0.13811086118221283,
|
| 384 |
+
"step": 19,
|
| 385 |
+
"timing/generation_ms": 4183.5604682564735,
|
| 386 |
+
"timing/scoring_ms": 30921.414978802204,
|
| 387 |
+
"timing/total_ms": 35104.97544705868,
|
| 388 |
+
"tokens/completion": 476.46875,
|
| 389 |
+
"tokens/masked_fraction": 0.0,
|
| 390 |
+
"wall_clock/generate_s": 69.38085293769836
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"advantage/absmean": 0.0018164062639698386,
|
| 394 |
+
"entropy": 0.004832141101360321,
|
| 395 |
+
"epoch": 0.04,
|
| 396 |
+
"grad_norm": 0.00028270733132418016,
|
| 397 |
+
"importance_ratio": 0.9995356202125549,
|
| 398 |
+
"learning_rate": 0.0001,
|
| 399 |
+
"loss": -0.0,
|
| 400 |
+
"mismatch_kl": 1.2251906582605443e-06,
|
| 401 |
+
"reward": 0.010937499813735485,
|
| 402 |
+
"reward/refusal_reward_func": 0.010937499813735485,
|
| 403 |
+
"reward/std": 0.005219778511673212,
|
| 404 |
+
"step": 20,
|
| 405 |
+
"timing/generation_ms": 469.4804549217224,
|
| 406 |
+
"timing/scoring_ms": 18075.80190896988,
|
| 407 |
+
"timing/total_ms": 18545.2823638916,
|
| 408 |
+
"tokens/completion": 13.0,
|
| 409 |
+
"tokens/masked_fraction": 0.0,
|
| 410 |
+
"wall_clock/generate_s": 22.75140905380249
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"advantage/absmean": 0.13593749701976776,
|
| 414 |
+
"entropy": 0.6113055944442749,
|
| 415 |
+
"epoch": 0.042,
|
| 416 |
+
"grad_norm": 0.14856988549338257,
|
| 417 |
+
"importance_ratio": 0.999966561794281,
|
| 418 |
+
"learning_rate": 0.0001,
|
| 419 |
+
"loss": -0.0022,
|
| 420 |
+
"mismatch_kl": 0.005093984771519899,
|
| 421 |
+
"reward": 0.7350000143051147,
|
| 422 |
+
"reward/refusal_reward_func": 0.7350000143051147,
|
| 423 |
+
"reward/std": 0.23318447172641754,
|
| 424 |
+
"step": 21,
|
| 425 |
+
"timing/generation_ms": 6330.163478851318,
|
| 426 |
+
"timing/scoring_ms": 28551.29039287567,
|
| 427 |
+
"timing/total_ms": 34881.45387172699,
|
| 428 |
+
"tokens/completion": 715.9375,
|
| 429 |
+
"tokens/masked_fraction": 0.0,
|
| 430 |
+
"wall_clock/generate_s": 84.07915687561035
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"advantage/absmean": 0.01318359375,
|
| 434 |
+
"entropy": 0.5851391553878784,
|
| 435 |
+
"epoch": 0.044,
|
| 436 |
+
"grad_norm": 0.040096615934913975,
|
| 437 |
+
"importance_ratio": 1.0004969835281372,
|
| 438 |
+
"learning_rate": 0.0001,
|
| 439 |
+
"loss": -0.0018,
|
| 440 |
+
"mismatch_kl": 0.009555388242006302,
|
| 441 |
+
"reward": 0.017812499776482582,
|
| 442 |
+
"reward/refusal_reward_func": 0.017812499776482582,
|
| 443 |
+
"reward/std": 0.020575225353240967,
|
| 444 |
+
"step": 22,
|
| 445 |
+
"timing/generation_ms": 2347.38065302372,
|
| 446 |
+
"timing/scoring_ms": 26964.800156652927,
|
| 447 |
+
"timing/total_ms": 29312.180809676647,
|
| 448 |
+
"tokens/completion": 255.15625,
|
| 449 |
+
"tokens/masked_fraction": 0.0,
|
| 450 |
+
"wall_clock/generate_s": 150.09117031097412
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"advantage/absmean": 0.01718750037252903,
|
| 454 |
+
"entropy": 0.5837696194648743,
|
| 455 |
+
"epoch": 0.046,
|
| 456 |
+
"grad_norm": 0.027639162796687714,
|
| 457 |
+
"importance_ratio": 1.0036453008651733,
|
| 458 |
+
"learning_rate": 0.0001,
|
| 459 |
+
"loss": -0.0037,
|
| 460 |
+
"mismatch_kl": 0.017472539097070694,
|
| 461 |
+
"reward": 0.022499999031424522,
|
| 462 |
+
"reward/refusal_reward_func": 0.022499999031424522,
|
| 463 |
+
"reward/std": 0.021650634706020355,
|
| 464 |
+
"step": 23,
|
| 465 |
+
"timing/generation_ms": 2000.5059093236923,
|
| 466 |
+
"timing/scoring_ms": 24497.070513665676,
|
| 467 |
+
"timing/total_ms": 26497.57642298937,
|
| 468 |
+
"tokens/completion": 222.25,
|
| 469 |
+
"tokens/masked_fraction": 0.0,
|
| 470 |
+
"wall_clock/generate_s": 39.751105070114136
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"advantage/absmean": 0.09375,
|
| 474 |
+
"entropy": 0.6268512010574341,
|
| 475 |
+
"epoch": 0.048,
|
| 476 |
+
"grad_norm": 0.18640619697968583,
|
| 477 |
+
"importance_ratio": 0.9990081787109375,
|
| 478 |
+
"learning_rate": 0.0001,
|
| 479 |
+
"loss": -0.0019,
|
| 480 |
+
"mismatch_kl": 0.005103914998471737,
|
| 481 |
+
"reward": 0.7599999904632568,
|
| 482 |
+
"reward/refusal_reward_func": 0.7599999904632568,
|
| 483 |
+
"reward/std": 0.19364915788173676,
|
| 484 |
+
"step": 24,
|
| 485 |
+
"timing/generation_ms": 5919.098302721977,
|
| 486 |
+
"timing/scoring_ms": 32965.33615142107,
|
| 487 |
+
"timing/total_ms": 38884.43445414305,
|
| 488 |
+
"tokens/completion": 674.9375,
|
| 489 |
+
"tokens/masked_fraction": 0.0,
|
| 490 |
+
"wall_clock/generate_s": 105.22681593894958
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"advantage/absmean": 0.007929687388241291,
|
| 494 |
+
"entropy": 0.657754123210907,
|
| 495 |
+
"epoch": 0.05,
|
| 496 |
+
"grad_norm": 0.025336375406596286,
|
| 497 |
+
"importance_ratio": 1.0006119012832642,
|
| 498 |
+
"learning_rate": 0.0001,
|
| 499 |
+
"loss": -0.0006,
|
| 500 |
+
"mismatch_kl": 0.006333181634545326,
|
| 501 |
+
"reward": 0.014374999329447746,
|
| 502 |
+
"reward/refusal_reward_func": 0.014374999329447746,
|
| 503 |
+
"reward/std": 0.015398357063531876,
|
| 504 |
+
"step": 25,
|
| 505 |
+
"timing/generation_ms": 4377.142012119293,
|
| 506 |
+
"timing/scoring_ms": 25255.830891430378,
|
| 507 |
+
"timing/total_ms": 29632.97290354967,
|
| 508 |
+
"tokens/completion": 501.125,
|
| 509 |
+
"tokens/masked_fraction": 0.0,
|
| 510 |
+
"wall_clock/generate_s": 47.89186096191406
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"advantage/absmean": 0.0,
|
| 514 |
+
"entropy": 0.17220205068588257,
|
| 515 |
+
"epoch": 0.052,
|
| 516 |
+
"grad_norm": 0.0,
|
| 517 |
+
"importance_ratio": 1.0002461671829224,
|
| 518 |
+
"learning_rate": 0.0001,
|
| 519 |
+
"loss": 0.0,
|
| 520 |
+
"mismatch_kl": 0.0026462471578270197,
|
| 521 |
+
"reward": 0.009999999776482582,
|
| 522 |
+
"reward/refusal_reward_func": 0.009999999776482582,
|
| 523 |
+
"reward/std": 0.0,
|
| 524 |
+
"step": 26,
|
| 525 |
+
"timing/generation_ms": 497.48579412698746,
|
| 526 |
+
"timing/scoring_ms": 18527.822844684124,
|
| 527 |
+
"timing/total_ms": 19025.30863881111,
|
| 528 |
+
"tokens/completion": 19.40625,
|
| 529 |
+
"tokens/masked_fraction": 0.0,
|
| 530 |
+
"wall_clock/generate_s": 22.794724941253662
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"advantage/absmean": 0.17499999701976776,
|
| 534 |
+
"entropy": 0.546366810798645,
|
| 535 |
+
"epoch": 0.054,
|
| 536 |
+
"grad_norm": 0.22199732245019957,
|
| 537 |
+
"importance_ratio": 0.999916672706604,
|
| 538 |
+
"learning_rate": 0.0001,
|
| 539 |
+
"loss": -0.0027,
|
| 540 |
+
"mismatch_kl": 0.005502623040229082,
|
| 541 |
+
"reward": 0.7100000381469727,
|
| 542 |
+
"reward/refusal_reward_func": 0.7100000381469727,
|
| 543 |
+
"reward/std": 0.26457512378692627,
|
| 544 |
+
"step": 27,
|
| 545 |
+
"timing/generation_ms": 7701.031573116779,
|
| 546 |
+
"timing/scoring_ms": 28537.479266524315,
|
| 547 |
+
"timing/total_ms": 36238.510839641094,
|
| 548 |
+
"tokens/completion": 881.84375,
|
| 549 |
+
"tokens/masked_fraction": 0.0,
|
| 550 |
+
"wall_clock/generate_s": 68.64078164100647
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"advantage/absmean": 0.06175781041383743,
|
| 554 |
+
"entropy": 0.6991814374923706,
|
| 555 |
+
"epoch": 0.056,
|
| 556 |
+
"grad_norm": 0.04213571392317473,
|
| 557 |
+
"importance_ratio": 0.9988569617271423,
|
| 558 |
+
"learning_rate": 0.0001,
|
| 559 |
+
"loss": -0.0026,
|
| 560 |
+
"mismatch_kl": 0.00713223684579134,
|
| 561 |
+
"reward": 0.07656250149011612,
|
| 562 |
+
"reward/refusal_reward_func": 0.07656250149011612,
|
| 563 |
+
"reward/std": 0.1120995506644249,
|
| 564 |
+
"step": 28,
|
| 565 |
+
"timing/generation_ms": 4908.71948748827,
|
| 566 |
+
"timing/scoring_ms": 35273.585848510265,
|
| 567 |
+
"timing/total_ms": 40182.305335998535,
|
| 568 |
+
"tokens/completion": 557.53125,
|
| 569 |
+
"tokens/masked_fraction": 0.0,
|
| 570 |
+
"wall_clock/generate_s": 92.86418867111206
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"advantage/absmean": 0.2201562523841858,
|
| 574 |
+
"entropy": 0.8785261511802673,
|
| 575 |
+
"epoch": 0.058,
|
| 576 |
+
"grad_norm": 0.3194279303513244,
|
| 577 |
+
"importance_ratio": 0.998691737651825,
|
| 578 |
+
"learning_rate": 0.0001,
|
| 579 |
+
"loss": 0.0079,
|
| 580 |
+
"mismatch_kl": 0.006111000664532185,
|
| 581 |
+
"reward": 0.5774999856948853,
|
| 582 |
+
"reward/refusal_reward_func": 0.5774999856948853,
|
| 583 |
+
"reward/std": 0.2741691768169403,
|
| 584 |
+
"step": 29,
|
| 585 |
+
"timing/generation_ms": 4181.9010972976685,
|
| 586 |
+
"timing/scoring_ms": 29512.903429567814,
|
| 587 |
+
"timing/total_ms": 33694.80452686548,
|
| 588 |
+
"tokens/completion": 478.875,
|
| 589 |
+
"tokens/masked_fraction": 0.0,
|
| 590 |
+
"wall_clock/generate_s": 72.91563320159912
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"advantage/absmean": 0.20125000178813934,
|
| 594 |
+
"entropy": 0.6921765208244324,
|
| 595 |
+
"epoch": 0.06,
|
| 596 |
+
"grad_norm": 0.4406053771041607,
|
| 597 |
+
"importance_ratio": 0.9964888095855713,
|
| 598 |
+
"learning_rate": 0.0001,
|
| 599 |
+
"loss": -0.0386,
|
| 600 |
+
"mismatch_kl": 0.01204030029475689,
|
| 601 |
+
"reward": 0.17999999225139618,
|
| 602 |
+
"reward/refusal_reward_func": 0.17999999225139618,
|
| 603 |
+
"reward/std": 0.23286262154579163,
|
| 604 |
+
"step": 30,
|
| 605 |
+
"timing/generation_ms": 3321.3287368416786,
|
| 606 |
+
"timing/scoring_ms": 29757.627181708813,
|
| 607 |
+
"timing/total_ms": 33078.95591855049,
|
| 608 |
+
"tokens/completion": 382.8125,
|
| 609 |
+
"tokens/masked_fraction": 0.0,
|
| 610 |
+
"wall_clock/generate_s": 53.77872610092163
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"advantage/absmean": 0.0018164062639698386,
|
| 614 |
+
"entropy": 0.5210146903991699,
|
| 615 |
+
"epoch": 0.062,
|
| 616 |
+
"grad_norm": 0.01658749077570669,
|
| 617 |
+
"importance_ratio": 0.9975528120994568,
|
| 618 |
+
"learning_rate": 0.0001,
|
| 619 |
+
"loss": 0.0,
|
| 620 |
+
"mismatch_kl": 0.024219391867518425,
|
| 621 |
+
"reward": 0.010937499813735485,
|
| 622 |
+
"reward/refusal_reward_func": 0.010937499813735485,
|
| 623 |
+
"reward/std": 0.005219778511673212,
|
| 624 |
+
"step": 31,
|
| 625 |
+
"timing/generation_ms": 1940.1119500398636,
|
| 626 |
+
"timing/scoring_ms": 23180.305778980255,
|
| 627 |
+
"timing/total_ms": 25120.41772902012,
|
| 628 |
+
"tokens/completion": 190.5625,
|
| 629 |
+
"tokens/masked_fraction": 0.0,
|
| 630 |
+
"wall_clock/generate_s": 33.192246198654175
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"advantage/absmean": 0.04218750074505806,
|
| 634 |
+
"entropy": 0.7747635245323181,
|
| 635 |
+
"epoch": 0.064,
|
| 636 |
+
"grad_norm": 0.05856219259172721,
|
| 637 |
+
"importance_ratio": 1.0011621713638306,
|
| 638 |
+
"learning_rate": 0.0001,
|
| 639 |
+
"loss": 0.0001,
|
| 640 |
+
"mismatch_kl": 0.008931240066885948,
|
| 641 |
+
"reward": 0.0456249974668026,
|
| 642 |
+
"reward/refusal_reward_func": 0.0456249974668026,
|
| 643 |
+
"reward/std": 0.06082441285252571,
|
| 644 |
+
"step": 32,
|
| 645 |
+
"timing/generation_ms": 6628.670156002045,
|
| 646 |
+
"timing/scoring_ms": 29644.853502511978,
|
| 647 |
+
"timing/total_ms": 36273.52365851402,
|
| 648 |
+
"tokens/completion": 758.0625,
|
| 649 |
+
"tokens/masked_fraction": 0.0,
|
| 650 |
+
"wall_clock/generate_s": 44.213383197784424
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"advantage/absmean": 0.005097656510770321,
|
| 654 |
+
"entropy": 0.7032718062400818,
|
| 655 |
+
"epoch": 0.066,
|
| 656 |
+
"grad_norm": 0.004871385581572528,
|
| 657 |
+
"importance_ratio": 0.9980704188346863,
|
| 658 |
+
"learning_rate": 0.0001,
|
| 659 |
+
"loss": -0.0002,
|
| 660 |
+
"mismatch_kl": 0.017929747700691223,
|
| 661 |
+
"reward": 0.012812498956918716,
|
| 662 |
+
"reward/refusal_reward_func": 0.012812498956918716,
|
| 663 |
+
"reward/std": 0.008744417689740658,
|
| 664 |
+
"step": 33,
|
| 665 |
+
"timing/generation_ms": 2495.2172189950943,
|
| 666 |
+
"timing/scoring_ms": 23352.533906698227,
|
| 667 |
+
"timing/total_ms": 25847.75112569332,
|
| 668 |
+
"tokens/completion": 271.1875,
|
| 669 |
+
"tokens/masked_fraction": 0.0,
|
| 670 |
+
"wall_clock/generate_s": 146.50593042373657
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"advantage/absmean": 0.20359376072883606,
|
| 674 |
+
"entropy": 0.7547333240509033,
|
| 675 |
+
"epoch": 0.068,
|
| 676 |
+
"grad_norm": 0.2936937114262587,
|
| 677 |
+
"importance_ratio": 1.0006296634674072,
|
| 678 |
+
"learning_rate": 0.0001,
|
| 679 |
+
"loss": -0.0116,
|
| 680 |
+
"mismatch_kl": 0.011033565737307072,
|
| 681 |
+
"reward": 0.47718751430511475,
|
| 682 |
+
"reward/refusal_reward_func": 0.47718751430511475,
|
| 683 |
+
"reward/std": 0.26260992884635925,
|
| 684 |
+
"step": 34,
|
| 685 |
+
"timing/generation_ms": 4908.003121614456,
|
| 686 |
+
"timing/scoring_ms": 30874.776013195515,
|
| 687 |
+
"timing/total_ms": 35782.77913480997,
|
| 688 |
+
"tokens/completion": 551.65625,
|
| 689 |
+
"tokens/masked_fraction": 0.0,
|
| 690 |
+
"wall_clock/generate_s": 60.00288248062134
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"advantage/absmean": 0.11302733421325684,
|
| 694 |
+
"entropy": 0.7300074696540833,
|
| 695 |
+
"epoch": 0.07,
|
| 696 |
+
"grad_norm": 0.32493885105188386,
|
| 697 |
+
"importance_ratio": 0.9997804164886475,
|
| 698 |
+
"learning_rate": 0.0001,
|
| 699 |
+
"loss": -0.0011,
|
| 700 |
+
"mismatch_kl": 0.016147281974554062,
|
| 701 |
+
"reward": 0.10593750327825546,
|
| 702 |
+
"reward/refusal_reward_func": 0.10593750327825546,
|
| 703 |
+
"reward/std": 0.16747872531414032,
|
| 704 |
+
"step": 35,
|
| 705 |
+
"timing/generation_ms": 3645.879790186882,
|
| 706 |
+
"timing/scoring_ms": 28717.37616509199,
|
| 707 |
+
"timing/total_ms": 32363.255955278873,
|
| 708 |
+
"tokens/completion": 418.03125,
|
| 709 |
+
"tokens/masked_fraction": 0.0,
|
| 710 |
+
"wall_clock/generate_s": 58.10710334777832
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"advantage/absmean": 0.015136717818677425,
|
| 714 |
+
"entropy": 0.6982847452163696,
|
| 715 |
+
"epoch": 0.072,
|
| 716 |
+
"grad_norm": 0.04051473715939973,
|
| 717 |
+
"importance_ratio": 0.9992015957832336,
|
| 718 |
+
"learning_rate": 0.0001,
|
| 719 |
+
"loss": -0.0006,
|
| 720 |
+
"mismatch_kl": 0.01328412164002657,
|
| 721 |
+
"reward": 0.019687499850988388,
|
| 722 |
+
"reward/refusal_reward_func": 0.019687499850988388,
|
| 723 |
+
"reward/std": 0.02113710716366768,
|
| 724 |
+
"step": 36,
|
| 725 |
+
"timing/generation_ms": 3033.673010766506,
|
| 726 |
+
"timing/scoring_ms": 29509.141087532043,
|
| 727 |
+
"timing/total_ms": 32542.81409829855,
|
| 728 |
+
"tokens/completion": 339.375,
|
| 729 |
+
"tokens/masked_fraction": 0.0,
|
| 730 |
+
"wall_clock/generate_s": 151.12143540382385
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"advantage/absmean": 0.013593749143183231,
|
| 734 |
+
"entropy": 0.6699934005737305,
|
| 735 |
+
"epoch": 0.074,
|
| 736 |
+
"grad_norm": 0.02768782924464237,
|
| 737 |
+
"importance_ratio": 1.0014734268188477,
|
| 738 |
+
"learning_rate": 0.0001,
|
| 739 |
+
"loss": -0.0003,
|
| 740 |
+
"mismatch_kl": 0.00942978449165821,
|
| 741 |
+
"reward": 0.019062498584389687,
|
| 742 |
+
"reward/refusal_reward_func": 0.019062498584389687,
|
| 743 |
+
"reward/std": 0.017741085961461067,
|
| 744 |
+
"step": 37,
|
| 745 |
+
"timing/generation_ms": 4092.4242958426476,
|
| 746 |
+
"timing/scoring_ms": 29670.215159654617,
|
| 747 |
+
"timing/total_ms": 33762.639455497265,
|
| 748 |
+
"tokens/completion": 466.8125,
|
| 749 |
+
"tokens/masked_fraction": 0.0,
|
| 750 |
+
"wall_clock/generate_s": 77.67575216293335
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"advantage/absmean": 0.09351562708616257,
|
| 754 |
+
"entropy": 0.6306953430175781,
|
| 755 |
+
"epoch": 0.076,
|
| 756 |
+
"grad_norm": 0.2799348001151443,
|
| 757 |
+
"importance_ratio": 0.9985800981521606,
|
| 758 |
+
"learning_rate": 0.0001,
|
| 759 |
+
"loss": -0.0037,
|
| 760 |
+
"mismatch_kl": 0.009972751140594482,
|
| 761 |
+
"reward": 0.06187500059604645,
|
| 762 |
+
"reward/refusal_reward_func": 0.06187500059604645,
|
| 763 |
+
"reward/std": 0.19330088794231415,
|
| 764 |
+
"step": 38,
|
| 765 |
+
"timing/generation_ms": 5255.660645663738,
|
| 766 |
+
"timing/scoring_ms": 29534.79740768671,
|
| 767 |
+
"timing/total_ms": 34790.45805335045,
|
| 768 |
+
"tokens/completion": 602.75,
|
| 769 |
+
"tokens/masked_fraction": 0.0,
|
| 770 |
+
"wall_clock/generate_s": 56.81032872200012
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"advantage/absmean": 0.021894531324505806,
|
| 774 |
+
"entropy": 0.6860240697860718,
|
| 775 |
+
"epoch": 0.078,
|
| 776 |
+
"grad_norm": 0.021804850572780036,
|
| 777 |
+
"importance_ratio": 0.999793291091919,
|
| 778 |
+
"learning_rate": 0.0001,
|
| 779 |
+
"loss": 0.0008,
|
| 780 |
+
"mismatch_kl": 0.008842560462653637,
|
| 781 |
+
"reward": 0.028437498956918716,
|
| 782 |
+
"reward/refusal_reward_func": 0.028437498956918716,
|
| 783 |
+
"reward/std": 0.02670549787580967,
|
| 784 |
+
"step": 39,
|
| 785 |
+
"timing/generation_ms": 5519.4277837872505,
|
| 786 |
+
"timing/scoring_ms": 28062.6777485013,
|
| 787 |
+
"timing/total_ms": 33582.10553228855,
|
| 788 |
+
"tokens/completion": 635.6875,
|
| 789 |
+
"tokens/masked_fraction": 0.0,
|
| 790 |
+
"wall_clock/generate_s": 49.89069700241089
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"advantage/absmean": 0.04843749850988388,
|
| 794 |
+
"entropy": 0.7330471873283386,
|
| 795 |
+
"epoch": 0.08,
|
| 796 |
+
"grad_norm": 0.23560813153480506,
|
| 797 |
+
"importance_ratio": 1.0003836154937744,
|
| 798 |
+
"learning_rate": 0.0001,
|
| 799 |
+
"loss": 0.0014,
|
| 800 |
+
"mismatch_kl": 0.008081368170678616,
|
| 801 |
+
"reward": 0.03500000014901161,
|
| 802 |
+
"reward/refusal_reward_func": 0.03500000014901161,
|
| 803 |
+
"reward/std": 0.13919411599636078,
|
| 804 |
+
"step": 40,
|
| 805 |
+
"timing/generation_ms": 3926.6494438052177,
|
| 806 |
+
"timing/scoring_ms": 27580.46282082796,
|
| 807 |
+
"timing/total_ms": 31507.11226463318,
|
| 808 |
+
"tokens/completion": 445.78125,
|
| 809 |
+
"tokens/masked_fraction": 0.0,
|
| 810 |
+
"wall_clock/generate_s": 70.10888338088989
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"advantage/absmean": 0.12544921040534973,
|
| 814 |
+
"entropy": 0.5206725597381592,
|
| 815 |
+
"epoch": 0.082,
|
| 816 |
+
"grad_norm": 0.5033391726424777,
|
| 817 |
+
"importance_ratio": 1.0012822151184082,
|
| 818 |
+
"learning_rate": 0.0001,
|
| 819 |
+
"loss": -0.0225,
|
| 820 |
+
"mismatch_kl": 0.007961519993841648,
|
| 821 |
+
"reward": 0.1446875035762787,
|
| 822 |
+
"reward/refusal_reward_func": 0.1446875035762787,
|
| 823 |
+
"reward/std": 0.1963731199502945,
|
| 824 |
+
"step": 41,
|
| 825 |
+
"timing/generation_ms": 1945.6753060221672,
|
| 826 |
+
"timing/scoring_ms": 24715.371668338776,
|
| 827 |
+
"timing/total_ms": 26661.046974360943,
|
| 828 |
+
"tokens/completion": 203.96875,
|
| 829 |
+
"tokens/masked_fraction": 0.0,
|
| 830 |
+
"wall_clock/generate_s": 41.8072030544281
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"advantage/absmean": 0.0018164062639698386,
|
| 834 |
+
"entropy": 0.82820063829422,
|
| 835 |
+
"epoch": 0.084,
|
| 836 |
+
"grad_norm": 0.0008894764900618824,
|
| 837 |
+
"importance_ratio": 0.9987862706184387,
|
| 838 |
+
"learning_rate": 0.0001,
|
| 839 |
+
"loss": 0.0002,
|
| 840 |
+
"mismatch_kl": 0.008028813637793064,
|
| 841 |
+
"reward": 0.010937499813735485,
|
| 842 |
+
"reward/refusal_reward_func": 0.010937499813735485,
|
| 843 |
+
"reward/std": 0.005219778511673212,
|
| 844 |
+
"step": 42,
|
| 845 |
+
"timing/generation_ms": 6800.906598567963,
|
| 846 |
+
"timing/scoring_ms": 31876.1548101902,
|
| 847 |
+
"timing/total_ms": 38677.06140875816,
|
| 848 |
+
"tokens/completion": 785.28125,
|
| 849 |
+
"tokens/masked_fraction": 0.0,
|
| 850 |
+
"wall_clock/generate_s": 59.81179714202881
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"advantage/absmean": 0.050859373062849045,
|
| 854 |
+
"entropy": 0.829409658908844,
|
| 855 |
+
"epoch": 0.086,
|
| 856 |
+
"grad_norm": 0.1901939415206169,
|
| 857 |
+
"importance_ratio": 1.0000019073486328,
|
| 858 |
+
"learning_rate": 0.0001,
|
| 859 |
+
"loss": 0.003,
|
| 860 |
+
"mismatch_kl": 0.009588975459337234,
|
| 861 |
+
"reward": 0.04312499612569809,
|
| 862 |
+
"reward/refusal_reward_func": 0.04312499612569809,
|
| 863 |
+
"reward/std": 0.1388217806816101,
|
| 864 |
+
"step": 43,
|
| 865 |
+
"timing/generation_ms": 6013.470813632011,
|
| 866 |
+
"timing/scoring_ms": 28884.994342923164,
|
| 867 |
+
"timing/total_ms": 34898.465156555176,
|
| 868 |
+
"tokens/completion": 691.1875,
|
| 869 |
+
"tokens/masked_fraction": 0.0,
|
| 870 |
+
"wall_clock/generate_s": 51.95971155166626
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"advantage/absmean": 0.01318359375,
|
| 874 |
+
"entropy": 0.9210640788078308,
|
| 875 |
+
"epoch": 0.088,
|
| 876 |
+
"grad_norm": 0.03744765208460247,
|
| 877 |
+
"importance_ratio": 0.9966821670532227,
|
| 878 |
+
"learning_rate": 0.0001,
|
| 879 |
+
"loss": -0.0019,
|
| 880 |
+
"mismatch_kl": 0.018160995095968246,
|
| 881 |
+
"reward": 0.017812499776482582,
|
| 882 |
+
"reward/refusal_reward_func": 0.017812499776482582,
|
| 883 |
+
"reward/std": 0.020575225353240967,
|
| 884 |
+
"step": 44,
|
| 885 |
+
"timing/generation_ms": 2924.579069018364,
|
| 886 |
+
"timing/scoring_ms": 20441.766560077667,
|
| 887 |
+
"timing/total_ms": 23366.34562909603,
|
| 888 |
+
"tokens/completion": 343.75,
|
| 889 |
+
"tokens/masked_fraction": 0.0,
|
| 890 |
+
"wall_clock/generate_s": 84.02694988250732
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"advantage/absmean": 0.04843749850988388,
|
| 894 |
+
"entropy": 0.8891170024871826,
|
| 895 |
+
"epoch": 0.09,
|
| 896 |
+
"grad_norm": 0.12978747422641476,
|
| 897 |
+
"importance_ratio": 0.9995278120040894,
|
| 898 |
+
"learning_rate": 0.0001,
|
| 899 |
+
"loss": -0.0021,
|
| 900 |
+
"mismatch_kl": 0.004716904368251562,
|
| 901 |
+
"reward": 0.7849999666213989,
|
| 902 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 903 |
+
"reward/std": 0.13919411599636078,
|
| 904 |
+
"step": 45,
|
| 905 |
+
"timing/generation_ms": 12784.472778439522,
|
| 906 |
+
"timing/scoring_ms": 41995.29768526554,
|
| 907 |
+
"timing/total_ms": 54779.77046370506,
|
| 908 |
+
"tokens/completion": 1424.53125,
|
| 909 |
+
"tokens/masked_fraction": 0.0,
|
| 910 |
+
"wall_clock/generate_s": 159.0035264492035
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"advantage/absmean": 0.19917967915534973,
|
| 914 |
+
"entropy": 1.1817245483398438,
|
| 915 |
+
"epoch": 0.092,
|
| 916 |
+
"grad_norm": 0.18725081241592767,
|
| 917 |
+
"importance_ratio": 1.0018072128295898,
|
| 918 |
+
"learning_rate": 0.0001,
|
| 919 |
+
"loss": -0.0256,
|
| 920 |
+
"mismatch_kl": 0.008767618797719479,
|
| 921 |
+
"reward": 0.5290625095367432,
|
| 922 |
+
"reward/refusal_reward_func": 0.5290625095367432,
|
| 923 |
+
"reward/std": 0.2553595304489136,
|
| 924 |
+
"step": 46,
|
| 925 |
+
"timing/generation_ms": 11124.341294169426,
|
| 926 |
+
"timing/scoring_ms": 37177.75782942772,
|
| 927 |
+
"timing/total_ms": 48302.099123597145,
|
| 928 |
+
"tokens/completion": 1282.71875,
|
| 929 |
+
"tokens/masked_fraction": 0.0,
|
| 930 |
+
"wall_clock/generate_s": 67.16122078895569
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"advantage/absmean": 0.02968750149011612,
|
| 934 |
+
"entropy": 0.9758523106575012,
|
| 935 |
+
"epoch": 0.094,
|
| 936 |
+
"grad_norm": 0.038752578543884246,
|
| 937 |
+
"importance_ratio": 1.0004377365112305,
|
| 938 |
+
"learning_rate": 0.0001,
|
| 939 |
+
"loss": 0.0042,
|
| 940 |
+
"mismatch_kl": 0.007721519563347101,
|
| 941 |
+
"reward": 0.03500000014901161,
|
| 942 |
+
"reward/refusal_reward_func": 0.03500000014901161,
|
| 943 |
+
"reward/std": 0.04690415784716606,
|
| 944 |
+
"step": 47,
|
| 945 |
+
"timing/generation_ms": 8879.09684330225,
|
| 946 |
+
"timing/scoring_ms": 37631.5980181098,
|
| 947 |
+
"timing/total_ms": 46510.69486141205,
|
| 948 |
+
"tokens/completion": 1009.46875,
|
| 949 |
+
"tokens/masked_fraction": 0.0,
|
| 950 |
+
"wall_clock/generate_s": 73.34895396232605
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"advantage/absmean": 0.13593749701976776,
|
| 954 |
+
"entropy": 0.918021559715271,
|
| 955 |
+
"epoch": 0.096,
|
| 956 |
+
"grad_norm": 0.19107099447712278,
|
| 957 |
+
"importance_ratio": 0.9999390244483948,
|
| 958 |
+
"learning_rate": 0.0001,
|
| 959 |
+
"loss": -0.0005,
|
| 960 |
+
"mismatch_kl": 0.007433234713971615,
|
| 961 |
+
"reward": 0.7350000143051147,
|
| 962 |
+
"reward/refusal_reward_func": 0.7350000143051147,
|
| 963 |
+
"reward/std": 0.23318448662757874,
|
| 964 |
+
"step": 48,
|
| 965 |
+
"timing/generation_ms": 10426.40034854412,
|
| 966 |
+
"timing/scoring_ms": 37421.48996144533,
|
| 967 |
+
"timing/total_ms": 47847.89030998945,
|
| 968 |
+
"tokens/completion": 1180.71875,
|
| 969 |
+
"tokens/masked_fraction": 0.0,
|
| 970 |
+
"wall_clock/generate_s": 93.57710862159729
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"advantage/absmean": 0.17765624821186066,
|
| 974 |
+
"entropy": 1.0576484203338623,
|
| 975 |
+
"epoch": 0.098,
|
| 976 |
+
"grad_norm": 0.16558979067438845,
|
| 977 |
+
"importance_ratio": 1.0013474225997925,
|
| 978 |
+
"learning_rate": 0.0001,
|
| 979 |
+
"loss": -0.006,
|
| 980 |
+
"mismatch_kl": 0.006344192661345005,
|
| 981 |
+
"reward": 0.6915624737739563,
|
| 982 |
+
"reward/refusal_reward_func": 0.6915624737739563,
|
| 983 |
+
"reward/std": 0.24835848808288574,
|
| 984 |
+
"step": 49,
|
| 985 |
+
"timing/generation_ms": 8644.713327288628,
|
| 986 |
+
"timing/scoring_ms": 35074.74631816149,
|
| 987 |
+
"timing/total_ms": 43719.459645450115,
|
| 988 |
+
"tokens/completion": 984.625,
|
| 989 |
+
"tokens/masked_fraction": 0.0,
|
| 990 |
+
"wall_clock/generate_s": 63.12021732330322
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"advantage/absmean": 0.12312500178813934,
|
| 994 |
+
"entropy": 0.8099650144577026,
|
| 995 |
+
"epoch": 0.1,
|
| 996 |
+
"grad_norm": 0.1253616039663461,
|
| 997 |
+
"importance_ratio": 1.0017279386520386,
|
| 998 |
+
"learning_rate": 0.0001,
|
| 999 |
+
"loss": -0.0052,
|
| 1000 |
+
"mismatch_kl": 0.008134027011692524,
|
| 1001 |
+
"reward": 0.14249999821186066,
|
| 1002 |
+
"reward/refusal_reward_func": 0.14249999821186066,
|
| 1003 |
+
"reward/std": 0.1628841608762741,
|
| 1004 |
+
"step": 50,
|
| 1005 |
+
"timing/generation_ms": 8775.396101176739,
|
| 1006 |
+
"timing/scoring_ms": 40764.19186592102,
|
| 1007 |
+
"timing/total_ms": 49539.58796709776,
|
| 1008 |
+
"tokens/completion": 1000.1875,
|
| 1009 |
+
"tokens/masked_fraction": 0.0,
|
| 1010 |
+
"wall_clock/generate_s": 70.7229483127594
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"advantage/absmean": 0.04843749850988388,
|
| 1014 |
+
"entropy": 0.8692086338996887,
|
| 1015 |
+
"epoch": 0.102,
|
| 1016 |
+
"grad_norm": 0.016399389830222273,
|
| 1017 |
+
"importance_ratio": 1.001124620437622,
|
| 1018 |
+
"learning_rate": 0.0001,
|
| 1019 |
+
"loss": -0.0013,
|
| 1020 |
+
"mismatch_kl": 0.010322043672204018,
|
| 1021 |
+
"reward": 0.7849999666213989,
|
| 1022 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 1023 |
+
"reward/std": 0.13919411599636078,
|
| 1024 |
+
"step": 51,
|
| 1025 |
+
"timing/generation_ms": 14720.608927309513,
|
| 1026 |
+
"timing/scoring_ms": 41765.84377884865,
|
| 1027 |
+
"timing/total_ms": 56486.45270615816,
|
| 1028 |
+
"tokens/completion": 1616.15625,
|
| 1029 |
+
"tokens/masked_fraction": 0.0,
|
| 1030 |
+
"wall_clock/generate_s": 89.23512148857117
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"advantage/absmean": 0.2948242425918579,
|
| 1034 |
+
"entropy": 1.1107399463653564,
|
| 1035 |
+
"epoch": 0.104,
|
| 1036 |
+
"grad_norm": 0.18323432452024488,
|
| 1037 |
+
"importance_ratio": 1.0023913383483887,
|
| 1038 |
+
"learning_rate": 0.0001,
|
| 1039 |
+
"loss": 0.0017,
|
| 1040 |
+
"mismatch_kl": 0.0076919617131352425,
|
| 1041 |
+
"reward": 0.32218748331069946,
|
| 1042 |
+
"reward/refusal_reward_func": 0.32218748331069946,
|
| 1043 |
+
"reward/std": 0.3130169212818146,
|
| 1044 |
+
"step": 52,
|
| 1045 |
+
"timing/generation_ms": 15175.773054361343,
|
| 1046 |
+
"timing/scoring_ms": 57348.27160835266,
|
| 1047 |
+
"timing/total_ms": 72524.044662714,
|
| 1048 |
+
"tokens/completion": 1662.0,
|
| 1049 |
+
"tokens/masked_fraction": 0.0,
|
| 1050 |
+
"wall_clock/generate_s": 182.7874722480774
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"advantage/absmean": 0.15343749523162842,
|
| 1054 |
+
"entropy": 0.9647155404090881,
|
| 1055 |
+
"epoch": 0.106,
|
| 1056 |
+
"grad_norm": 0.1470849270360251,
|
| 1057 |
+
"importance_ratio": 1.0000090599060059,
|
| 1058 |
+
"learning_rate": 0.0001,
|
| 1059 |
+
"loss": 0.0102,
|
| 1060 |
+
"mismatch_kl": 0.00862339697778225,
|
| 1061 |
+
"reward": 0.6565625071525574,
|
| 1062 |
+
"reward/refusal_reward_func": 0.6565625071525574,
|
| 1063 |
+
"reward/std": 0.22976359724998474,
|
| 1064 |
+
"step": 53,
|
| 1065 |
+
"timing/generation_ms": 10986.380942165852,
|
| 1066 |
+
"timing/scoring_ms": 38820.41800022125,
|
| 1067 |
+
"timing/total_ms": 49806.798942387104,
|
| 1068 |
+
"tokens/completion": 1229.09375,
|
| 1069 |
+
"tokens/masked_fraction": 0.0,
|
| 1070 |
+
"wall_clock/generate_s": 157.4189648628235
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"advantage/absmean": 0.04843749850988388,
|
| 1074 |
+
"entropy": 0.8174250721931458,
|
| 1075 |
+
"epoch": 0.108,
|
| 1076 |
+
"grad_norm": 0.017013624591187354,
|
| 1077 |
+
"importance_ratio": 1.000490427017212,
|
| 1078 |
+
"learning_rate": 0.0001,
|
| 1079 |
+
"loss": -0.0008,
|
| 1080 |
+
"mismatch_kl": 0.004199406132102013,
|
| 1081 |
+
"reward": 0.7849999666213989,
|
| 1082 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 1083 |
+
"reward/std": 0.13919411599636078,
|
| 1084 |
+
"step": 54,
|
| 1085 |
+
"timing/generation_ms": 13110.579743981361,
|
| 1086 |
+
"timing/scoring_ms": 39618.385925889015,
|
| 1087 |
+
"timing/total_ms": 52728.96566987038,
|
| 1088 |
+
"tokens/completion": 1453.71875,
|
| 1089 |
+
"tokens/masked_fraction": 0.0,
|
| 1090 |
+
"wall_clock/generate_s": 159.44240832328796
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"advantage/absmean": 0.07984375208616257,
|
| 1094 |
+
"entropy": 0.8634753227233887,
|
| 1095 |
+
"epoch": 0.11,
|
| 1096 |
+
"grad_norm": 0.06408238305252761,
|
| 1097 |
+
"importance_ratio": 1.0001016855239868,
|
| 1098 |
+
"learning_rate": 0.0001,
|
| 1099 |
+
"loss": -0.0018,
|
| 1100 |
+
"mismatch_kl": 0.006974226329475641,
|
| 1101 |
+
"reward": 0.7643749713897705,
|
| 1102 |
+
"reward/refusal_reward_func": 0.7643749713897705,
|
| 1103 |
+
"reward/std": 0.15140874683856964,
|
| 1104 |
+
"step": 55,
|
| 1105 |
+
"timing/generation_ms": 13839.490927755833,
|
| 1106 |
+
"timing/scoring_ms": 43094.41144019365,
|
| 1107 |
+
"timing/total_ms": 56933.902367949486,
|
| 1108 |
+
"tokens/completion": 1539.5625,
|
| 1109 |
+
"tokens/masked_fraction": 0.0,
|
| 1110 |
+
"wall_clock/generate_s": 162.71979093551636
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"advantage/absmean": 0.22667968273162842,
|
| 1114 |
+
"entropy": 0.7791604399681091,
|
| 1115 |
+
"epoch": 0.112,
|
| 1116 |
+
"grad_norm": 0.16725001444914697,
|
| 1117 |
+
"importance_ratio": 1.0006935596466064,
|
| 1118 |
+
"learning_rate": 0.0001,
|
| 1119 |
+
"loss": -0.0115,
|
| 1120 |
+
"mismatch_kl": 0.007046831306070089,
|
| 1121 |
+
"reward": 0.47468751668930054,
|
| 1122 |
+
"reward/refusal_reward_func": 0.47468751668930054,
|
| 1123 |
+
"reward/std": 0.2708203196525574,
|
| 1124 |
+
"step": 56,
|
| 1125 |
+
"timing/generation_ms": 16662.443839013577,
|
| 1126 |
+
"timing/scoring_ms": 45676.41341686249,
|
| 1127 |
+
"timing/total_ms": 62338.857255876064,
|
| 1128 |
+
"tokens/completion": 1816.03125,
|
| 1129 |
+
"tokens/masked_fraction": 0.0,
|
| 1130 |
+
"wall_clock/generate_s": 124.58920621871948
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"advantage/absmean": 0.11601562052965164,
|
| 1134 |
+
"entropy": 0.7439562678337097,
|
| 1135 |
+
"epoch": 0.114,
|
| 1136 |
+
"grad_norm": 0.13019703908354843,
|
| 1137 |
+
"importance_ratio": 1.0008944272994995,
|
| 1138 |
+
"learning_rate": 0.0001,
|
| 1139 |
+
"loss": -0.0026,
|
| 1140 |
+
"mismatch_kl": 0.009258158504962921,
|
| 1141 |
+
"reward": 0.7171875238418579,
|
| 1142 |
+
"reward/refusal_reward_func": 0.7171875238418579,
|
| 1143 |
+
"reward/std": 0.16097815334796906,
|
| 1144 |
+
"step": 57,
|
| 1145 |
+
"timing/generation_ms": 12729.440599679947,
|
| 1146 |
+
"timing/scoring_ms": 44226.92193090916,
|
| 1147 |
+
"timing/total_ms": 56956.362530589104,
|
| 1148 |
+
"tokens/completion": 1423.09375,
|
| 1149 |
+
"tokens/masked_fraction": 0.0,
|
| 1150 |
+
"wall_clock/generate_s": 158.57960319519043
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"advantage/absmean": 0.15726563334465027,
|
| 1154 |
+
"entropy": 0.8287293910980225,
|
| 1155 |
+
"epoch": 0.116,
|
| 1156 |
+
"grad_norm": 0.14055232526730846,
|
| 1157 |
+
"importance_ratio": 1.001518726348877,
|
| 1158 |
+
"learning_rate": 0.0001,
|
| 1159 |
+
"loss": -0.0132,
|
| 1160 |
+
"mismatch_kl": 0.008714662864804268,
|
| 1161 |
+
"reward": 0.6956250071525574,
|
| 1162 |
+
"reward/refusal_reward_func": 0.6956250071525574,
|
| 1163 |
+
"reward/std": 0.2274579405784607,
|
| 1164 |
+
"step": 58,
|
| 1165 |
+
"timing/generation_ms": 10148.803442716599,
|
| 1166 |
+
"timing/scoring_ms": 45419.282242655754,
|
| 1167 |
+
"timing/total_ms": 55568.08568537235,
|
| 1168 |
+
"tokens/completion": 1146.34375,
|
| 1169 |
+
"tokens/masked_fraction": 0.0,
|
| 1170 |
+
"wall_clock/generate_s": 88.32079148292542
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"advantage/absmean": 0.18738281726837158,
|
| 1174 |
+
"entropy": 0.6767469644546509,
|
| 1175 |
+
"epoch": 0.118,
|
| 1176 |
+
"grad_norm": 0.1800028526601889,
|
| 1177 |
+
"importance_ratio": 0.9996235370635986,
|
| 1178 |
+
"learning_rate": 0.0001,
|
| 1179 |
+
"loss": -0.0004,
|
| 1180 |
+
"mismatch_kl": 0.00572241609916091,
|
| 1181 |
+
"reward": 0.6946874856948853,
|
| 1182 |
+
"reward/refusal_reward_func": 0.6946874856948853,
|
| 1183 |
+
"reward/std": 0.26609423756599426,
|
| 1184 |
+
"step": 59,
|
| 1185 |
+
"timing/generation_ms": 5857.890740036964,
|
| 1186 |
+
"timing/scoring_ms": 28364.16070908308,
|
| 1187 |
+
"timing/total_ms": 34222.051449120045,
|
| 1188 |
+
"tokens/completion": 674.78125,
|
| 1189 |
+
"tokens/masked_fraction": 0.0,
|
| 1190 |
+
"wall_clock/generate_s": 153.28042459487915
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"advantage/absmean": 0.04843749850988388,
|
| 1194 |
+
"entropy": 0.48612505197525024,
|
| 1195 |
+
"epoch": 0.12,
|
| 1196 |
+
"grad_norm": 0.017636908815215933,
|
| 1197 |
+
"importance_ratio": 1.001957893371582,
|
| 1198 |
+
"learning_rate": 0.0001,
|
| 1199 |
+
"loss": -0.0003,
|
| 1200 |
+
"mismatch_kl": 0.0077150240540504456,
|
| 1201 |
+
"reward": 0.7849999666213989,
|
| 1202 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 1203 |
+
"reward/std": 0.13919411599636078,
|
| 1204 |
+
"step": 60,
|
| 1205 |
+
"timing/generation_ms": 13039.215676486492,
|
| 1206 |
+
"timing/scoring_ms": 37324.29302483797,
|
| 1207 |
+
"timing/total_ms": 50363.50870132446,
|
| 1208 |
+
"tokens/completion": 1435.25,
|
| 1209 |
+
"tokens/masked_fraction": 0.0,
|
| 1210 |
+
"wall_clock/generate_s": 155.66002011299133
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"advantage/absmean": 0.033906251192092896,
|
| 1214 |
+
"entropy": 0.8923248648643494,
|
| 1215 |
+
"epoch": 0.122,
|
| 1216 |
+
"grad_norm": 0.01138046317613093,
|
| 1217 |
+
"importance_ratio": 1.0010781288146973,
|
| 1218 |
+
"learning_rate": 0.0001,
|
| 1219 |
+
"loss": -0.0001,
|
| 1220 |
+
"mismatch_kl": 0.009470692835748196,
|
| 1221 |
+
"reward": 0.7925000190734863,
|
| 1222 |
+
"reward/refusal_reward_func": 0.7925000190734863,
|
| 1223 |
+
"reward/std": 0.09743587672710419,
|
| 1224 |
+
"step": 61,
|
| 1225 |
+
"timing/generation_ms": 20387.244410812855,
|
| 1226 |
+
"timing/scoring_ms": 49000.582568347454,
|
| 1227 |
+
"timing/total_ms": 69387.82697916031,
|
| 1228 |
+
"tokens/completion": 2047.65625,
|
| 1229 |
+
"tokens/masked_fraction": 0.0,
|
| 1230 |
+
"wall_clock/generate_s": 179.17891383171082
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"advantage/absmean": 0.14195312559604645,
|
| 1234 |
+
"entropy": 0.6803052425384521,
|
| 1235 |
+
"epoch": 0.124,
|
| 1236 |
+
"grad_norm": 0.1396860758416004,
|
| 1237 |
+
"importance_ratio": 1.0004676580429077,
|
| 1238 |
+
"learning_rate": 0.0001,
|
| 1239 |
+
"loss": 0.0005,
|
| 1240 |
+
"mismatch_kl": 0.009484711103141308,
|
| 1241 |
+
"reward": 0.7112500071525574,
|
| 1242 |
+
"reward/refusal_reward_func": 0.7112500071525574,
|
| 1243 |
+
"reward/std": 0.19915054738521576,
|
| 1244 |
+
"step": 62,
|
| 1245 |
+
"timing/generation_ms": 20013.910226523876,
|
| 1246 |
+
"timing/scoring_ms": 51556.99533224106,
|
| 1247 |
+
"timing/total_ms": 71570.90555876493,
|
| 1248 |
+
"tokens/completion": 2029.46875,
|
| 1249 |
+
"tokens/masked_fraction": 0.0,
|
| 1250 |
+
"wall_clock/generate_s": 281.4015655517578
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"advantage/absmean": 0.14208984375,
|
| 1254 |
+
"entropy": 0.7867326736450195,
|
| 1255 |
+
"epoch": 0.126,
|
| 1256 |
+
"grad_norm": 0.07770627410364793,
|
| 1257 |
+
"importance_ratio": 1.0014797449111938,
|
| 1258 |
+
"learning_rate": 0.0001,
|
| 1259 |
+
"loss": -0.0003,
|
| 1260 |
+
"mismatch_kl": 0.014644050039350986,
|
| 1261 |
+
"reward": 0.6584374904632568,
|
| 1262 |
+
"reward/refusal_reward_func": 0.6584374904632568,
|
| 1263 |
+
"reward/std": 0.21947535872459412,
|
| 1264 |
+
"step": 63,
|
| 1265 |
+
"timing/generation_ms": 19855.214461684227,
|
| 1266 |
+
"timing/scoring_ms": 52655.76823055744,
|
| 1267 |
+
"timing/total_ms": 72510.98269224167,
|
| 1268 |
+
"tokens/completion": 2026.1875,
|
| 1269 |
+
"tokens/masked_fraction": 0.0,
|
| 1270 |
+
"wall_clock/generate_s": 168.5356569290161
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"advantage/absmean": 0.09375,
|
| 1274 |
+
"entropy": 0.722605288028717,
|
| 1275 |
+
"epoch": 0.128,
|
| 1276 |
+
"grad_norm": 0.10541623752837016,
|
| 1277 |
+
"importance_ratio": 1.0017896890640259,
|
| 1278 |
+
"learning_rate": 0.0001,
|
| 1279 |
+
"loss": 0.0002,
|
| 1280 |
+
"mismatch_kl": 0.009318462572991848,
|
| 1281 |
+
"reward": 0.7599999904632568,
|
| 1282 |
+
"reward/refusal_reward_func": 0.7599999904632568,
|
| 1283 |
+
"reward/std": 0.19364915788173676,
|
| 1284 |
+
"step": 64,
|
| 1285 |
+
"timing/generation_ms": 20329.02915775776,
|
| 1286 |
+
"timing/scoring_ms": 43974.1270840168,
|
| 1287 |
+
"timing/total_ms": 64303.15624177456,
|
| 1288 |
+
"tokens/completion": 2046.09375,
|
| 1289 |
+
"tokens/masked_fraction": 0.0,
|
| 1290 |
+
"wall_clock/generate_s": 69.69067120552063
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"advantage/absmean": 0.13593751192092896,
|
| 1294 |
+
"entropy": 0.6431602239608765,
|
| 1295 |
+
"epoch": 0.13,
|
| 1296 |
+
"grad_norm": 0.04996864946933639,
|
| 1297 |
+
"importance_ratio": 1.0007299184799194,
|
| 1298 |
+
"learning_rate": 0.0001,
|
| 1299 |
+
"loss": 0.0042,
|
| 1300 |
+
"mismatch_kl": 0.01128534134477377,
|
| 1301 |
+
"reward": 0.7350000143051147,
|
| 1302 |
+
"reward/refusal_reward_func": 0.7350000143051147,
|
| 1303 |
+
"reward/std": 0.23318447172641754,
|
| 1304 |
+
"step": 65,
|
| 1305 |
+
"timing/generation_ms": 17292.446829378605,
|
| 1306 |
+
"timing/scoring_ms": 48005.3500905633,
|
| 1307 |
+
"timing/total_ms": 65297.7969199419,
|
| 1308 |
+
"tokens/completion": 1868.75,
|
| 1309 |
+
"tokens/masked_fraction": 0.0,
|
| 1310 |
+
"wall_clock/generate_s": 137.745671749115
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"advantage/absmean": 0.043593745678663254,
|
| 1314 |
+
"entropy": 0.6788095831871033,
|
| 1315 |
+
"epoch": 0.132,
|
| 1316 |
+
"grad_norm": 0.1104672773690437,
|
| 1317 |
+
"importance_ratio": 1.0008015632629395,
|
| 1318 |
+
"learning_rate": 0.0001,
|
| 1319 |
+
"loss": -0.0,
|
| 1320 |
+
"mismatch_kl": 0.011132912710309029,
|
| 1321 |
+
"reward": 0.7875000238418579,
|
| 1322 |
+
"reward/refusal_reward_func": 0.7875000238418579,
|
| 1323 |
+
"reward/std": 0.1252746880054474,
|
| 1324 |
+
"step": 66,
|
| 1325 |
+
"timing/generation_ms": 20402.18196809292,
|
| 1326 |
+
"timing/scoring_ms": 42341.174609959126,
|
| 1327 |
+
"timing/total_ms": 62743.356578052044,
|
| 1328 |
+
"tokens/completion": 2048.0,
|
| 1329 |
+
"tokens/masked_fraction": 0.0,
|
| 1330 |
+
"wall_clock/generate_s": 162.58158588409424
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"advantage/absmean": 0.05214843899011612,
|
| 1334 |
+
"entropy": 0.5562156438827515,
|
| 1335 |
+
"epoch": 0.134,
|
| 1336 |
+
"grad_norm": 0.016212117765327168,
|
| 1337 |
+
"importance_ratio": 0.9997291564941406,
|
| 1338 |
+
"learning_rate": 0.0001,
|
| 1339 |
+
"loss": 0.0,
|
| 1340 |
+
"mismatch_kl": 0.008888973854482174,
|
| 1341 |
+
"reward": 0.7821874618530273,
|
| 1342 |
+
"reward/refusal_reward_func": 0.7821874618530273,
|
| 1343 |
+
"reward/std": 0.1277872771024704,
|
| 1344 |
+
"step": 67,
|
| 1345 |
+
"timing/generation_ms": 20475.26439279318,
|
| 1346 |
+
"timing/scoring_ms": 46698.052957654,
|
| 1347 |
+
"timing/total_ms": 67173.31735044718,
|
| 1348 |
+
"tokens/completion": 2048.0,
|
| 1349 |
+
"tokens/masked_fraction": 0.0,
|
| 1350 |
+
"wall_clock/generate_s": 151.99999165534973
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"advantage/absmean": 0.08964844048023224,
|
| 1354 |
+
"entropy": 0.6966589093208313,
|
| 1355 |
+
"epoch": 0.136,
|
| 1356 |
+
"grad_norm": 0.1577321206922815,
|
| 1357 |
+
"importance_ratio": 1.0010697841644287,
|
| 1358 |
+
"learning_rate": 0.0001,
|
| 1359 |
+
"loss": -0.0002,
|
| 1360 |
+
"mismatch_kl": 0.00940707977861166,
|
| 1361 |
+
"reward": 0.7621874809265137,
|
| 1362 |
+
"reward/refusal_reward_func": 0.7621874809265137,
|
| 1363 |
+
"reward/std": 0.18551842868328094,
|
| 1364 |
+
"step": 68,
|
| 1365 |
+
"timing/generation_ms": 20498.32931160927,
|
| 1366 |
+
"timing/scoring_ms": 54109.15730148554,
|
| 1367 |
+
"timing/total_ms": 74607.4866130948,
|
| 1368 |
+
"tokens/completion": 2048.0,
|
| 1369 |
+
"tokens/masked_fraction": 0.0,
|
| 1370 |
+
"wall_clock/generate_s": 395.203547000885
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"advantage/absmean": 0.27099609375,
|
| 1374 |
+
"entropy": 0.5135282278060913,
|
| 1375 |
+
"epoch": 0.138,
|
| 1376 |
+
"grad_norm": 0.2079259512896872,
|
| 1377 |
+
"importance_ratio": 1.0020484924316406,
|
| 1378 |
+
"learning_rate": 0.0001,
|
| 1379 |
+
"loss": -0.0,
|
| 1380 |
+
"mismatch_kl": 0.012067537754774094,
|
| 1381 |
+
"reward": 0.6365625262260437,
|
| 1382 |
+
"reward/refusal_reward_func": 0.6365625262260437,
|
| 1383 |
+
"reward/std": 0.32783961296081543,
|
| 1384 |
+
"step": 69,
|
| 1385 |
+
"timing/generation_ms": 20592.435374855995,
|
| 1386 |
+
"timing/scoring_ms": 59319.189973175526,
|
| 1387 |
+
"timing/total_ms": 79911.62534803152,
|
| 1388 |
+
"tokens/completion": 2048.0,
|
| 1389 |
+
"tokens/masked_fraction": 0.0,
|
| 1390 |
+
"wall_clock/generate_s": 397.43536710739136
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"advantage/absmean": 0.0,
|
| 1394 |
+
"entropy": 0.4793013036251068,
|
| 1395 |
+
"epoch": 0.14,
|
| 1396 |
+
"grad_norm": 0.0,
|
| 1397 |
+
"importance_ratio": 1.0002073049545288,
|
| 1398 |
+
"learning_rate": 0.0001,
|
| 1399 |
+
"loss": 0.0,
|
| 1400 |
+
"mismatch_kl": 0.01224527694284916,
|
| 1401 |
+
"reward": 0.8100000023841858,
|
| 1402 |
+
"reward/refusal_reward_func": 0.8100000023841858,
|
| 1403 |
+
"reward/std": 0.0,
|
| 1404 |
+
"step": 70,
|
| 1405 |
+
"timing/generation_ms": 20546.609550714493,
|
| 1406 |
+
"timing/scoring_ms": 42320.83362340927,
|
| 1407 |
+
"timing/total_ms": 62867.443174123764,
|
| 1408 |
+
"tokens/completion": 2048.0,
|
| 1409 |
+
"tokens/masked_fraction": 0.0,
|
| 1410 |
+
"wall_clock/generate_s": 161.7110676765442
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"advantage/absmean": 0.053593751043081284,
|
| 1414 |
+
"entropy": 0.35620445013046265,
|
| 1415 |
+
"epoch": 0.142,
|
| 1416 |
+
"grad_norm": 0.06178490851261073,
|
| 1417 |
+
"importance_ratio": 1.0008577108383179,
|
| 1418 |
+
"learning_rate": 0.0001,
|
| 1419 |
+
"loss": 0.0001,
|
| 1420 |
+
"mismatch_kl": 0.007034921087324619,
|
| 1421 |
+
"reward": 0.7793750166893005,
|
| 1422 |
+
"reward/refusal_reward_func": 0.7793750166893005,
|
| 1423 |
+
"reward/std": 0.08525467664003372,
|
| 1424 |
+
"step": 71,
|
| 1425 |
+
"timing/generation_ms": 20559.83528494835,
|
| 1426 |
+
"timing/scoring_ms": 50445.407539606094,
|
| 1427 |
+
"timing/total_ms": 71005.24282455444,
|
| 1428 |
+
"tokens/completion": 2047.65625,
|
| 1429 |
+
"tokens/masked_fraction": 0.0,
|
| 1430 |
+
"wall_clock/generate_s": 165.77110528945923
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"advantage/absmean": 0.06621094048023224,
|
| 1434 |
+
"entropy": 0.34386953711509705,
|
| 1435 |
+
"epoch": 0.144,
|
| 1436 |
+
"grad_norm": 0.06068478204359807,
|
| 1437 |
+
"importance_ratio": 1.0005115270614624,
|
| 1438 |
+
"learning_rate": 0.0001,
|
| 1439 |
+
"loss": 0.0001,
|
| 1440 |
+
"mismatch_kl": 0.006632550619542599,
|
| 1441 |
+
"reward": 0.7746874690055847,
|
| 1442 |
+
"reward/refusal_reward_func": 0.7746874690055847,
|
| 1443 |
+
"reward/std": 0.14985378086566925,
|
| 1444 |
+
"step": 72,
|
| 1445 |
+
"timing/generation_ms": 20651.6492664814,
|
| 1446 |
+
"timing/scoring_ms": 53449.473068118095,
|
| 1447 |
+
"timing/total_ms": 74101.1223345995,
|
| 1448 |
+
"tokens/completion": 2048.0,
|
| 1449 |
+
"tokens/masked_fraction": 0.0,
|
| 1450 |
+
"wall_clock/generate_s": 396.06008672714233
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"advantage/absmean": 0.08964844048023224,
|
| 1454 |
+
"entropy": 0.3689025640487671,
|
| 1455 |
+
"epoch": 0.146,
|
| 1456 |
+
"grad_norm": 0.028592002076965557,
|
| 1457 |
+
"importance_ratio": 1.0001220703125,
|
| 1458 |
+
"learning_rate": 0.0001,
|
| 1459 |
+
"loss": 0.0,
|
| 1460 |
+
"mismatch_kl": 0.009937528520822525,
|
| 1461 |
+
"reward": 0.7621874809265137,
|
| 1462 |
+
"reward/refusal_reward_func": 0.7621874809265137,
|
| 1463 |
+
"reward/std": 0.18551842868328094,
|
| 1464 |
+
"step": 73,
|
| 1465 |
+
"timing/generation_ms": 20704.90287989378,
|
| 1466 |
+
"timing/scoring_ms": 56059.35876071453,
|
| 1467 |
+
"timing/total_ms": 76764.26164060831,
|
| 1468 |
+
"tokens/completion": 2048.0,
|
| 1469 |
+
"tokens/masked_fraction": 0.0,
|
| 1470 |
+
"wall_clock/generate_s": 395.15700674057007
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"advantage/absmean": 0.10025390982627869,
|
| 1474 |
+
"entropy": 0.3531518578529358,
|
| 1475 |
+
"epoch": 0.148,
|
| 1476 |
+
"grad_norm": 0.0966128922725595,
|
| 1477 |
+
"importance_ratio": 0.9996236562728882,
|
| 1478 |
+
"learning_rate": 0.0001,
|
| 1479 |
+
"loss": -0.0001,
|
| 1480 |
+
"mismatch_kl": 0.00807525310665369,
|
| 1481 |
+
"reward": 0.754687488079071,
|
| 1482 |
+
"reward/refusal_reward_func": 0.754687488079071,
|
| 1483 |
+
"reward/std": 0.19453445076942444,
|
| 1484 |
+
"step": 74,
|
| 1485 |
+
"timing/generation_ms": 20194.012761116028,
|
| 1486 |
+
"timing/scoring_ms": 45799.67290908098,
|
| 1487 |
+
"timing/total_ms": 65993.68567019701,
|
| 1488 |
+
"tokens/completion": 2048.0,
|
| 1489 |
+
"tokens/masked_fraction": 0.0,
|
| 1490 |
+
"wall_clock/generate_s": 150.22299551963806
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"advantage/absmean": 0.08964844048023224,
|
| 1494 |
+
"entropy": 0.3910990059375763,
|
| 1495 |
+
"epoch": 0.15,
|
| 1496 |
+
"grad_norm": 0.14646197667696922,
|
| 1497 |
+
"importance_ratio": 0.999146580696106,
|
| 1498 |
+
"learning_rate": 0.0001,
|
| 1499 |
+
"loss": -0.0003,
|
| 1500 |
+
"mismatch_kl": 0.008608575910329819,
|
| 1501 |
+
"reward": 0.7621874809265137,
|
| 1502 |
+
"reward/refusal_reward_func": 0.7621874809265137,
|
| 1503 |
+
"reward/std": 0.18551842868328094,
|
| 1504 |
+
"step": 75,
|
| 1505 |
+
"timing/generation_ms": 20237.79760301113,
|
| 1506 |
+
"timing/scoring_ms": 51105.7443395257,
|
| 1507 |
+
"timing/total_ms": 71343.54194253683,
|
| 1508 |
+
"tokens/completion": 2048.0,
|
| 1509 |
+
"tokens/masked_fraction": 0.0,
|
| 1510 |
+
"wall_clock/generate_s": 394.9270164966583
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"advantage/absmean": 0.13537108898162842,
|
| 1514 |
+
"entropy": 0.2564987540245056,
|
| 1515 |
+
"epoch": 0.152,
|
| 1516 |
+
"grad_norm": 0.10767344052989877,
|
| 1517 |
+
"importance_ratio": 1.000230073928833,
|
| 1518 |
+
"learning_rate": 0.0001,
|
| 1519 |
+
"loss": 0.0002,
|
| 1520 |
+
"mismatch_kl": 0.00910898856818676,
|
| 1521 |
+
"reward": 0.7353124618530273,
|
| 1522 |
+
"reward/refusal_reward_func": 0.7353124618530273,
|
| 1523 |
+
"reward/std": 0.23228463530540466,
|
| 1524 |
+
"step": 76,
|
| 1525 |
+
"timing/generation_ms": 20231.063432991505,
|
| 1526 |
+
"timing/scoring_ms": 64505.957297980785,
|
| 1527 |
+
"timing/total_ms": 84737.02073097229,
|
| 1528 |
+
"tokens/completion": 2048.0,
|
| 1529 |
+
"tokens/masked_fraction": 0.0,
|
| 1530 |
+
"wall_clock/generate_s": 394.6906681060791
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"advantage/absmean": 0.04843749850988388,
|
| 1534 |
+
"entropy": 0.35127270221710205,
|
| 1535 |
+
"epoch": 0.154,
|
| 1536 |
+
"grad_norm": 0.05430162481667564,
|
| 1537 |
+
"importance_ratio": 1.001720666885376,
|
| 1538 |
+
"learning_rate": 0.0001,
|
| 1539 |
+
"loss": -0.0112,
|
| 1540 |
+
"mismatch_kl": 0.02907688170671463,
|
| 1541 |
+
"reward": 0.7849999666213989,
|
| 1542 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 1543 |
+
"reward/std": 0.13919411599636078,
|
| 1544 |
+
"step": 77,
|
| 1545 |
+
"timing/generation_ms": 2299.7111305594444,
|
| 1546 |
+
"timing/scoring_ms": 24145.363181829453,
|
| 1547 |
+
"timing/total_ms": 26445.074312388897,
|
| 1548 |
+
"tokens/completion": 256.90625,
|
| 1549 |
+
"tokens/masked_fraction": 0.0,
|
| 1550 |
+
"wall_clock/generate_s": 43.695470094680786
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"advantage/absmean": 0.17824219167232513,
|
| 1554 |
+
"entropy": 0.2108859121799469,
|
| 1555 |
+
"epoch": 0.156,
|
| 1556 |
+
"grad_norm": 0.1343157239601839,
|
| 1557 |
+
"importance_ratio": 1.000138521194458,
|
| 1558 |
+
"learning_rate": 0.0001,
|
| 1559 |
+
"loss": -0.0061,
|
| 1560 |
+
"mismatch_kl": 0.006567842327058315,
|
| 1561 |
+
"reward": 0.7043750286102295,
|
| 1562 |
+
"reward/refusal_reward_func": 0.7043750286102295,
|
| 1563 |
+
"reward/std": 0.26504644751548767,
|
| 1564 |
+
"step": 78,
|
| 1565 |
+
"timing/generation_ms": 19559.65828895569,
|
| 1566 |
+
"timing/scoring_ms": 56106.24121129513,
|
| 1567 |
+
"timing/total_ms": 75665.89950025082,
|
| 1568 |
+
"tokens/completion": 2012.8125,
|
| 1569 |
+
"tokens/masked_fraction": 0.0,
|
| 1570 |
+
"wall_clock/generate_s": 397.83462166786194
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"advantage/absmean": 0.14109376072883606,
|
| 1574 |
+
"entropy": 0.2118162214756012,
|
| 1575 |
+
"epoch": 0.158,
|
| 1576 |
+
"grad_norm": 0.04314766069392161,
|
| 1577 |
+
"importance_ratio": 0.999754011631012,
|
| 1578 |
+
"learning_rate": 0.0001,
|
| 1579 |
+
"loss": 0.0,
|
| 1580 |
+
"mismatch_kl": 0.005713644903153181,
|
| 1581 |
+
"reward": 0.7293750047683716,
|
| 1582 |
+
"reward/refusal_reward_func": 0.7293750047683716,
|
| 1583 |
+
"reward/std": 0.23431998491287231,
|
| 1584 |
+
"step": 79,
|
| 1585 |
+
"timing/generation_ms": 20274.492114782333,
|
| 1586 |
+
"timing/scoring_ms": 53302.132822573185,
|
| 1587 |
+
"timing/total_ms": 73576.62493735552,
|
| 1588 |
+
"tokens/completion": 2048.0,
|
| 1589 |
+
"tokens/masked_fraction": 0.0,
|
| 1590 |
+
"wall_clock/generate_s": 396.22464632987976
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"advantage/absmean": 0.13593749701976776,
|
| 1594 |
+
"entropy": 0.32609474658966064,
|
| 1595 |
+
"epoch": 0.16,
|
| 1596 |
+
"grad_norm": 0.10083540436117842,
|
| 1597 |
+
"importance_ratio": 1.0001963376998901,
|
| 1598 |
+
"learning_rate": 0.0001,
|
| 1599 |
+
"loss": -0.0002,
|
| 1600 |
+
"mismatch_kl": 0.008770663291215897,
|
| 1601 |
+
"reward": 0.7350000143051147,
|
| 1602 |
+
"reward/refusal_reward_func": 0.7350000143051147,
|
| 1603 |
+
"reward/std": 0.23318448662757874,
|
| 1604 |
+
"step": 80,
|
| 1605 |
+
"timing/generation_ms": 20281.54794126749,
|
| 1606 |
+
"timing/scoring_ms": 45572.31470942497,
|
| 1607 |
+
"timing/total_ms": 65853.86265069246,
|
| 1608 |
+
"tokens/completion": 2048.0,
|
| 1609 |
+
"tokens/masked_fraction": 0.0,
|
| 1610 |
+
"wall_clock/generate_s": 157.70197463035583
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"advantage/absmean": 0.13197265565395355,
|
| 1614 |
+
"entropy": 0.349658727645874,
|
| 1615 |
+
"epoch": 0.162,
|
| 1616 |
+
"grad_norm": 0.15801403288716265,
|
| 1617 |
+
"importance_ratio": 0.9997016191482544,
|
| 1618 |
+
"learning_rate": 0.0001,
|
| 1619 |
+
"loss": 0.0003,
|
| 1620 |
+
"mismatch_kl": 0.007918323390185833,
|
| 1621 |
+
"reward": 0.7371875047683716,
|
| 1622 |
+
"reward/refusal_reward_func": 0.7371875047683716,
|
| 1623 |
+
"reward/std": 0.22671890258789062,
|
| 1624 |
+
"step": 81,
|
| 1625 |
+
"timing/generation_ms": 19729.195773601532,
|
| 1626 |
+
"timing/scoring_ms": 52965.313747525215,
|
| 1627 |
+
"timing/total_ms": 72694.50952112675,
|
| 1628 |
+
"tokens/completion": 2031.71875,
|
| 1629 |
+
"tokens/masked_fraction": 0.0,
|
| 1630 |
+
"wall_clock/generate_s": 395.87234902381897
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"advantage/absmean": 0.19093748927116394,
|
| 1634 |
+
"entropy": 0.30307242274284363,
|
| 1635 |
+
"epoch": 0.164,
|
| 1636 |
+
"grad_norm": 0.05658978916858572,
|
| 1637 |
+
"importance_ratio": 0.9997415542602539,
|
| 1638 |
+
"learning_rate": 0.0001,
|
| 1639 |
+
"loss": 0.0,
|
| 1640 |
+
"mismatch_kl": 0.009000571444630623,
|
| 1641 |
+
"reward": 0.6924999952316284,
|
| 1642 |
+
"reward/refusal_reward_func": 0.6924999952316284,
|
| 1643 |
+
"reward/std": 0.2553306818008423,
|
| 1644 |
+
"step": 82,
|
| 1645 |
+
"timing/generation_ms": 20152.134649455547,
|
| 1646 |
+
"timing/scoring_ms": 55147.81706035137,
|
| 1647 |
+
"timing/total_ms": 75299.95170980692,
|
| 1648 |
+
"tokens/completion": 2048.0,
|
| 1649 |
+
"tokens/masked_fraction": 0.0,
|
| 1650 |
+
"wall_clock/generate_s": 276.00095558166504
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"advantage/absmean": 0.13197265565395355,
|
| 1654 |
+
"entropy": 0.2418041229248047,
|
| 1655 |
+
"epoch": 0.166,
|
| 1656 |
+
"grad_norm": 0.09124961991568047,
|
| 1657 |
+
"importance_ratio": 0.9993461966514587,
|
| 1658 |
+
"learning_rate": 0.0001,
|
| 1659 |
+
"loss": 0.0,
|
| 1660 |
+
"mismatch_kl": 0.008319162763655186,
|
| 1661 |
+
"reward": 0.7371875047683716,
|
| 1662 |
+
"reward/refusal_reward_func": 0.7371875047683716,
|
| 1663 |
+
"reward/std": 0.22671890258789062,
|
| 1664 |
+
"step": 83,
|
| 1665 |
+
"timing/generation_ms": 20235.445871949196,
|
| 1666 |
+
"timing/scoring_ms": 60683.41539800167,
|
| 1667 |
+
"timing/total_ms": 80918.86126995087,
|
| 1668 |
+
"tokens/completion": 2048.0,
|
| 1669 |
+
"tokens/masked_fraction": 0.0,
|
| 1670 |
+
"wall_clock/generate_s": 394.98225951194763
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"advantage/absmean": 0.09492187201976776,
|
| 1674 |
+
"entropy": 0.36278918385505676,
|
| 1675 |
+
"epoch": 0.168,
|
| 1676 |
+
"grad_norm": 0.10428837192295014,
|
| 1677 |
+
"importance_ratio": 1.0009691715240479,
|
| 1678 |
+
"learning_rate": 0.0001,
|
| 1679 |
+
"loss": 0.0001,
|
| 1680 |
+
"mismatch_kl": 0.006821990944445133,
|
| 1681 |
+
"reward": 0.7593749761581421,
|
| 1682 |
+
"reward/refusal_reward_func": 0.7593749761581421,
|
| 1683 |
+
"reward/std": 0.19606979191303253,
|
| 1684 |
+
"step": 84,
|
| 1685 |
+
"timing/generation_ms": 20079.659663140774,
|
| 1686 |
+
"timing/scoring_ms": 62456.88313245773,
|
| 1687 |
+
"timing/total_ms": 82536.5427955985,
|
| 1688 |
+
"tokens/completion": 2047.4375,
|
| 1689 |
+
"tokens/masked_fraction": 0.0,
|
| 1690 |
+
"wall_clock/generate_s": 395.080442905426
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"advantage/absmean": 0.16296875476837158,
|
| 1694 |
+
"entropy": 0.2741187810897827,
|
| 1695 |
+
"epoch": 0.17,
|
| 1696 |
+
"grad_norm": 0.05559178148711944,
|
| 1697 |
+
"importance_ratio": 1.0001330375671387,
|
| 1698 |
+
"learning_rate": 0.0001,
|
| 1699 |
+
"loss": 0.0002,
|
| 1700 |
+
"mismatch_kl": 0.008196860551834106,
|
| 1701 |
+
"reward": 0.7168750166893005,
|
| 1702 |
+
"reward/refusal_reward_func": 0.7168750166893005,
|
| 1703 |
+
"reward/std": 0.2492668777704239,
|
| 1704 |
+
"step": 85,
|
| 1705 |
+
"timing/generation_ms": 20191.432282328606,
|
| 1706 |
+
"timing/scoring_ms": 66640.07867872715,
|
| 1707 |
+
"timing/total_ms": 86831.51096105576,
|
| 1708 |
+
"tokens/completion": 2048.0,
|
| 1709 |
+
"tokens/masked_fraction": 0.0,
|
| 1710 |
+
"wall_clock/generate_s": 395.3853757381439
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"advantage/absmean": 0.06513672322034836,
|
| 1714 |
+
"entropy": 0.3744433522224426,
|
| 1715 |
+
"epoch": 0.172,
|
| 1716 |
+
"grad_norm": 0.02027358693373687,
|
| 1717 |
+
"importance_ratio": 0.9997438192367554,
|
| 1718 |
+
"learning_rate": 0.0001,
|
| 1719 |
+
"loss": 0.0001,
|
| 1720 |
+
"mismatch_kl": 0.0079119261354208,
|
| 1721 |
+
"reward": 0.7740625143051147,
|
| 1722 |
+
"reward/refusal_reward_func": 0.7740625143051147,
|
| 1723 |
+
"reward/std": 0.1449754238128662,
|
| 1724 |
+
"step": 86,
|
| 1725 |
+
"timing/generation_ms": 20130.67189604044,
|
| 1726 |
+
"timing/scoring_ms": 63243.81287395954,
|
| 1727 |
+
"timing/total_ms": 83374.48476999998,
|
| 1728 |
+
"tokens/completion": 2048.0,
|
| 1729 |
+
"tokens/masked_fraction": 0.0,
|
| 1730 |
+
"wall_clock/generate_s": 394.553094625473
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"advantage/absmean": 0.06621094048023224,
|
| 1734 |
+
"entropy": 0.2802242636680603,
|
| 1735 |
+
"epoch": 0.174,
|
| 1736 |
+
"grad_norm": 0.11007226741752094,
|
| 1737 |
+
"importance_ratio": 1.0002408027648926,
|
| 1738 |
+
"learning_rate": 0.0001,
|
| 1739 |
+
"loss": -0.0,
|
| 1740 |
+
"mismatch_kl": 0.010727161541581154,
|
| 1741 |
+
"reward": 0.7746875286102295,
|
| 1742 |
+
"reward/refusal_reward_func": 0.7746875286102295,
|
| 1743 |
+
"reward/std": 0.14985376596450806,
|
| 1744 |
+
"step": 87,
|
| 1745 |
+
"timing/generation_ms": 20197.582133114338,
|
| 1746 |
+
"timing/scoring_ms": 60626.82098895311,
|
| 1747 |
+
"timing/total_ms": 80824.40312206745,
|
| 1748 |
+
"tokens/completion": 2048.0,
|
| 1749 |
+
"tokens/masked_fraction": 0.0,
|
| 1750 |
+
"wall_clock/generate_s": 394.6186418533325
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"advantage/absmean": 0.15476563572883606,
|
| 1754 |
+
"entropy": 0.15260648727416992,
|
| 1755 |
+
"epoch": 0.176,
|
| 1756 |
+
"grad_norm": 0.09113868718318835,
|
| 1757 |
+
"importance_ratio": 0.9991167187690735,
|
| 1758 |
+
"learning_rate": 0.0001,
|
| 1759 |
+
"loss": -0.0,
|
| 1760 |
+
"mismatch_kl": 0.007635745219886303,
|
| 1761 |
+
"reward": 0.7215625047683716,
|
| 1762 |
+
"reward/refusal_reward_func": 0.7215625047683716,
|
| 1763 |
+
"reward/std": 0.2370404750108719,
|
| 1764 |
+
"step": 88,
|
| 1765 |
+
"timing/generation_ms": 20186.022453010082,
|
| 1766 |
+
"timing/scoring_ms": 65319.87015157938,
|
| 1767 |
+
"timing/total_ms": 85505.89260458946,
|
| 1768 |
+
"tokens/completion": 2048.0,
|
| 1769 |
+
"tokens/masked_fraction": 0.0,
|
| 1770 |
+
"wall_clock/generate_s": 394.704843044281
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"advantage/absmean": 0.09375,
|
| 1774 |
+
"entropy": 0.2925874888896942,
|
| 1775 |
+
"epoch": 0.178,
|
| 1776 |
+
"grad_norm": 0.10900817571461202,
|
| 1777 |
+
"importance_ratio": 1.0003485679626465,
|
| 1778 |
+
"learning_rate": 0.0001,
|
| 1779 |
+
"loss": 0.0003,
|
| 1780 |
+
"mismatch_kl": 0.01015115063637495,
|
| 1781 |
+
"reward": 0.7599999904632568,
|
| 1782 |
+
"reward/refusal_reward_func": 0.7599999904632568,
|
| 1783 |
+
"reward/std": 0.19364915788173676,
|
| 1784 |
+
"step": 89,
|
| 1785 |
+
"timing/generation_ms": 19985.6186658144,
|
| 1786 |
+
"timing/scoring_ms": 50727.93058305979,
|
| 1787 |
+
"timing/total_ms": 70713.54924887419,
|
| 1788 |
+
"tokens/completion": 2039.15625,
|
| 1789 |
+
"tokens/masked_fraction": 0.0,
|
| 1790 |
+
"wall_clock/generate_s": 292.40805864334106
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"advantage/absmean": 0.24515625834465027,
|
| 1794 |
+
"entropy": 0.24920716881752014,
|
| 1795 |
+
"epoch": 0.18,
|
| 1796 |
+
"grad_norm": 0.12621044924209393,
|
| 1797 |
+
"importance_ratio": 1.000746250152588,
|
| 1798 |
+
"learning_rate": 0.0001,
|
| 1799 |
+
"loss": 0.0,
|
| 1800 |
+
"mismatch_kl": 0.008628414012491703,
|
| 1801 |
+
"reward": 0.6162500381469727,
|
| 1802 |
+
"reward/refusal_reward_func": 0.6162500381469727,
|
| 1803 |
+
"reward/std": 0.2840307056903839,
|
| 1804 |
+
"step": 90,
|
| 1805 |
+
"timing/generation_ms": 20263.45807313919,
|
| 1806 |
+
"timing/scoring_ms": 69397.22065627575,
|
| 1807 |
+
"timing/total_ms": 89660.67872941494,
|
| 1808 |
+
"tokens/completion": 2048.0,
|
| 1809 |
+
"tokens/masked_fraction": 0.0,
|
| 1810 |
+
"wall_clock/generate_s": 394.5488703250885
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"advantage/absmean": 0.10718750208616257,
|
| 1814 |
+
"entropy": 0.3780635893344879,
|
| 1815 |
+
"epoch": 0.182,
|
| 1816 |
+
"grad_norm": 0.041883076718277436,
|
| 1817 |
+
"importance_ratio": 1.0010097026824951,
|
| 1818 |
+
"learning_rate": 0.0001,
|
| 1819 |
+
"loss": 0.0001,
|
| 1820 |
+
"mismatch_kl": 0.009139418601989746,
|
| 1821 |
+
"reward": 0.7487499713897705,
|
| 1822 |
+
"reward/refusal_reward_func": 0.7487499713897705,
|
| 1823 |
+
"reward/std": 0.1976384073495865,
|
| 1824 |
+
"step": 91,
|
| 1825 |
+
"timing/generation_ms": 20657.530024647713,
|
| 1826 |
+
"timing/scoring_ms": 65376.88625603914,
|
| 1827 |
+
"timing/total_ms": 86034.41628068686,
|
| 1828 |
+
"tokens/completion": 2048.0,
|
| 1829 |
+
"tokens/masked_fraction": 0.0,
|
| 1830 |
+
"wall_clock/generate_s": 395.1356108188629
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"advantage/absmean": 0.13974609971046448,
|
| 1834 |
+
"entropy": 0.3699057102203369,
|
| 1835 |
+
"epoch": 0.184,
|
| 1836 |
+
"grad_norm": 0.06598040996573111,
|
| 1837 |
+
"importance_ratio": 0.9998457431793213,
|
| 1838 |
+
"learning_rate": 0.0001,
|
| 1839 |
+
"loss": 0.0,
|
| 1840 |
+
"mismatch_kl": 0.0077532450668513775,
|
| 1841 |
+
"reward": 0.7271875143051147,
|
| 1842 |
+
"reward/refusal_reward_func": 0.7271875143051147,
|
| 1843 |
+
"reward/std": 0.2168990969657898,
|
| 1844 |
+
"step": 92,
|
| 1845 |
+
"timing/generation_ms": 20616.587534546852,
|
| 1846 |
+
"timing/scoring_ms": 47124.867990612984,
|
| 1847 |
+
"timing/total_ms": 67741.45552515984,
|
| 1848 |
+
"tokens/completion": 2048.0,
|
| 1849 |
+
"tokens/masked_fraction": 0.0,
|
| 1850 |
+
"wall_clock/generate_s": 168.30030918121338
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"advantage/absmean": 0.19189453125,
|
| 1854 |
+
"entropy": 0.34180349111557007,
|
| 1855 |
+
"epoch": 0.186,
|
| 1856 |
+
"grad_norm": 0.14232699624625103,
|
| 1857 |
+
"importance_ratio": 0.999165952205658,
|
| 1858 |
+
"learning_rate": 0.0001,
|
| 1859 |
+
"loss": 0.0003,
|
| 1860 |
+
"mismatch_kl": 0.008044413290917873,
|
| 1861 |
+
"reward": 0.6871874928474426,
|
| 1862 |
+
"reward/refusal_reward_func": 0.6871874928474426,
|
| 1863 |
+
"reward/std": 0.2622169256210327,
|
| 1864 |
+
"step": 93,
|
| 1865 |
+
"timing/generation_ms": 20529.827870428562,
|
| 1866 |
+
"timing/scoring_ms": 60052.588775753975,
|
| 1867 |
+
"timing/total_ms": 80582.41664618254,
|
| 1868 |
+
"tokens/completion": 2048.0,
|
| 1869 |
+
"tokens/masked_fraction": 0.0,
|
| 1870 |
+
"wall_clock/generate_s": 396.85295939445496
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"advantage/absmean": 0.20988282561302185,
|
| 1874 |
+
"entropy": 0.315336674451828,
|
| 1875 |
+
"epoch": 0.188,
|
| 1876 |
+
"grad_norm": 0.10320818511601894,
|
| 1877 |
+
"importance_ratio": 1.000422477722168,
|
| 1878 |
+
"learning_rate": 0.0001,
|
| 1879 |
+
"loss": -0.0,
|
| 1880 |
+
"mismatch_kl": 0.007602104917168617,
|
| 1881 |
+
"reward": 0.6856250166893005,
|
| 1882 |
+
"reward/refusal_reward_func": 0.6856250166893005,
|
| 1883 |
+
"reward/std": 0.28907111287117004,
|
| 1884 |
+
"step": 94,
|
| 1885 |
+
"timing/generation_ms": 20579.548463225365,
|
| 1886 |
+
"timing/scoring_ms": 51010.259330272675,
|
| 1887 |
+
"timing/total_ms": 71589.80779349804,
|
| 1888 |
+
"tokens/completion": 2048.0,
|
| 1889 |
+
"tokens/masked_fraction": 0.0,
|
| 1890 |
+
"wall_clock/generate_s": 395.95818734169006
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"advantage/absmean": 0.15345704555511475,
|
| 1894 |
+
"entropy": 0.5262030363082886,
|
| 1895 |
+
"epoch": 0.19,
|
| 1896 |
+
"grad_norm": 0.05870104388608926,
|
| 1897 |
+
"importance_ratio": 0.9999489188194275,
|
| 1898 |
+
"learning_rate": 0.0001,
|
| 1899 |
+
"loss": 0.0002,
|
| 1900 |
+
"mismatch_kl": 0.0076672472059726715,
|
| 1901 |
+
"reward": 0.7190625071525574,
|
| 1902 |
+
"reward/refusal_reward_func": 0.7190625071525574,
|
| 1903 |
+
"reward/std": 0.2384108603000641,
|
| 1904 |
+
"step": 95,
|
| 1905 |
+
"timing/generation_ms": 20614.634588360786,
|
| 1906 |
+
"timing/scoring_ms": 65994.11156028509,
|
| 1907 |
+
"timing/total_ms": 86608.74614864588,
|
| 1908 |
+
"tokens/completion": 2048.0,
|
| 1909 |
+
"tokens/masked_fraction": 0.0,
|
| 1910 |
+
"wall_clock/generate_s": 395.0603678226471
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"advantage/absmean": 0.23779296875,
|
| 1914 |
+
"entropy": 0.3904465436935425,
|
| 1915 |
+
"epoch": 0.192,
|
| 1916 |
+
"grad_norm": 0.1935352935904031,
|
| 1917 |
+
"importance_ratio": 1.000299334526062,
|
| 1918 |
+
"learning_rate": 0.0001,
|
| 1919 |
+
"loss": 0.0,
|
| 1920 |
+
"mismatch_kl": 0.008834589272737503,
|
| 1921 |
+
"reward": 0.6578124761581421,
|
| 1922 |
+
"reward/refusal_reward_func": 0.6578124761581421,
|
| 1923 |
+
"reward/std": 0.2928708493709564,
|
| 1924 |
+
"step": 96,
|
| 1925 |
+
"timing/generation_ms": 20604.460656642914,
|
| 1926 |
+
"timing/scoring_ms": 67186.7751404643,
|
| 1927 |
+
"timing/total_ms": 87791.23579710722,
|
| 1928 |
+
"tokens/completion": 2048.0,
|
| 1929 |
+
"tokens/masked_fraction": 0.0,
|
| 1930 |
+
"wall_clock/generate_s": 394.96749925613403
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"advantage/absmean": 0.2852538824081421,
|
| 1934 |
+
"entropy": 0.32257264852523804,
|
| 1935 |
+
"epoch": 0.194,
|
| 1936 |
+
"grad_norm": 0.11248909611986616,
|
| 1937 |
+
"importance_ratio": 0.9994723796844482,
|
| 1938 |
+
"learning_rate": 0.0001,
|
| 1939 |
+
"loss": -0.0,
|
| 1940 |
+
"mismatch_kl": 0.008222612552344799,
|
| 1941 |
+
"reward": 0.6115624904632568,
|
| 1942 |
+
"reward/refusal_reward_func": 0.6115624904632568,
|
| 1943 |
+
"reward/std": 0.3214528560638428,
|
| 1944 |
+
"step": 97,
|
| 1945 |
+
"timing/generation_ms": 20211.18316054344,
|
| 1946 |
+
"timing/scoring_ms": 60143.2975307107,
|
| 1947 |
+
"timing/total_ms": 80354.48069125414,
|
| 1948 |
+
"tokens/completion": 2048.0,
|
| 1949 |
+
"tokens/masked_fraction": 0.0,
|
| 1950 |
+
"wall_clock/generate_s": 395.2164263725281
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"advantage/absmean": 0.09375,
|
| 1954 |
+
"entropy": 0.45740675926208496,
|
| 1955 |
+
"epoch": 0.196,
|
| 1956 |
+
"grad_norm": 0.12144158106340411,
|
| 1957 |
+
"importance_ratio": 0.9998034834861755,
|
| 1958 |
+
"learning_rate": 0.0001,
|
| 1959 |
+
"loss": 0.0001,
|
| 1960 |
+
"mismatch_kl": 0.008089970797300339,
|
| 1961 |
+
"reward": 0.7599999904632568,
|
| 1962 |
+
"reward/refusal_reward_func": 0.7599999904632568,
|
| 1963 |
+
"reward/std": 0.19364915788173676,
|
| 1964 |
+
"step": 98,
|
| 1965 |
+
"timing/generation_ms": 20260.59687882662,
|
| 1966 |
+
"timing/scoring_ms": 44623.64313751459,
|
| 1967 |
+
"timing/total_ms": 64884.24001634121,
|
| 1968 |
+
"tokens/completion": 2048.0,
|
| 1969 |
+
"tokens/masked_fraction": 0.0,
|
| 1970 |
+
"wall_clock/generate_s": 131.37106108665466
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"advantage/absmean": 0.04843749850988388,
|
| 1974 |
+
"entropy": 0.5326197147369385,
|
| 1975 |
+
"epoch": 0.198,
|
| 1976 |
+
"grad_norm": 0.12835217845348193,
|
| 1977 |
+
"importance_ratio": 0.9994455575942993,
|
| 1978 |
+
"learning_rate": 0.0001,
|
| 1979 |
+
"loss": -0.0,
|
| 1980 |
+
"mismatch_kl": 0.009172793477773666,
|
| 1981 |
+
"reward": 0.7849999666213989,
|
| 1982 |
+
"reward/refusal_reward_func": 0.7849999666213989,
|
| 1983 |
+
"reward/std": 0.13919411599636078,
|
| 1984 |
+
"step": 99,
|
| 1985 |
+
"timing/generation_ms": 20207.647144794464,
|
| 1986 |
+
"timing/scoring_ms": 49733.94272476435,
|
| 1987 |
+
"timing/total_ms": 69941.58986955881,
|
| 1988 |
+
"tokens/completion": 2048.0,
|
| 1989 |
+
"tokens/masked_fraction": 0.0,
|
| 1990 |
+
"wall_clock/generate_s": 161.61080026626587
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"advantage/absmean": 0.10171875357627869,
|
| 1994 |
+
"entropy": 0.5042125582695007,
|
| 1995 |
+
"epoch": 0.2,
|
| 1996 |
+
"grad_norm": 0.08743211269759214,
|
| 1997 |
+
"importance_ratio": 1.0003973245620728,
|
| 1998 |
+
"learning_rate": 0.0001,
|
| 1999 |
+
"loss": -0.0002,
|
| 2000 |
+
"mismatch_kl": 0.01178868766874075,
|
| 2001 |
+
"reward": 0.7518749833106995,
|
| 2002 |
+
"reward/refusal_reward_func": 0.7518749833106995,
|
| 2003 |
+
"reward/std": 0.17614690959453583,
|
| 2004 |
+
"step": 100,
|
| 2005 |
+
"timing/generation_ms": 20365.172304213047,
|
| 2006 |
+
"timing/scoring_ms": 65265.56546241045,
|
| 2007 |
+
"timing/total_ms": 85630.7377666235,
|
| 2008 |
+
"tokens/completion": 2048.0,
|
| 2009 |
+
"tokens/masked_fraction": 0.0,
|
| 2010 |
+
"wall_clock/generate_s": 394.74688720703125
|
| 2011 |
+
}
|
| 2012 |
+
],
|
| 2013 |
+
"logging_steps": 1,
|
| 2014 |
+
"max_steps": 500,
|
| 2015 |
+
"num_input_tokens_seen": 0,
|
| 2016 |
+
"num_train_epochs": 1,
|
| 2017 |
+
"save_steps": 100,
|
| 2018 |
+
"stateful_callbacks": {
|
| 2019 |
+
"TrainerControl": {
|
| 2020 |
+
"args": {
|
| 2021 |
+
"should_epoch_stop": false,
|
| 2022 |
+
"should_evaluate": false,
|
| 2023 |
+
"should_log": false,
|
| 2024 |
+
"should_save": true,
|
| 2025 |
+
"should_training_stop": false
|
| 2026 |
+
},
|
| 2027 |
+
"attributes": {}
|
| 2028 |
+
}
|
| 2029 |
+
},
|
| 2030 |
+
"total_flos": 0.0,
|
| 2031 |
+
"train_batch_size": 4,
|
| 2032 |
+
"trial_name": null,
|
| 2033 |
+
"trial_params": null
|
| 2034 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faac6e2dc8c57c18fbf4e65139a1a87a4ce991484f1c480da9878818cb07d4fe
|
| 3 |
+
size 9681
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|