Upload sfo model
Browse files- .gitattributes +2 -0
- README.md +67 -0
- added_tokens.json +24 -0
- chat_template.jinja +54 -0
- checkpoint-107/added_tokens.json +24 -0
- checkpoint-107/chat_template.jinja +54 -0
- checkpoint-107/config.json +58 -0
- checkpoint-107/generation_config.json +9 -0
- checkpoint-107/merges.txt +0 -0
- checkpoint-107/model.safetensors +3 -0
- checkpoint-107/optimizer.pt +3 -0
- checkpoint-107/rng_state.pth +3 -0
- checkpoint-107/scheduler.pt +3 -0
- checkpoint-107/special_tokens_map.json +31 -0
- checkpoint-107/tokenizer.json +3 -0
- checkpoint-107/tokenizer_config.json +207 -0
- checkpoint-107/trainer_state.json +2282 -0
- checkpoint-107/training_args.bin +3 -0
- checkpoint-107/vocab.json +0 -0
- config.json +58 -0
- generation_config.json +9 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- special_tokens_map.json +31 -0
- tokenizer.json +3 -0
- tokenizer_config.json +207 -0
- training_args.bin +3 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,5 @@ 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 |
+
checkpoint-107/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: MMattaparthy/sft_finetined_final
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: ppo_final_model
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- ppo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for ppo_final_model
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [MMattaparthy/sft_finetined_final](https://huggingface.co/MMattaparthy/sft_finetined_final).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="None", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with PPO, a method introduced in [Fine-Tuning Language Models from Human Preferences](https://huggingface.co/papers/1909.08593).
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.24.0
|
| 38 |
+
- Transformers: 4.57.1
|
| 39 |
+
- Pytorch: 2.8.0+cu126
|
| 40 |
+
- Datasets: 4.0.0
|
| 41 |
+
- Tokenizers: 0.22.1
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite PPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{mziegler2019fine-tuning,
|
| 49 |
+
title = {{Fine-Tuning Language Models from Human Preferences}},
|
| 50 |
+
author = {Daniel M. Ziegler and Nisan Stiennon and Jeffrey Wu and Tom B. Brown and Alec Radford and Dario Amodei and Paul F. Christiano and Geoffrey Irving},
|
| 51 |
+
year = 2019,
|
| 52 |
+
eprint = {arXiv:1909.08593}
|
| 53 |
+
}
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
Cite TRL as:
|
| 57 |
+
|
| 58 |
+
```bibtex
|
| 59 |
+
@misc{vonwerra2022trl,
|
| 60 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 61 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
|
| 62 |
+
year = 2020,
|
| 63 |
+
journal = {GitHub repository},
|
| 64 |
+
publisher = {GitHub},
|
| 65 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 66 |
+
}
|
| 67 |
+
```
|
added_tokens.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</tool_call>": 151658,
|
| 3 |
+
"<tool_call>": 151657,
|
| 4 |
+
"<|box_end|>": 151649,
|
| 5 |
+
"<|box_start|>": 151648,
|
| 6 |
+
"<|endoftext|>": 151643,
|
| 7 |
+
"<|file_sep|>": 151664,
|
| 8 |
+
"<|fim_middle|>": 151660,
|
| 9 |
+
"<|fim_pad|>": 151662,
|
| 10 |
+
"<|fim_prefix|>": 151659,
|
| 11 |
+
"<|fim_suffix|>": 151661,
|
| 12 |
+
"<|im_end|>": 151645,
|
| 13 |
+
"<|im_start|>": 151644,
|
| 14 |
+
"<|image_pad|>": 151655,
|
| 15 |
+
"<|object_ref_end|>": 151647,
|
| 16 |
+
"<|object_ref_start|>": 151646,
|
| 17 |
+
"<|quad_end|>": 151651,
|
| 18 |
+
"<|quad_start|>": 151650,
|
| 19 |
+
"<|repo_name|>": 151663,
|
| 20 |
+
"<|video_pad|>": 151656,
|
| 21 |
+
"<|vision_end|>": 151653,
|
| 22 |
+
"<|vision_pad|>": 151654,
|
| 23 |
+
"<|vision_start|>": 151652
|
| 24 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\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" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
checkpoint-107/added_tokens.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</tool_call>": 151658,
|
| 3 |
+
"<tool_call>": 151657,
|
| 4 |
+
"<|box_end|>": 151649,
|
| 5 |
+
"<|box_start|>": 151648,
|
| 6 |
+
"<|endoftext|>": 151643,
|
| 7 |
+
"<|file_sep|>": 151664,
|
| 8 |
+
"<|fim_middle|>": 151660,
|
| 9 |
+
"<|fim_pad|>": 151662,
|
| 10 |
+
"<|fim_prefix|>": 151659,
|
| 11 |
+
"<|fim_suffix|>": 151661,
|
| 12 |
+
"<|im_end|>": 151645,
|
| 13 |
+
"<|im_start|>": 151644,
|
| 14 |
+
"<|image_pad|>": 151655,
|
| 15 |
+
"<|object_ref_end|>": 151647,
|
| 16 |
+
"<|object_ref_start|>": 151646,
|
| 17 |
+
"<|quad_end|>": 151651,
|
| 18 |
+
"<|quad_start|>": 151650,
|
| 19 |
+
"<|repo_name|>": 151663,
|
| 20 |
+
"<|video_pad|>": 151656,
|
| 21 |
+
"<|vision_end|>": 151653,
|
| 22 |
+
"<|vision_pad|>": 151654,
|
| 23 |
+
"<|vision_start|>": 151652
|
| 24 |
+
}
|
checkpoint-107/chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\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" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
checkpoint-107/config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen2ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"dtype": "bfloat16",
|
| 7 |
+
"eos_token_id": 151645,
|
| 8 |
+
"hidden_act": "silu",
|
| 9 |
+
"hidden_size": 1536,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 8960,
|
| 12 |
+
"layer_types": [
|
| 13 |
+
"full_attention",
|
| 14 |
+
"full_attention",
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention"
|
| 41 |
+
],
|
| 42 |
+
"max_position_embeddings": 32768,
|
| 43 |
+
"max_window_layers": 21,
|
| 44 |
+
"model_type": "qwen2",
|
| 45 |
+
"num_attention_heads": 12,
|
| 46 |
+
"num_hidden_layers": 28,
|
| 47 |
+
"num_key_value_heads": 2,
|
| 48 |
+
"pad_token_id": 151643,
|
| 49 |
+
"rms_norm_eps": 1e-06,
|
| 50 |
+
"rope_scaling": null,
|
| 51 |
+
"rope_theta": 1000000.0,
|
| 52 |
+
"sliding_window": null,
|
| 53 |
+
"tie_word_embeddings": true,
|
| 54 |
+
"transformers_version": "4.57.1",
|
| 55 |
+
"use_cache": false,
|
| 56 |
+
"use_sliding_window": false,
|
| 57 |
+
"vocab_size": 151936
|
| 58 |
+
}
|
checkpoint-107/generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"pad_token_id": 151643,
|
| 4 |
+
"repetition_penalty": 1.1,
|
| 5 |
+
"temperature": 0.7,
|
| 6 |
+
"top_k": 20,
|
| 7 |
+
"top_p": 0.8,
|
| 8 |
+
"transformers_version": "4.57.1"
|
| 9 |
+
}
|
checkpoint-107/merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-107/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c93779623565bec515d60d0d5ba3ac8bb8d891825cd93bd6ffb072dbf265550
|
| 3 |
+
size 3087467144
|
checkpoint-107/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3cf57a2c37522bc1897e9890c7c7403fc7f35354df7532984151af4cbfc2c3c5
|
| 3 |
+
size 12350320650
|
checkpoint-107/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec9e44811f24164e11bec7b150f7b4179380583f524ef1a6d16b4838107e5d16
|
| 3 |
+
size 14709
|
checkpoint-107/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9a67ed9857c6e6351f390426ac7a185bbcb92711588dc22a658de8cf5976ff88
|
| 3 |
+
size 1465
|
checkpoint-107/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 |
+
}
|
checkpoint-107/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f650899cdee1a74abb093b96e45774316750373414f9e11647b8480290d3937f
|
| 3 |
+
size 11421988
|
checkpoint-107/tokenizer_config.json
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
},
|
| 182 |
+
"additional_special_tokens": [
|
| 183 |
+
"<|im_start|>",
|
| 184 |
+
"<|im_end|>",
|
| 185 |
+
"<|object_ref_start|>",
|
| 186 |
+
"<|object_ref_end|>",
|
| 187 |
+
"<|box_start|>",
|
| 188 |
+
"<|box_end|>",
|
| 189 |
+
"<|quad_start|>",
|
| 190 |
+
"<|quad_end|>",
|
| 191 |
+
"<|vision_start|>",
|
| 192 |
+
"<|vision_end|>",
|
| 193 |
+
"<|vision_pad|>",
|
| 194 |
+
"<|image_pad|>",
|
| 195 |
+
"<|video_pad|>"
|
| 196 |
+
],
|
| 197 |
+
"bos_token": null,
|
| 198 |
+
"clean_up_tokenization_spaces": false,
|
| 199 |
+
"eos_token": "<|im_end|>",
|
| 200 |
+
"errors": "replace",
|
| 201 |
+
"extra_special_tokens": {},
|
| 202 |
+
"model_max_length": 131072,
|
| 203 |
+
"pad_token": "<|endoftext|>",
|
| 204 |
+
"split_special_tokens": false,
|
| 205 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 206 |
+
"unk_token": null
|
| 207 |
+
}
|
checkpoint-107/trainer_state.json
ADDED
|
@@ -0,0 +1,2282 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"episode": 1712,
|
| 6 |
+
"epoch": 3.0035087719298246,
|
| 7 |
+
"eval_steps": 500,
|
| 8 |
+
"global_step": 107,
|
| 9 |
+
"is_hyper_param_search": false,
|
| 10 |
+
"is_local_process_zero": true,
|
| 11 |
+
"is_world_process_zero": true,
|
| 12 |
+
"log_history": [
|
| 13 |
+
{
|
| 14 |
+
"episode": 16,
|
| 15 |
+
"epoch": 0.028070175438596492,
|
| 16 |
+
"eps": 2,
|
| 17 |
+
"loss/policy_avg": 0.040254347026348114,
|
| 18 |
+
"loss/value_avg": 4.365694046020508,
|
| 19 |
+
"lr": 1.41e-05,
|
| 20 |
+
"objective/entropy": 20.665752410888672,
|
| 21 |
+
"objective/kl": 38.31879425048828,
|
| 22 |
+
"objective/non_score_reward": -1.9159398078918457,
|
| 23 |
+
"objective/rlhf_reward": -0.5565648078918457,
|
| 24 |
+
"objective/scores": 1.359375,
|
| 25 |
+
"policy/approxkl_avg": 5.953035354614258,
|
| 26 |
+
"policy/clipfrac_avg": 0.2399764209985733,
|
| 27 |
+
"policy/entropy_avg": 0.6761884689331055,
|
| 28 |
+
"step": 1,
|
| 29 |
+
"val/clipfrac_avg": 0.26650944352149963,
|
| 30 |
+
"val/num_eos_tokens": 0,
|
| 31 |
+
"val/ratio": 0.8525064587593079,
|
| 32 |
+
"val/ratio_var": 0.0006113838753663003
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"episode": 32,
|
| 36 |
+
"epoch": 0.056140350877192984,
|
| 37 |
+
"eps": 2,
|
| 38 |
+
"loss/policy_avg": 0.0446447990834713,
|
| 39 |
+
"loss/value_avg": 3.0273280143737793,
|
| 40 |
+
"lr": 1.3968224299065421e-05,
|
| 41 |
+
"objective/entropy": 21.37685203552246,
|
| 42 |
+
"objective/kl": 73.40159606933594,
|
| 43 |
+
"objective/non_score_reward": -3.6700797080993652,
|
| 44 |
+
"objective/rlhf_reward": -2.5294547080993652,
|
| 45 |
+
"objective/scores": 1.140625,
|
| 46 |
+
"policy/approxkl_avg": 6.8038129806518555,
|
| 47 |
+
"policy/clipfrac_avg": 0.22936320304870605,
|
| 48 |
+
"policy/entropy_avg": 0.6141112446784973,
|
| 49 |
+
"step": 2,
|
| 50 |
+
"val/clipfrac_avg": 0.2458726465702057,
|
| 51 |
+
"val/num_eos_tokens": 0,
|
| 52 |
+
"val/ratio": 0.8650269508361816,
|
| 53 |
+
"val/ratio_var": 0.0003330775070935488
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"episode": 48,
|
| 57 |
+
"epoch": 0.08421052631578947,
|
| 58 |
+
"eps": 3,
|
| 59 |
+
"loss/policy_avg": 0.07702315598726273,
|
| 60 |
+
"loss/value_avg": 2.1064209938049316,
|
| 61 |
+
"lr": 1.3836448598130842e-05,
|
| 62 |
+
"objective/entropy": 26.10182762145996,
|
| 63 |
+
"objective/kl": 74.03025817871094,
|
| 64 |
+
"objective/non_score_reward": -3.7015130519866943,
|
| 65 |
+
"objective/rlhf_reward": -2.6546380519866943,
|
| 66 |
+
"objective/scores": 1.046875,
|
| 67 |
+
"policy/approxkl_avg": 5.6432695388793945,
|
| 68 |
+
"policy/clipfrac_avg": 0.2087264060974121,
|
| 69 |
+
"policy/entropy_avg": 0.7391780614852905,
|
| 70 |
+
"step": 3,
|
| 71 |
+
"val/clipfrac_avg": 0.21639150381088257,
|
| 72 |
+
"val/num_eos_tokens": 0,
|
| 73 |
+
"val/ratio": 0.8773487210273743,
|
| 74 |
+
"val/ratio_var": 0.00039911610656417906
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"episode": 64,
|
| 78 |
+
"epoch": 0.11228070175438597,
|
| 79 |
+
"eps": 3,
|
| 80 |
+
"loss/policy_avg": 0.03978118300437927,
|
| 81 |
+
"loss/value_avg": 1.5315862894058228,
|
| 82 |
+
"lr": 1.3704672897196262e-05,
|
| 83 |
+
"objective/entropy": 20.781639099121094,
|
| 84 |
+
"objective/kl": 68.94087219238281,
|
| 85 |
+
"objective/non_score_reward": -3.4470434188842773,
|
| 86 |
+
"objective/rlhf_reward": -2.4470434188842773,
|
| 87 |
+
"objective/scores": 1.0,
|
| 88 |
+
"policy/approxkl_avg": 4.726678848266602,
|
| 89 |
+
"policy/clipfrac_avg": 0.20341980457305908,
|
| 90 |
+
"policy/entropy_avg": 0.7119085788726807,
|
| 91 |
+
"step": 4,
|
| 92 |
+
"val/clipfrac_avg": 0.16096699237823486,
|
| 93 |
+
"val/num_eos_tokens": 0,
|
| 94 |
+
"val/ratio": 0.8485063314437866,
|
| 95 |
+
"val/ratio_var": 0.0003752955235540867
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"episode": 80,
|
| 99 |
+
"epoch": 0.14035087719298245,
|
| 100 |
+
"eps": 3,
|
| 101 |
+
"loss/policy_avg": 0.06184825301170349,
|
| 102 |
+
"loss/value_avg": 0.9904976487159729,
|
| 103 |
+
"lr": 1.3572897196261683e-05,
|
| 104 |
+
"objective/entropy": 22.48508071899414,
|
| 105 |
+
"objective/kl": 75.60542297363281,
|
| 106 |
+
"objective/non_score_reward": -3.780271291732788,
|
| 107 |
+
"objective/rlhf_reward": -3.272458791732788,
|
| 108 |
+
"objective/scores": 0.5078125,
|
| 109 |
+
"policy/approxkl_avg": 4.352260589599609,
|
| 110 |
+
"policy/clipfrac_avg": 0.2146226465702057,
|
| 111 |
+
"policy/entropy_avg": 0.7200251817703247,
|
| 112 |
+
"step": 5,
|
| 113 |
+
"val/clipfrac_avg": 0.09669811278581619,
|
| 114 |
+
"val/num_eos_tokens": 0,
|
| 115 |
+
"val/ratio": 0.8546276092529297,
|
| 116 |
+
"val/ratio_var": 0.00010198648669756949
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"episode": 96,
|
| 120 |
+
"epoch": 0.16842105263157894,
|
| 121 |
+
"eps": 3,
|
| 122 |
+
"loss/policy_avg": 0.046236515045166016,
|
| 123 |
+
"loss/value_avg": 1.2219572067260742,
|
| 124 |
+
"lr": 1.3441121495327103e-05,
|
| 125 |
+
"objective/entropy": 22.197113037109375,
|
| 126 |
+
"objective/kl": 82.56051635742188,
|
| 127 |
+
"objective/non_score_reward": -4.128026008605957,
|
| 128 |
+
"objective/rlhf_reward": -3.253026008605957,
|
| 129 |
+
"objective/scores": 0.875,
|
| 130 |
+
"policy/approxkl_avg": 3.8577029705047607,
|
| 131 |
+
"policy/clipfrac_avg": 0.22641509771347046,
|
| 132 |
+
"policy/entropy_avg": 0.7238099575042725,
|
| 133 |
+
"step": 6,
|
| 134 |
+
"val/clipfrac_avg": 0.0383254736661911,
|
| 135 |
+
"val/num_eos_tokens": 0,
|
| 136 |
+
"val/ratio": 0.8292837738990784,
|
| 137 |
+
"val/ratio_var": 6.185401434777305e-05
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"episode": 112,
|
| 141 |
+
"epoch": 0.19649122807017544,
|
| 142 |
+
"eps": 3,
|
| 143 |
+
"loss/policy_avg": 0.06606701016426086,
|
| 144 |
+
"loss/value_avg": 1.325202226638794,
|
| 145 |
+
"lr": 1.3309345794392524e-05,
|
| 146 |
+
"objective/entropy": 27.022249221801758,
|
| 147 |
+
"objective/kl": 101.50897216796875,
|
| 148 |
+
"objective/non_score_reward": -5.075448513031006,
|
| 149 |
+
"objective/rlhf_reward": -3.653573513031006,
|
| 150 |
+
"objective/scores": 1.421875,
|
| 151 |
+
"policy/approxkl_avg": 5.585095405578613,
|
| 152 |
+
"policy/clipfrac_avg": 0.24174529314041138,
|
| 153 |
+
"policy/entropy_avg": 0.8859995603561401,
|
| 154 |
+
"step": 7,
|
| 155 |
+
"val/clipfrac_avg": 0.028301887214183807,
|
| 156 |
+
"val/num_eos_tokens": 0,
|
| 157 |
+
"val/ratio": 0.8486474752426147,
|
| 158 |
+
"val/ratio_var": 0.0005388148711062968
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"episode": 128,
|
| 162 |
+
"epoch": 0.22456140350877193,
|
| 163 |
+
"eps": 3,
|
| 164 |
+
"loss/policy_avg": 0.06467999517917633,
|
| 165 |
+
"loss/value_avg": 0.9481044411659241,
|
| 166 |
+
"lr": 1.3177570093457945e-05,
|
| 167 |
+
"objective/entropy": 26.585155487060547,
|
| 168 |
+
"objective/kl": 116.76457214355469,
|
| 169 |
+
"objective/non_score_reward": -5.838229179382324,
|
| 170 |
+
"objective/rlhf_reward": -4.111666679382324,
|
| 171 |
+
"objective/scores": 1.7265625,
|
| 172 |
+
"policy/approxkl_avg": 5.351860046386719,
|
| 173 |
+
"policy/clipfrac_avg": 0.24469339847564697,
|
| 174 |
+
"policy/entropy_avg": 0.8991943597793579,
|
| 175 |
+
"step": 8,
|
| 176 |
+
"val/clipfrac_avg": 0.08785377442836761,
|
| 177 |
+
"val/num_eos_tokens": 0,
|
| 178 |
+
"val/ratio": 0.8524093627929688,
|
| 179 |
+
"val/ratio_var": 3.493872281978838e-05
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"episode": 144,
|
| 183 |
+
"epoch": 0.25263157894736843,
|
| 184 |
+
"eps": 3,
|
| 185 |
+
"loss/policy_avg": 0.03949737548828125,
|
| 186 |
+
"loss/value_avg": 0.9386715888977051,
|
| 187 |
+
"lr": 1.3045794392523365e-05,
|
| 188 |
+
"objective/entropy": 29.664459228515625,
|
| 189 |
+
"objective/kl": 132.19940185546875,
|
| 190 |
+
"objective/non_score_reward": -6.6099700927734375,
|
| 191 |
+
"objective/rlhf_reward": -5.6763763427734375,
|
| 192 |
+
"objective/scores": 0.93359375,
|
| 193 |
+
"policy/approxkl_avg": 5.967764377593994,
|
| 194 |
+
"policy/clipfrac_avg": 0.2057783007621765,
|
| 195 |
+
"policy/entropy_avg": 1.0122931003570557,
|
| 196 |
+
"step": 9,
|
| 197 |
+
"val/clipfrac_avg": 0.017099056392908096,
|
| 198 |
+
"val/num_eos_tokens": 0,
|
| 199 |
+
"val/ratio": 0.8205329179763794,
|
| 200 |
+
"val/ratio_var": 7.763963367324322e-05
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"episode": 160,
|
| 204 |
+
"epoch": 0.2807017543859649,
|
| 205 |
+
"eps": 3,
|
| 206 |
+
"loss/policy_avg": 0.0211679395288229,
|
| 207 |
+
"loss/value_avg": 1.213599681854248,
|
| 208 |
+
"lr": 1.2914018691588786e-05,
|
| 209 |
+
"objective/entropy": 31.076860427856445,
|
| 210 |
+
"objective/kl": 135.477294921875,
|
| 211 |
+
"objective/non_score_reward": -6.77386474609375,
|
| 212 |
+
"objective/rlhf_reward": -5.00042724609375,
|
| 213 |
+
"objective/scores": 1.7734375,
|
| 214 |
+
"policy/approxkl_avg": 3.5569186210632324,
|
| 215 |
+
"policy/clipfrac_avg": 0.21304652094841003,
|
| 216 |
+
"policy/entropy_avg": 1.1141610145568848,
|
| 217 |
+
"step": 10,
|
| 218 |
+
"val/clipfrac_avg": 0.06719152629375458,
|
| 219 |
+
"val/num_eos_tokens": 0,
|
| 220 |
+
"val/ratio": 0.8466310501098633,
|
| 221 |
+
"val/ratio_var": 0.0001414915022905916
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"episode": 176,
|
| 225 |
+
"epoch": 0.3087719298245614,
|
| 226 |
+
"eps": 3,
|
| 227 |
+
"loss/policy_avg": 0.0240048635751009,
|
| 228 |
+
"loss/value_avg": 1.3511584997177124,
|
| 229 |
+
"lr": 1.2782242990654206e-05,
|
| 230 |
+
"objective/entropy": 32.90422058105469,
|
| 231 |
+
"objective/kl": 146.03787231445312,
|
| 232 |
+
"objective/non_score_reward": -7.301893711090088,
|
| 233 |
+
"objective/rlhf_reward": -5.630018711090088,
|
| 234 |
+
"objective/scores": 1.671875,
|
| 235 |
+
"policy/approxkl_avg": 5.625584125518799,
|
| 236 |
+
"policy/clipfrac_avg": 0.20400942862033844,
|
| 237 |
+
"policy/entropy_avg": 1.1266499757766724,
|
| 238 |
+
"step": 11,
|
| 239 |
+
"val/clipfrac_avg": 0.05483490601181984,
|
| 240 |
+
"val/num_eos_tokens": 0,
|
| 241 |
+
"val/ratio": 0.8710312843322754,
|
| 242 |
+
"val/ratio_var": 0.00014447471767198294
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"episode": 192,
|
| 246 |
+
"epoch": 0.3368421052631579,
|
| 247 |
+
"eps": 3,
|
| 248 |
+
"loss/policy_avg": 0.05039631202816963,
|
| 249 |
+
"loss/value_avg": 1.4507163763046265,
|
| 250 |
+
"lr": 1.2650467289719627e-05,
|
| 251 |
+
"objective/entropy": 34.778846740722656,
|
| 252 |
+
"objective/kl": 152.17945861816406,
|
| 253 |
+
"objective/non_score_reward": -7.608973503112793,
|
| 254 |
+
"objective/rlhf_reward": -5.726161003112793,
|
| 255 |
+
"objective/scores": 1.8828125,
|
| 256 |
+
"policy/approxkl_avg": 4.985030174255371,
|
| 257 |
+
"policy/clipfrac_avg": 0.1987028270959854,
|
| 258 |
+
"policy/entropy_avg": 1.1480720043182373,
|
| 259 |
+
"step": 12,
|
| 260 |
+
"val/clipfrac_avg": 0.05188679322600365,
|
| 261 |
+
"val/num_eos_tokens": 0,
|
| 262 |
+
"val/ratio": 0.8588758111000061,
|
| 263 |
+
"val/ratio_var": 3.4296579542569816e-05
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"episode": 208,
|
| 267 |
+
"epoch": 0.3649122807017544,
|
| 268 |
+
"eps": 3,
|
| 269 |
+
"loss/policy_avg": 0.037938639521598816,
|
| 270 |
+
"loss/value_avg": 1.5314528942108154,
|
| 271 |
+
"lr": 1.2518691588785048e-05,
|
| 272 |
+
"objective/entropy": 45.65214538574219,
|
| 273 |
+
"objective/kl": 132.1256866455078,
|
| 274 |
+
"objective/non_score_reward": -6.606284141540527,
|
| 275 |
+
"objective/rlhf_reward": -6.803549766540527,
|
| 276 |
+
"objective/scores": -0.197265625,
|
| 277 |
+
"policy/approxkl_avg": 5.0863566398620605,
|
| 278 |
+
"policy/clipfrac_avg": 0.18867924809455872,
|
| 279 |
+
"policy/entropy_avg": 1.3655226230621338,
|
| 280 |
+
"step": 13,
|
| 281 |
+
"val/clipfrac_avg": 0.06367924809455872,
|
| 282 |
+
"val/num_eos_tokens": 0,
|
| 283 |
+
"val/ratio": 0.8282334804534912,
|
| 284 |
+
"val/ratio_var": 0.0004519254434853792
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"episode": 224,
|
| 288 |
+
"epoch": 0.3929824561403509,
|
| 289 |
+
"eps": 3,
|
| 290 |
+
"loss/policy_avg": 0.036732763051986694,
|
| 291 |
+
"loss/value_avg": 2.189356803894043,
|
| 292 |
+
"lr": 1.2386915887850468e-05,
|
| 293 |
+
"objective/entropy": 38.35211944580078,
|
| 294 |
+
"objective/kl": 113.13496398925781,
|
| 295 |
+
"objective/non_score_reward": -5.656748294830322,
|
| 296 |
+
"objective/rlhf_reward": -4.539560794830322,
|
| 297 |
+
"objective/scores": 1.1171875,
|
| 298 |
+
"policy/approxkl_avg": 4.408792972564697,
|
| 299 |
+
"policy/clipfrac_avg": 0.21201542019844055,
|
| 300 |
+
"policy/entropy_avg": 1.2410473823547363,
|
| 301 |
+
"step": 14,
|
| 302 |
+
"val/clipfrac_avg": 0.10222071409225464,
|
| 303 |
+
"val/num_eos_tokens": 0,
|
| 304 |
+
"val/ratio": 0.8754662275314331,
|
| 305 |
+
"val/ratio_var": 0.0007480247295461595
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"episode": 240,
|
| 309 |
+
"epoch": 0.42105263157894735,
|
| 310 |
+
"eps": 3,
|
| 311 |
+
"loss/policy_avg": 0.03690744563937187,
|
| 312 |
+
"loss/value_avg": 1.4401739835739136,
|
| 313 |
+
"lr": 1.2255140186915889e-05,
|
| 314 |
+
"objective/entropy": 38.69694519042969,
|
| 315 |
+
"objective/kl": 115.77085876464844,
|
| 316 |
+
"objective/non_score_reward": -5.788543224334717,
|
| 317 |
+
"objective/rlhf_reward": -5.627410411834717,
|
| 318 |
+
"objective/scores": 0.1611328125,
|
| 319 |
+
"policy/approxkl_avg": 4.673148155212402,
|
| 320 |
+
"policy/clipfrac_avg": 0.1833726465702057,
|
| 321 |
+
"policy/entropy_avg": 1.2108347415924072,
|
| 322 |
+
"step": 15,
|
| 323 |
+
"val/clipfrac_avg": 0.01179245300590992,
|
| 324 |
+
"val/num_eos_tokens": 0,
|
| 325 |
+
"val/ratio": 0.8487275838851929,
|
| 326 |
+
"val/ratio_var": 0.0008522844291292131
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"episode": 256,
|
| 330 |
+
"epoch": 0.44912280701754387,
|
| 331 |
+
"eps": 3,
|
| 332 |
+
"loss/policy_avg": 0.019956424832344055,
|
| 333 |
+
"loss/value_avg": 1.5383408069610596,
|
| 334 |
+
"lr": 1.212336448598131e-05,
|
| 335 |
+
"objective/entropy": 32.02195739746094,
|
| 336 |
+
"objective/kl": 122.87109375,
|
| 337 |
+
"objective/non_score_reward": -6.1435546875,
|
| 338 |
+
"objective/rlhf_reward": -5.7470703125,
|
| 339 |
+
"objective/scores": 0.396484375,
|
| 340 |
+
"policy/approxkl_avg": 5.454257965087891,
|
| 341 |
+
"policy/clipfrac_avg": 0.22314535081386566,
|
| 342 |
+
"policy/entropy_avg": 1.1098759174346924,
|
| 343 |
+
"step": 16,
|
| 344 |
+
"val/clipfrac_avg": 0.03384597226977348,
|
| 345 |
+
"val/num_eos_tokens": 0,
|
| 346 |
+
"val/ratio": 0.8839770555496216,
|
| 347 |
+
"val/ratio_var": 0.0004538022622000426
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"episode": 272,
|
| 351 |
+
"epoch": 0.47719298245614034,
|
| 352 |
+
"eps": 3,
|
| 353 |
+
"loss/policy_avg": 0.016002152115106583,
|
| 354 |
+
"loss/value_avg": 1.5554126501083374,
|
| 355 |
+
"lr": 1.199158878504673e-05,
|
| 356 |
+
"objective/entropy": 28.698740005493164,
|
| 357 |
+
"objective/kl": 129.80386352539062,
|
| 358 |
+
"objective/non_score_reward": -6.4901933670043945,
|
| 359 |
+
"objective/rlhf_reward": -6.1796464920043945,
|
| 360 |
+
"objective/scores": 0.310546875,
|
| 361 |
+
"policy/approxkl_avg": 3.0268969535827637,
|
| 362 |
+
"policy/clipfrac_avg": 0.19969519972801208,
|
| 363 |
+
"policy/entropy_avg": 0.9890843629837036,
|
| 364 |
+
"step": 17,
|
| 365 |
+
"val/clipfrac_avg": 0.07565668225288391,
|
| 366 |
+
"val/num_eos_tokens": 0,
|
| 367 |
+
"val/ratio": 0.8847370743751526,
|
| 368 |
+
"val/ratio_var": 0.000195541579159908
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"episode": 288,
|
| 372 |
+
"epoch": 0.5052631578947369,
|
| 373 |
+
"eps": 3,
|
| 374 |
+
"loss/policy_avg": 0.013964798301458359,
|
| 375 |
+
"loss/value_avg": 1.590545892715454,
|
| 376 |
+
"lr": 1.185981308411215e-05,
|
| 377 |
+
"objective/entropy": 12.885665893554688,
|
| 378 |
+
"objective/kl": 107.77202606201172,
|
| 379 |
+
"objective/non_score_reward": -5.388601303100586,
|
| 380 |
+
"objective/rlhf_reward": -4.169851303100586,
|
| 381 |
+
"objective/scores": 1.21875,
|
| 382 |
+
"policy/approxkl_avg": 5.9386420249938965,
|
| 383 |
+
"policy/clipfrac_avg": 0.18009786307811737,
|
| 384 |
+
"policy/entropy_avg": 0.6803157925605774,
|
| 385 |
+
"step": 18,
|
| 386 |
+
"val/clipfrac_avg": 0.12902730703353882,
|
| 387 |
+
"val/num_eos_tokens": 0,
|
| 388 |
+
"val/ratio": 0.9129149913787842,
|
| 389 |
+
"val/ratio_var": 0.0004151359898969531
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"episode": 304,
|
| 393 |
+
"epoch": 0.5333333333333333,
|
| 394 |
+
"eps": 3,
|
| 395 |
+
"loss/policy_avg": 0.02675745077431202,
|
| 396 |
+
"loss/value_avg": 1.745023488998413,
|
| 397 |
+
"lr": 1.1728037383177571e-05,
|
| 398 |
+
"objective/entropy": 22.75409698486328,
|
| 399 |
+
"objective/kl": 119.34423828125,
|
| 400 |
+
"objective/non_score_reward": -5.967212200164795,
|
| 401 |
+
"objective/rlhf_reward": -5.197680950164795,
|
| 402 |
+
"objective/scores": 0.76953125,
|
| 403 |
+
"policy/approxkl_avg": 6.532215118408203,
|
| 404 |
+
"policy/clipfrac_avg": 0.1759602427482605,
|
| 405 |
+
"policy/entropy_avg": 0.8050931692123413,
|
| 406 |
+
"step": 19,
|
| 407 |
+
"val/clipfrac_avg": 0.0052770450711250305,
|
| 408 |
+
"val/num_eos_tokens": 0,
|
| 409 |
+
"val/ratio": 0.9205665588378906,
|
| 410 |
+
"val/ratio_var": 0.0005024418351240456
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"episode": 320,
|
| 414 |
+
"epoch": 0.5614035087719298,
|
| 415 |
+
"eps": 3,
|
| 416 |
+
"loss/policy_avg": 0.014609305188059807,
|
| 417 |
+
"loss/value_avg": 1.7434797286987305,
|
| 418 |
+
"lr": 1.159626168224299e-05,
|
| 419 |
+
"objective/entropy": 29.933460235595703,
|
| 420 |
+
"objective/kl": 144.45672607421875,
|
| 421 |
+
"objective/non_score_reward": -7.222836494445801,
|
| 422 |
+
"objective/rlhf_reward": -6.675961494445801,
|
| 423 |
+
"objective/scores": 0.546875,
|
| 424 |
+
"policy/approxkl_avg": 7.733211517333984,
|
| 425 |
+
"policy/clipfrac_avg": 0.1762971729040146,
|
| 426 |
+
"policy/entropy_avg": 0.8804416656494141,
|
| 427 |
+
"step": 20,
|
| 428 |
+
"val/clipfrac_avg": 0.026533018797636032,
|
| 429 |
+
"val/num_eos_tokens": 0,
|
| 430 |
+
"val/ratio": 0.8693416714668274,
|
| 431 |
+
"val/ratio_var": 0.000557584862690419
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"episode": 336,
|
| 435 |
+
"epoch": 0.5894736842105263,
|
| 436 |
+
"eps": 3,
|
| 437 |
+
"loss/policy_avg": 0.019340746104717255,
|
| 438 |
+
"loss/value_avg": 1.471176266670227,
|
| 439 |
+
"lr": 1.146448598130841e-05,
|
| 440 |
+
"objective/entropy": 23.77098846435547,
|
| 441 |
+
"objective/kl": 149.46974182128906,
|
| 442 |
+
"objective/non_score_reward": -7.473487377166748,
|
| 443 |
+
"objective/rlhf_reward": -5.668799877166748,
|
| 444 |
+
"objective/scores": 1.8046875,
|
| 445 |
+
"policy/approxkl_avg": 5.010880470275879,
|
| 446 |
+
"policy/clipfrac_avg": 0.18341976404190063,
|
| 447 |
+
"policy/entropy_avg": 0.973499596118927,
|
| 448 |
+
"step": 21,
|
| 449 |
+
"val/clipfrac_avg": 0.016483133658766747,
|
| 450 |
+
"val/num_eos_tokens": 0,
|
| 451 |
+
"val/ratio": 0.8893705606460571,
|
| 452 |
+
"val/ratio_var": 0.0012118957238271832
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"episode": 352,
|
| 456 |
+
"epoch": 0.6175438596491228,
|
| 457 |
+
"eps": 3,
|
| 458 |
+
"loss/policy_avg": 0.009950436651706696,
|
| 459 |
+
"loss/value_avg": 1.1944406032562256,
|
| 460 |
+
"lr": 1.1332710280373831e-05,
|
| 461 |
+
"objective/entropy": 23.91471290588379,
|
| 462 |
+
"objective/kl": 155.48843383789062,
|
| 463 |
+
"objective/non_score_reward": -7.774421691894531,
|
| 464 |
+
"objective/rlhf_reward": -7.125984191894531,
|
| 465 |
+
"objective/scores": 0.6484375,
|
| 466 |
+
"policy/approxkl_avg": 6.8445940017700195,
|
| 467 |
+
"policy/clipfrac_avg": 0.16214622557163239,
|
| 468 |
+
"policy/entropy_avg": 0.837052047252655,
|
| 469 |
+
"step": 22,
|
| 470 |
+
"val/clipfrac_avg": 0.09257075190544128,
|
| 471 |
+
"val/num_eos_tokens": 0,
|
| 472 |
+
"val/ratio": 0.9027042984962463,
|
| 473 |
+
"val/ratio_var": 0.00019580482330638915
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"episode": 368,
|
| 477 |
+
"epoch": 0.6456140350877193,
|
| 478 |
+
"eps": 3,
|
| 479 |
+
"loss/policy_avg": 0.021877210587263107,
|
| 480 |
+
"loss/value_avg": 1.0937385559082031,
|
| 481 |
+
"lr": 1.1200934579439252e-05,
|
| 482 |
+
"objective/entropy": 23.552492141723633,
|
| 483 |
+
"objective/kl": 172.6247100830078,
|
| 484 |
+
"objective/non_score_reward": -8.631235122680664,
|
| 485 |
+
"objective/rlhf_reward": -6.787485122680664,
|
| 486 |
+
"objective/scores": 1.84375,
|
| 487 |
+
"policy/approxkl_avg": 6.227967262268066,
|
| 488 |
+
"policy/clipfrac_avg": 0.16203010082244873,
|
| 489 |
+
"policy/entropy_avg": 0.8671172857284546,
|
| 490 |
+
"step": 23,
|
| 491 |
+
"val/clipfrac_avg": 0.018410665914416313,
|
| 492 |
+
"val/num_eos_tokens": 0,
|
| 493 |
+
"val/ratio": 0.8919187784194946,
|
| 494 |
+
"val/ratio_var": 0.001309347921051085
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"episode": 384,
|
| 498 |
+
"epoch": 0.6736842105263158,
|
| 499 |
+
"eps": 3,
|
| 500 |
+
"loss/policy_avg": 0.036975178867578506,
|
| 501 |
+
"loss/value_avg": 1.1198090314865112,
|
| 502 |
+
"lr": 1.1069158878504672e-05,
|
| 503 |
+
"objective/entropy": 20.281631469726562,
|
| 504 |
+
"objective/kl": 170.099365234375,
|
| 505 |
+
"objective/non_score_reward": -8.504968643188477,
|
| 506 |
+
"objective/rlhf_reward": -7.262781143188477,
|
| 507 |
+
"objective/scores": 1.2421875,
|
| 508 |
+
"policy/approxkl_avg": 6.825319766998291,
|
| 509 |
+
"policy/clipfrac_avg": 0.14049983024597168,
|
| 510 |
+
"policy/entropy_avg": 0.802283525466919,
|
| 511 |
+
"step": 24,
|
| 512 |
+
"val/clipfrac_avg": 0.0785929411649704,
|
| 513 |
+
"val/num_eos_tokens": 0,
|
| 514 |
+
"val/ratio": 0.8865500092506409,
|
| 515 |
+
"val/ratio_var": 0.0006475243135355413
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"episode": 400,
|
| 519 |
+
"epoch": 0.7017543859649122,
|
| 520 |
+
"eps": 3,
|
| 521 |
+
"loss/policy_avg": 0.0237587820738554,
|
| 522 |
+
"loss/value_avg": 1.3415908813476562,
|
| 523 |
+
"lr": 1.0937383177570093e-05,
|
| 524 |
+
"objective/entropy": 21.36954689025879,
|
| 525 |
+
"objective/kl": 179.03819274902344,
|
| 526 |
+
"objective/non_score_reward": -8.951910018920898,
|
| 527 |
+
"objective/rlhf_reward": -7.334722518920898,
|
| 528 |
+
"objective/scores": 1.6171875,
|
| 529 |
+
"policy/approxkl_avg": 7.667660713195801,
|
| 530 |
+
"policy/clipfrac_avg": 0.11468379199504852,
|
| 531 |
+
"policy/entropy_avg": 0.7623839974403381,
|
| 532 |
+
"step": 25,
|
| 533 |
+
"val/clipfrac_avg": 0.27072370052337646,
|
| 534 |
+
"val/num_eos_tokens": 0,
|
| 535 |
+
"val/ratio": 0.8912988901138306,
|
| 536 |
+
"val/ratio_var": 2.9881122827646323e-05
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"episode": 416,
|
| 540 |
+
"epoch": 0.7298245614035088,
|
| 541 |
+
"eps": 3,
|
| 542 |
+
"loss/policy_avg": 0.018890127539634705,
|
| 543 |
+
"loss/value_avg": 1.2341463565826416,
|
| 544 |
+
"lr": 1.0805607476635514e-05,
|
| 545 |
+
"objective/entropy": 16.852466583251953,
|
| 546 |
+
"objective/kl": 179.8382568359375,
|
| 547 |
+
"objective/non_score_reward": -8.991912841796875,
|
| 548 |
+
"objective/rlhf_reward": -8.312225341796875,
|
| 549 |
+
"objective/scores": 0.6796875,
|
| 550 |
+
"policy/approxkl_avg": 7.357123374938965,
|
| 551 |
+
"policy/clipfrac_avg": 0.11944779008626938,
|
| 552 |
+
"policy/entropy_avg": 0.7555092573165894,
|
| 553 |
+
"step": 26,
|
| 554 |
+
"val/clipfrac_avg": 0.2508935034275055,
|
| 555 |
+
"val/num_eos_tokens": 0,
|
| 556 |
+
"val/ratio": 0.8802620768547058,
|
| 557 |
+
"val/ratio_var": 0.00018670025747269392
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"episode": 432,
|
| 561 |
+
"epoch": 0.7578947368421053,
|
| 562 |
+
"eps": 3,
|
| 563 |
+
"loss/policy_avg": 0.021100062876939774,
|
| 564 |
+
"loss/value_avg": 0.9362931847572327,
|
| 565 |
+
"lr": 1.0673831775700934e-05,
|
| 566 |
+
"objective/entropy": 21.68787384033203,
|
| 567 |
+
"objective/kl": 183.68655395507812,
|
| 568 |
+
"objective/non_score_reward": -9.184328079223633,
|
| 569 |
+
"objective/rlhf_reward": -7.199953079223633,
|
| 570 |
+
"objective/scores": 1.984375,
|
| 571 |
+
"policy/approxkl_avg": 5.103993892669678,
|
| 572 |
+
"policy/clipfrac_avg": 0.12146226316690445,
|
| 573 |
+
"policy/entropy_avg": 0.7987968921661377,
|
| 574 |
+
"step": 27,
|
| 575 |
+
"val/clipfrac_avg": 0.05424528568983078,
|
| 576 |
+
"val/num_eos_tokens": 0,
|
| 577 |
+
"val/ratio": 0.8928566575050354,
|
| 578 |
+
"val/ratio_var": 0.00013602118997368962
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"episode": 448,
|
| 582 |
+
"epoch": 0.7859649122807018,
|
| 583 |
+
"eps": 3,
|
| 584 |
+
"loss/policy_avg": 0.011967534199357033,
|
| 585 |
+
"loss/value_avg": 1.0544021129608154,
|
| 586 |
+
"lr": 1.0542056074766355e-05,
|
| 587 |
+
"objective/entropy": 23.05614471435547,
|
| 588 |
+
"objective/kl": 182.8275146484375,
|
| 589 |
+
"objective/non_score_reward": -9.141375541687012,
|
| 590 |
+
"objective/rlhf_reward": -7.789813041687012,
|
| 591 |
+
"objective/scores": 1.3515625,
|
| 592 |
+
"policy/approxkl_avg": 4.963105201721191,
|
| 593 |
+
"policy/clipfrac_avg": 0.14268869161605835,
|
| 594 |
+
"policy/entropy_avg": 0.8151739835739136,
|
| 595 |
+
"step": 28,
|
| 596 |
+
"val/clipfrac_avg": 0.2146226465702057,
|
| 597 |
+
"val/num_eos_tokens": 0,
|
| 598 |
+
"val/ratio": 0.8715541362762451,
|
| 599 |
+
"val/ratio_var": 7.266430475283414e-05
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"episode": 464,
|
| 603 |
+
"epoch": 0.8140350877192982,
|
| 604 |
+
"eps": 3,
|
| 605 |
+
"loss/policy_avg": 0.011134720407426357,
|
| 606 |
+
"loss/value_avg": 0.775411069393158,
|
| 607 |
+
"lr": 1.0410280373831775e-05,
|
| 608 |
+
"objective/entropy": 25.063724517822266,
|
| 609 |
+
"objective/kl": 188.51422119140625,
|
| 610 |
+
"objective/non_score_reward": -9.425710678100586,
|
| 611 |
+
"objective/rlhf_reward": -8.394460678100586,
|
| 612 |
+
"objective/scores": 1.03125,
|
| 613 |
+
"policy/approxkl_avg": 7.372692108154297,
|
| 614 |
+
"policy/clipfrac_avg": 0.1320754736661911,
|
| 615 |
+
"policy/entropy_avg": 0.8475193977355957,
|
| 616 |
+
"step": 29,
|
| 617 |
+
"val/clipfrac_avg": 0.04245283082127571,
|
| 618 |
+
"val/num_eos_tokens": 0,
|
| 619 |
+
"val/ratio": 0.891329824924469,
|
| 620 |
+
"val/ratio_var": 3.0236831207730575e-06
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"episode": 480,
|
| 624 |
+
"epoch": 0.8421052631578947,
|
| 625 |
+
"eps": 3,
|
| 626 |
+
"loss/policy_avg": 0.017270730808377266,
|
| 627 |
+
"loss/value_avg": 0.7942019701004028,
|
| 628 |
+
"lr": 1.0278504672897196e-05,
|
| 629 |
+
"objective/entropy": 22.402624130249023,
|
| 630 |
+
"objective/kl": 185.72421264648438,
|
| 631 |
+
"objective/non_score_reward": -9.286211013793945,
|
| 632 |
+
"objective/rlhf_reward": -7.606523513793945,
|
| 633 |
+
"objective/scores": 1.6796875,
|
| 634 |
+
"policy/approxkl_avg": 8.755260467529297,
|
| 635 |
+
"policy/clipfrac_avg": 0.11261792480945587,
|
| 636 |
+
"policy/entropy_avg": 0.799101710319519,
|
| 637 |
+
"step": 30,
|
| 638 |
+
"val/clipfrac_avg": 0.014740565791726112,
|
| 639 |
+
"val/num_eos_tokens": 0,
|
| 640 |
+
"val/ratio": 0.8892796635627747,
|
| 641 |
+
"val/ratio_var": 6.262218084884807e-05
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"episode": 496,
|
| 645 |
+
"epoch": 0.8701754385964913,
|
| 646 |
+
"eps": 3,
|
| 647 |
+
"loss/policy_avg": 0.019226763397455215,
|
| 648 |
+
"loss/value_avg": 0.6826229095458984,
|
| 649 |
+
"lr": 1.0146728971962616e-05,
|
| 650 |
+
"objective/entropy": 23.43070411682129,
|
| 651 |
+
"objective/kl": 201.67799377441406,
|
| 652 |
+
"objective/non_score_reward": -10.083900451660156,
|
| 653 |
+
"objective/rlhf_reward": -9.154212951660156,
|
| 654 |
+
"objective/scores": 0.9296875,
|
| 655 |
+
"policy/approxkl_avg": 8.144369125366211,
|
| 656 |
+
"policy/clipfrac_avg": 0.12264151126146317,
|
| 657 |
+
"policy/entropy_avg": 0.8400179147720337,
|
| 658 |
+
"step": 31,
|
| 659 |
+
"val/clipfrac_avg": 0.2228773534297943,
|
| 660 |
+
"val/num_eos_tokens": 0,
|
| 661 |
+
"val/ratio": 0.8812745213508606,
|
| 662 |
+
"val/ratio_var": 0.00010276544344378635
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"episode": 512,
|
| 666 |
+
"epoch": 0.8982456140350877,
|
| 667 |
+
"eps": 3,
|
| 668 |
+
"loss/policy_avg": 0.005793810822069645,
|
| 669 |
+
"loss/value_avg": 0.8381754159927368,
|
| 670 |
+
"lr": 1.0014953271028037e-05,
|
| 671 |
+
"objective/entropy": 21.22824478149414,
|
| 672 |
+
"objective/kl": 185.61614990234375,
|
| 673 |
+
"objective/non_score_reward": -9.280807495117188,
|
| 674 |
+
"objective/rlhf_reward": -7.3511199951171875,
|
| 675 |
+
"objective/scores": 1.9296875,
|
| 676 |
+
"policy/approxkl_avg": 7.533528804779053,
|
| 677 |
+
"policy/clipfrac_avg": 0.12205187976360321,
|
| 678 |
+
"policy/entropy_avg": 0.7801576256752014,
|
| 679 |
+
"step": 32,
|
| 680 |
+
"val/clipfrac_avg": 0.09787735342979431,
|
| 681 |
+
"val/num_eos_tokens": 0,
|
| 682 |
+
"val/ratio": 0.8987904787063599,
|
| 683 |
+
"val/ratio_var": 0.00018136559810955077
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"episode": 528,
|
| 687 |
+
"epoch": 0.9263157894736842,
|
| 688 |
+
"eps": 3,
|
| 689 |
+
"loss/policy_avg": 0.02254486456513405,
|
| 690 |
+
"loss/value_avg": 0.8877236843109131,
|
| 691 |
+
"lr": 9.883177570093458e-06,
|
| 692 |
+
"objective/entropy": 23.442203521728516,
|
| 693 |
+
"objective/kl": 188.65457153320312,
|
| 694 |
+
"objective/non_score_reward": -9.432729721069336,
|
| 695 |
+
"objective/rlhf_reward": -7.534292221069336,
|
| 696 |
+
"objective/scores": 1.8984375,
|
| 697 |
+
"policy/approxkl_avg": 5.329720497131348,
|
| 698 |
+
"policy/clipfrac_avg": 0.12323113530874252,
|
| 699 |
+
"policy/entropy_avg": 0.8007351160049438,
|
| 700 |
+
"step": 33,
|
| 701 |
+
"val/clipfrac_avg": 0.036556605249643326,
|
| 702 |
+
"val/num_eos_tokens": 0,
|
| 703 |
+
"val/ratio": 0.8945379257202148,
|
| 704 |
+
"val/ratio_var": 0.0001225806336151436
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"episode": 544,
|
| 708 |
+
"epoch": 0.9543859649122807,
|
| 709 |
+
"eps": 3,
|
| 710 |
+
"loss/policy_avg": 0.01662730611860752,
|
| 711 |
+
"loss/value_avg": 0.637324869632721,
|
| 712 |
+
"lr": 9.751401869158878e-06,
|
| 713 |
+
"objective/entropy": 20.620216369628906,
|
| 714 |
+
"objective/kl": 186.25180053710938,
|
| 715 |
+
"objective/non_score_reward": -9.312589645385742,
|
| 716 |
+
"objective/rlhf_reward": -7.851652145385742,
|
| 717 |
+
"objective/scores": 1.4609375,
|
| 718 |
+
"policy/approxkl_avg": 8.322406768798828,
|
| 719 |
+
"policy/clipfrac_avg": 0.125,
|
| 720 |
+
"policy/entropy_avg": 0.7402328848838806,
|
| 721 |
+
"step": 34,
|
| 722 |
+
"val/clipfrac_avg": 0.002358490601181984,
|
| 723 |
+
"val/num_eos_tokens": 0,
|
| 724 |
+
"val/ratio": 0.8899158239364624,
|
| 725 |
+
"val/ratio_var": 0.0002540757122915238
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"episode": 560,
|
| 729 |
+
"epoch": 0.9824561403508771,
|
| 730 |
+
"eps": 3,
|
| 731 |
+
"loss/policy_avg": 0.0003616265021264553,
|
| 732 |
+
"loss/value_avg": 0.6015822291374207,
|
| 733 |
+
"lr": 9.619626168224299e-06,
|
| 734 |
+
"objective/entropy": 21.12371063232422,
|
| 735 |
+
"objective/kl": 186.5103302001953,
|
| 736 |
+
"objective/non_score_reward": -9.325516700744629,
|
| 737 |
+
"objective/rlhf_reward": -7.817704200744629,
|
| 738 |
+
"objective/scores": 1.5078125,
|
| 739 |
+
"policy/approxkl_avg": 6.682015419006348,
|
| 740 |
+
"policy/clipfrac_avg": 0.13443395495414734,
|
| 741 |
+
"policy/entropy_avg": 0.7727504372596741,
|
| 742 |
+
"step": 35,
|
| 743 |
+
"val/clipfrac_avg": 0.037146229296922684,
|
| 744 |
+
"val/num_eos_tokens": 0,
|
| 745 |
+
"val/ratio": 0.8780511021614075,
|
| 746 |
+
"val/ratio_var": 6.755981303285807e-05
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"episode": 576,
|
| 750 |
+
"epoch": 1.0105263157894737,
|
| 751 |
+
"eps": 3,
|
| 752 |
+
"loss/policy_avg": 0.009338408708572388,
|
| 753 |
+
"loss/value_avg": 0.6681860685348511,
|
| 754 |
+
"lr": 9.48785046728972e-06,
|
| 755 |
+
"objective/entropy": 21.184894561767578,
|
| 756 |
+
"objective/kl": 192.3019256591797,
|
| 757 |
+
"objective/non_score_reward": -9.615096092224121,
|
| 758 |
+
"objective/rlhf_reward": -7.443221092224121,
|
| 759 |
+
"objective/scores": 2.171875,
|
| 760 |
+
"policy/approxkl_avg": 6.053627967834473,
|
| 761 |
+
"policy/clipfrac_avg": 0.1179245263338089,
|
| 762 |
+
"policy/entropy_avg": 0.7489176392555237,
|
| 763 |
+
"step": 36,
|
| 764 |
+
"val/clipfrac_avg": 0.15683962404727936,
|
| 765 |
+
"val/num_eos_tokens": 0,
|
| 766 |
+
"val/ratio": 0.8922820091247559,
|
| 767 |
+
"val/ratio_var": 0.00026253468240611255
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"episode": 592,
|
| 771 |
+
"epoch": 1.03859649122807,
|
| 772 |
+
"eps": 3,
|
| 773 |
+
"loss/policy_avg": 0.006294197402894497,
|
| 774 |
+
"loss/value_avg": 0.8030184507369995,
|
| 775 |
+
"lr": 9.35607476635514e-06,
|
| 776 |
+
"objective/entropy": 20.368942260742188,
|
| 777 |
+
"objective/kl": 190.313232421875,
|
| 778 |
+
"objective/non_score_reward": -9.51566219329834,
|
| 779 |
+
"objective/rlhf_reward": -8.02347469329834,
|
| 780 |
+
"objective/scores": 1.4921875,
|
| 781 |
+
"policy/approxkl_avg": 6.838142395019531,
|
| 782 |
+
"policy/clipfrac_avg": 0.11615566164255142,
|
| 783 |
+
"policy/entropy_avg": 0.7397478222846985,
|
| 784 |
+
"step": 37,
|
| 785 |
+
"val/clipfrac_avg": 0.003537735901772976,
|
| 786 |
+
"val/num_eos_tokens": 0,
|
| 787 |
+
"val/ratio": 0.8849948644638062,
|
| 788 |
+
"val/ratio_var": 8.990589412860572e-05
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"episode": 608,
|
| 792 |
+
"epoch": 1.0666666666666667,
|
| 793 |
+
"eps": 3,
|
| 794 |
+
"loss/policy_avg": 0.016940509900450706,
|
| 795 |
+
"loss/value_avg": 0.5882998704910278,
|
| 796 |
+
"lr": 9.22429906542056e-06,
|
| 797 |
+
"objective/entropy": 23.183269500732422,
|
| 798 |
+
"objective/kl": 186.7628631591797,
|
| 799 |
+
"objective/non_score_reward": -9.338143348693848,
|
| 800 |
+
"objective/rlhf_reward": -7.564705848693848,
|
| 801 |
+
"objective/scores": 1.7734375,
|
| 802 |
+
"policy/approxkl_avg": 3.9737157821655273,
|
| 803 |
+
"policy/clipfrac_avg": 0.13089622557163239,
|
| 804 |
+
"policy/entropy_avg": 0.7805944681167603,
|
| 805 |
+
"step": 38,
|
| 806 |
+
"val/clipfrac_avg": 0.004127358552068472,
|
| 807 |
+
"val/num_eos_tokens": 0,
|
| 808 |
+
"val/ratio": 0.8967232704162598,
|
| 809 |
+
"val/ratio_var": 0.00010195688810199499
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"episode": 624,
|
| 813 |
+
"epoch": 1.0947368421052632,
|
| 814 |
+
"eps": 3,
|
| 815 |
+
"loss/policy_avg": 0.011485239490866661,
|
| 816 |
+
"loss/value_avg": 0.5787074565887451,
|
| 817 |
+
"lr": 9.092523364485981e-06,
|
| 818 |
+
"objective/entropy": 17.85469627380371,
|
| 819 |
+
"objective/kl": 187.50631713867188,
|
| 820 |
+
"objective/non_score_reward": -9.375316619873047,
|
| 821 |
+
"objective/rlhf_reward": -7.203441619873047,
|
| 822 |
+
"objective/scores": 2.171875,
|
| 823 |
+
"policy/approxkl_avg": 5.274375915527344,
|
| 824 |
+
"policy/clipfrac_avg": 0.11556603759527206,
|
| 825 |
+
"policy/entropy_avg": 0.7150323390960693,
|
| 826 |
+
"step": 39,
|
| 827 |
+
"val/clipfrac_avg": 0.16509434580802917,
|
| 828 |
+
"val/num_eos_tokens": 0,
|
| 829 |
+
"val/ratio": 0.8858749866485596,
|
| 830 |
+
"val/ratio_var": 0.0001809587120078504
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"episode": 640,
|
| 834 |
+
"epoch": 1.1228070175438596,
|
| 835 |
+
"eps": 3,
|
| 836 |
+
"loss/policy_avg": 0.017689252272248268,
|
| 837 |
+
"loss/value_avg": 0.7021454572677612,
|
| 838 |
+
"lr": 8.960747663551402e-06,
|
| 839 |
+
"objective/entropy": 21.994365692138672,
|
| 840 |
+
"objective/kl": 182.05810546875,
|
| 841 |
+
"objective/non_score_reward": -9.1029052734375,
|
| 842 |
+
"objective/rlhf_reward": -6.8841552734375,
|
| 843 |
+
"objective/scores": 2.21875,
|
| 844 |
+
"policy/approxkl_avg": 4.578630447387695,
|
| 845 |
+
"policy/clipfrac_avg": 0.12558962404727936,
|
| 846 |
+
"policy/entropy_avg": 0.7208189964294434,
|
| 847 |
+
"step": 40,
|
| 848 |
+
"val/clipfrac_avg": 0.06426886469125748,
|
| 849 |
+
"val/num_eos_tokens": 0,
|
| 850 |
+
"val/ratio": 0.8957177400588989,
|
| 851 |
+
"val/ratio_var": 7.126481068553403e-05
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"episode": 656,
|
| 855 |
+
"epoch": 1.1508771929824562,
|
| 856 |
+
"eps": 3,
|
| 857 |
+
"loss/policy_avg": 0.005923585034906864,
|
| 858 |
+
"loss/value_avg": 0.5532969236373901,
|
| 859 |
+
"lr": 8.828971962616822e-06,
|
| 860 |
+
"objective/entropy": 18.056703567504883,
|
| 861 |
+
"objective/kl": 170.11863708496094,
|
| 862 |
+
"objective/non_score_reward": -8.505931854248047,
|
| 863 |
+
"objective/rlhf_reward": -6.380931854248047,
|
| 864 |
+
"objective/scores": 2.125,
|
| 865 |
+
"policy/approxkl_avg": 3.3685710430145264,
|
| 866 |
+
"policy/clipfrac_avg": 0.12028301507234573,
|
| 867 |
+
"policy/entropy_avg": 0.6859033107757568,
|
| 868 |
+
"step": 41,
|
| 869 |
+
"val/clipfrac_avg": 0.0,
|
| 870 |
+
"val/num_eos_tokens": 0,
|
| 871 |
+
"val/ratio": 0.8874868154525757,
|
| 872 |
+
"val/ratio_var": 4.332972093834542e-05
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"episode": 672,
|
| 876 |
+
"epoch": 1.1789473684210525,
|
| 877 |
+
"eps": 3,
|
| 878 |
+
"loss/policy_avg": 0.012987145222723484,
|
| 879 |
+
"loss/value_avg": 0.632028341293335,
|
| 880 |
+
"lr": 8.697196261682243e-06,
|
| 881 |
+
"objective/entropy": 19.865787506103516,
|
| 882 |
+
"objective/kl": 177.42971801757812,
|
| 883 |
+
"objective/non_score_reward": -8.871485710144043,
|
| 884 |
+
"objective/rlhf_reward": -6.527735710144043,
|
| 885 |
+
"objective/scores": 2.34375,
|
| 886 |
+
"policy/approxkl_avg": 6.186136245727539,
|
| 887 |
+
"policy/clipfrac_avg": 0.13325470685958862,
|
| 888 |
+
"policy/entropy_avg": 0.6923173069953918,
|
| 889 |
+
"step": 42,
|
| 890 |
+
"val/clipfrac_avg": 0.125,
|
| 891 |
+
"val/num_eos_tokens": 0,
|
| 892 |
+
"val/ratio": 0.8991193771362305,
|
| 893 |
+
"val/ratio_var": 0.0009607934043742716
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"episode": 688,
|
| 897 |
+
"epoch": 1.207017543859649,
|
| 898 |
+
"eps": 3,
|
| 899 |
+
"loss/policy_avg": 0.011711956933140755,
|
| 900 |
+
"loss/value_avg": 0.4474850296974182,
|
| 901 |
+
"lr": 8.565420560747664e-06,
|
| 902 |
+
"objective/entropy": 19.79934310913086,
|
| 903 |
+
"objective/kl": 176.83395385742188,
|
| 904 |
+
"objective/non_score_reward": -8.84169864654541,
|
| 905 |
+
"objective/rlhf_reward": -7.02138614654541,
|
| 906 |
+
"objective/scores": 1.8203125,
|
| 907 |
+
"policy/approxkl_avg": 5.904331207275391,
|
| 908 |
+
"policy/clipfrac_avg": 0.13089622557163239,
|
| 909 |
+
"policy/entropy_avg": 0.6940422654151917,
|
| 910 |
+
"step": 43,
|
| 911 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 912 |
+
"val/num_eos_tokens": 0,
|
| 913 |
+
"val/ratio": 0.8869062066078186,
|
| 914 |
+
"val/ratio_var": 0.0008940807892940938
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"episode": 704,
|
| 918 |
+
"epoch": 1.2350877192982457,
|
| 919 |
+
"eps": 3,
|
| 920 |
+
"loss/policy_avg": 0.014619714580476284,
|
| 921 |
+
"loss/value_avg": 0.625725507736206,
|
| 922 |
+
"lr": 8.433644859813084e-06,
|
| 923 |
+
"objective/entropy": 22.385128021240234,
|
| 924 |
+
"objective/kl": 181.2716064453125,
|
| 925 |
+
"objective/non_score_reward": -9.063579559326172,
|
| 926 |
+
"objective/rlhf_reward": -7.493267059326172,
|
| 927 |
+
"objective/scores": 1.5703125,
|
| 928 |
+
"policy/approxkl_avg": 4.966976165771484,
|
| 929 |
+
"policy/clipfrac_avg": 0.14799527823925018,
|
| 930 |
+
"policy/entropy_avg": 0.7825338840484619,
|
| 931 |
+
"step": 44,
|
| 932 |
+
"val/clipfrac_avg": 0.14681604504585266,
|
| 933 |
+
"val/num_eos_tokens": 0,
|
| 934 |
+
"val/ratio": 0.883050799369812,
|
| 935 |
+
"val/ratio_var": 1.7675925846560858e-05
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"episode": 720,
|
| 939 |
+
"epoch": 1.263157894736842,
|
| 940 |
+
"eps": 3,
|
| 941 |
+
"loss/policy_avg": 0.024550937116146088,
|
| 942 |
+
"loss/value_avg": 0.7726404070854187,
|
| 943 |
+
"lr": 8.301869158878505e-06,
|
| 944 |
+
"objective/entropy": 19.87116050720215,
|
| 945 |
+
"objective/kl": 172.38674926757812,
|
| 946 |
+
"objective/non_score_reward": -8.619338035583496,
|
| 947 |
+
"objective/rlhf_reward": -6.572463035583496,
|
| 948 |
+
"objective/scores": 2.046875,
|
| 949 |
+
"policy/approxkl_avg": 4.95554256439209,
|
| 950 |
+
"policy/clipfrac_avg": 0.13089622557163239,
|
| 951 |
+
"policy/entropy_avg": 0.6934086680412292,
|
| 952 |
+
"step": 45,
|
| 953 |
+
"val/clipfrac_avg": 0.009433962404727936,
|
| 954 |
+
"val/num_eos_tokens": 0,
|
| 955 |
+
"val/ratio": 0.8879209756851196,
|
| 956 |
+
"val/ratio_var": 0.0003393842780496925
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"episode": 736,
|
| 960 |
+
"epoch": 1.2912280701754386,
|
| 961 |
+
"eps": 3,
|
| 962 |
+
"loss/policy_avg": 0.019363895058631897,
|
| 963 |
+
"loss/value_avg": 0.5529497861862183,
|
| 964 |
+
"lr": 8.170093457943925e-06,
|
| 965 |
+
"objective/entropy": 17.47795295715332,
|
| 966 |
+
"objective/kl": 174.22576904296875,
|
| 967 |
+
"objective/non_score_reward": -8.711288452148438,
|
| 968 |
+
"objective/rlhf_reward": -6.3206634521484375,
|
| 969 |
+
"objective/scores": 2.390625,
|
| 970 |
+
"policy/approxkl_avg": 3.8177051544189453,
|
| 971 |
+
"policy/clipfrac_avg": 0.11851415038108826,
|
| 972 |
+
"policy/entropy_avg": 0.6579099297523499,
|
| 973 |
+
"step": 46,
|
| 974 |
+
"val/clipfrac_avg": 0.001179245300590992,
|
| 975 |
+
"val/num_eos_tokens": 0,
|
| 976 |
+
"val/ratio": 0.9020024538040161,
|
| 977 |
+
"val/ratio_var": 0.00011536524834809825
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"episode": 752,
|
| 981 |
+
"epoch": 1.3192982456140352,
|
| 982 |
+
"eps": 3,
|
| 983 |
+
"loss/policy_avg": 0.012542858719825745,
|
| 984 |
+
"loss/value_avg": 0.6548900604248047,
|
| 985 |
+
"lr": 8.038317757009346e-06,
|
| 986 |
+
"objective/entropy": 17.823394775390625,
|
| 987 |
+
"objective/kl": 171.56602478027344,
|
| 988 |
+
"objective/non_score_reward": -8.578301429748535,
|
| 989 |
+
"objective/rlhf_reward": -6.828301429748535,
|
| 990 |
+
"objective/scores": 1.75,
|
| 991 |
+
"policy/approxkl_avg": 5.147519111633301,
|
| 992 |
+
"policy/clipfrac_avg": 0.14563679695129395,
|
| 993 |
+
"policy/entropy_avg": 0.6417368650436401,
|
| 994 |
+
"step": 47,
|
| 995 |
+
"val/clipfrac_avg": 0.004127358552068472,
|
| 996 |
+
"val/num_eos_tokens": 0,
|
| 997 |
+
"val/ratio": 0.8848338723182678,
|
| 998 |
+
"val/ratio_var": 0.000510354817379266
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"episode": 768,
|
| 1002 |
+
"epoch": 1.3473684210526315,
|
| 1003 |
+
"eps": 3,
|
| 1004 |
+
"loss/policy_avg": 0.013358336873352528,
|
| 1005 |
+
"loss/value_avg": 0.3859623968601227,
|
| 1006 |
+
"lr": 7.906542056074766e-06,
|
| 1007 |
+
"objective/entropy": 16.128374099731445,
|
| 1008 |
+
"objective/kl": 168.9840087890625,
|
| 1009 |
+
"objective/non_score_reward": -8.449200630187988,
|
| 1010 |
+
"objective/rlhf_reward": -6.847638130187988,
|
| 1011 |
+
"objective/scores": 1.6015625,
|
| 1012 |
+
"policy/approxkl_avg": 4.781813144683838,
|
| 1013 |
+
"policy/clipfrac_avg": 0.12028302252292633,
|
| 1014 |
+
"policy/entropy_avg": 0.5887731313705444,
|
| 1015 |
+
"step": 48,
|
| 1016 |
+
"val/clipfrac_avg": 0.01179245300590992,
|
| 1017 |
+
"val/num_eos_tokens": 0,
|
| 1018 |
+
"val/ratio": 0.9012677669525146,
|
| 1019 |
+
"val/ratio_var": 0.00041765146306715906
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"episode": 784,
|
| 1023 |
+
"epoch": 1.3754385964912281,
|
| 1024 |
+
"eps": 3,
|
| 1025 |
+
"loss/policy_avg": 0.017281489446759224,
|
| 1026 |
+
"loss/value_avg": 0.5614770650863647,
|
| 1027 |
+
"lr": 7.774766355140187e-06,
|
| 1028 |
+
"objective/entropy": 20.177621841430664,
|
| 1029 |
+
"objective/kl": 166.16497802734375,
|
| 1030 |
+
"objective/non_score_reward": -8.308249473571777,
|
| 1031 |
+
"objective/rlhf_reward": -7.081686973571777,
|
| 1032 |
+
"objective/scores": 1.2265625,
|
| 1033 |
+
"policy/approxkl_avg": 3.1813056468963623,
|
| 1034 |
+
"policy/clipfrac_avg": 0.1432783007621765,
|
| 1035 |
+
"policy/entropy_avg": 0.7520714998245239,
|
| 1036 |
+
"step": 49,
|
| 1037 |
+
"val/clipfrac_avg": 0.1845518946647644,
|
| 1038 |
+
"val/num_eos_tokens": 0,
|
| 1039 |
+
"val/ratio": 0.8848077654838562,
|
| 1040 |
+
"val/ratio_var": 0.0005292592104524374
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"episode": 800,
|
| 1044 |
+
"epoch": 1.4035087719298245,
|
| 1045 |
+
"eps": 3,
|
| 1046 |
+
"loss/policy_avg": 0.015007663518190384,
|
| 1047 |
+
"loss/value_avg": 0.918763279914856,
|
| 1048 |
+
"lr": 7.642990654205608e-06,
|
| 1049 |
+
"objective/entropy": 17.82724380493164,
|
| 1050 |
+
"objective/kl": 179.90628051757812,
|
| 1051 |
+
"objective/non_score_reward": -8.99531364440918,
|
| 1052 |
+
"objective/rlhf_reward": -7.37812614440918,
|
| 1053 |
+
"objective/scores": 1.6171875,
|
| 1054 |
+
"policy/approxkl_avg": 5.0764312744140625,
|
| 1055 |
+
"policy/clipfrac_avg": 0.1149764209985733,
|
| 1056 |
+
"policy/entropy_avg": 0.6353041529655457,
|
| 1057 |
+
"step": 50,
|
| 1058 |
+
"val/clipfrac_avg": 0.014740565791726112,
|
| 1059 |
+
"val/num_eos_tokens": 0,
|
| 1060 |
+
"val/ratio": 0.9066751599311829,
|
| 1061 |
+
"val/ratio_var": 0.00023197307018563151
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"episode": 816,
|
| 1065 |
+
"epoch": 1.431578947368421,
|
| 1066 |
+
"eps": 3,
|
| 1067 |
+
"loss/policy_avg": 0.013827439397573471,
|
| 1068 |
+
"loss/value_avg": 0.6559504270553589,
|
| 1069 |
+
"lr": 7.511214953271027e-06,
|
| 1070 |
+
"objective/entropy": 17.98678207397461,
|
| 1071 |
+
"objective/kl": 173.85345458984375,
|
| 1072 |
+
"objective/non_score_reward": -8.692672729492188,
|
| 1073 |
+
"objective/rlhf_reward": -7.0911102294921875,
|
| 1074 |
+
"objective/scores": 1.6015625,
|
| 1075 |
+
"policy/approxkl_avg": 4.863851547241211,
|
| 1076 |
+
"policy/clipfrac_avg": 0.12382075190544128,
|
| 1077 |
+
"policy/entropy_avg": 0.6412214040756226,
|
| 1078 |
+
"step": 51,
|
| 1079 |
+
"val/clipfrac_avg": 0.03419811278581619,
|
| 1080 |
+
"val/num_eos_tokens": 0,
|
| 1081 |
+
"val/ratio": 0.8822081089019775,
|
| 1082 |
+
"val/ratio_var": 5.379146841733018e-06
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"episode": 832,
|
| 1086 |
+
"epoch": 1.4596491228070176,
|
| 1087 |
+
"eps": 3,
|
| 1088 |
+
"loss/policy_avg": 0.009551126509904861,
|
| 1089 |
+
"loss/value_avg": 0.46615320444107056,
|
| 1090 |
+
"lr": 7.379439252336448e-06,
|
| 1091 |
+
"objective/entropy": 14.611004829406738,
|
| 1092 |
+
"objective/kl": 169.49464416503906,
|
| 1093 |
+
"objective/non_score_reward": -8.4747314453125,
|
| 1094 |
+
"objective/rlhf_reward": -7.0137939453125,
|
| 1095 |
+
"objective/scores": 1.4609375,
|
| 1096 |
+
"policy/approxkl_avg": 4.765644073486328,
|
| 1097 |
+
"policy/clipfrac_avg": 0.09964622557163239,
|
| 1098 |
+
"policy/entropy_avg": 0.5645979642868042,
|
| 1099 |
+
"step": 52,
|
| 1100 |
+
"val/clipfrac_avg": 0.0,
|
| 1101 |
+
"val/num_eos_tokens": 0,
|
| 1102 |
+
"val/ratio": 0.9012112021446228,
|
| 1103 |
+
"val/ratio_var": 1.3516472790797707e-05
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"episode": 848,
|
| 1107 |
+
"epoch": 1.487719298245614,
|
| 1108 |
+
"eps": 3,
|
| 1109 |
+
"loss/policy_avg": 0.009903261438012123,
|
| 1110 |
+
"loss/value_avg": 1.024169683456421,
|
| 1111 |
+
"lr": 7.2476635514018685e-06,
|
| 1112 |
+
"objective/entropy": 16.012420654296875,
|
| 1113 |
+
"objective/kl": 173.99917602539062,
|
| 1114 |
+
"objective/non_score_reward": -8.699958801269531,
|
| 1115 |
+
"objective/rlhf_reward": -6.371833801269531,
|
| 1116 |
+
"objective/scores": 2.328125,
|
| 1117 |
+
"policy/approxkl_avg": 4.889924049377441,
|
| 1118 |
+
"policy/clipfrac_avg": 0.11261792480945587,
|
| 1119 |
+
"policy/entropy_avg": 0.5950597524642944,
|
| 1120 |
+
"step": 53,
|
| 1121 |
+
"val/clipfrac_avg": 0.20518869161605835,
|
| 1122 |
+
"val/num_eos_tokens": 0,
|
| 1123 |
+
"val/ratio": 0.8948594331741333,
|
| 1124 |
+
"val/ratio_var": 6.705896521452814e-05
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"episode": 864,
|
| 1128 |
+
"epoch": 1.5157894736842106,
|
| 1129 |
+
"eps": 3,
|
| 1130 |
+
"loss/policy_avg": -0.007397271227091551,
|
| 1131 |
+
"loss/value_avg": 0.4296436011791229,
|
| 1132 |
+
"lr": 7.115887850467289e-06,
|
| 1133 |
+
"objective/entropy": 12.315929412841797,
|
| 1134 |
+
"objective/kl": 175.46511840820312,
|
| 1135 |
+
"objective/non_score_reward": -8.773256301879883,
|
| 1136 |
+
"objective/rlhf_reward": -6.570131301879883,
|
| 1137 |
+
"objective/scores": 2.203125,
|
| 1138 |
+
"policy/approxkl_avg": 5.532078742980957,
|
| 1139 |
+
"policy/clipfrac_avg": 0.09375,
|
| 1140 |
+
"policy/entropy_avg": 0.5138251781463623,
|
| 1141 |
+
"step": 54,
|
| 1142 |
+
"val/clipfrac_avg": 0.003537735901772976,
|
| 1143 |
+
"val/num_eos_tokens": 0,
|
| 1144 |
+
"val/ratio": 0.9130541086196899,
|
| 1145 |
+
"val/ratio_var": 8.332962897839025e-05
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"episode": 880,
|
| 1149 |
+
"epoch": 1.543859649122807,
|
| 1150 |
+
"eps": 3,
|
| 1151 |
+
"loss/policy_avg": 0.0055680545046925545,
|
| 1152 |
+
"loss/value_avg": 0.44329357147216797,
|
| 1153 |
+
"lr": 6.9841121495327106e-06,
|
| 1154 |
+
"objective/entropy": 13.753085136413574,
|
| 1155 |
+
"objective/kl": 162.1046905517578,
|
| 1156 |
+
"objective/non_score_reward": -8.10523509979248,
|
| 1157 |
+
"objective/rlhf_reward": -6.1052350997924805,
|
| 1158 |
+
"objective/scores": 2.0,
|
| 1159 |
+
"policy/approxkl_avg": 4.2778730392456055,
|
| 1160 |
+
"policy/clipfrac_avg": 0.10200471431016922,
|
| 1161 |
+
"policy/entropy_avg": 0.538252592086792,
|
| 1162 |
+
"step": 55,
|
| 1163 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1164 |
+
"val/num_eos_tokens": 0,
|
| 1165 |
+
"val/ratio": 0.89923495054245,
|
| 1166 |
+
"val/ratio_var": 0.00013667276652995497
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"episode": 896,
|
| 1170 |
+
"epoch": 1.5719298245614035,
|
| 1171 |
+
"eps": 3,
|
| 1172 |
+
"loss/policy_avg": 0.0029763877391815186,
|
| 1173 |
+
"loss/value_avg": 0.5969923734664917,
|
| 1174 |
+
"lr": 6.852336448598131e-06,
|
| 1175 |
+
"objective/entropy": 10.386423110961914,
|
| 1176 |
+
"objective/kl": 170.64817810058594,
|
| 1177 |
+
"objective/non_score_reward": -8.53240966796875,
|
| 1178 |
+
"objective/rlhf_reward": -5.84490966796875,
|
| 1179 |
+
"objective/scores": 2.6875,
|
| 1180 |
+
"policy/approxkl_avg": 5.515145301818848,
|
| 1181 |
+
"policy/clipfrac_avg": 0.0695754736661911,
|
| 1182 |
+
"policy/entropy_avg": 0.4759911596775055,
|
| 1183 |
+
"step": 56,
|
| 1184 |
+
"val/clipfrac_avg": 0.22051887214183807,
|
| 1185 |
+
"val/num_eos_tokens": 0,
|
| 1186 |
+
"val/ratio": 0.9136029481887817,
|
| 1187 |
+
"val/ratio_var": 0.00015939133299980313
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"episode": 912,
|
| 1191 |
+
"epoch": 1.6,
|
| 1192 |
+
"eps": 3,
|
| 1193 |
+
"loss/policy_avg": -0.0002519981935620308,
|
| 1194 |
+
"loss/value_avg": 0.6188120245933533,
|
| 1195 |
+
"lr": 6.720560747663552e-06,
|
| 1196 |
+
"objective/entropy": 9.047847747802734,
|
| 1197 |
+
"objective/kl": 162.95162963867188,
|
| 1198 |
+
"objective/non_score_reward": -8.147581100463867,
|
| 1199 |
+
"objective/rlhf_reward": -5.835081100463867,
|
| 1200 |
+
"objective/scores": 2.3125,
|
| 1201 |
+
"policy/approxkl_avg": 5.942928314208984,
|
| 1202 |
+
"policy/clipfrac_avg": 0.06721697747707367,
|
| 1203 |
+
"policy/entropy_avg": 0.43892478942871094,
|
| 1204 |
+
"step": 57,
|
| 1205 |
+
"val/clipfrac_avg": 0.03242924436926842,
|
| 1206 |
+
"val/num_eos_tokens": 0,
|
| 1207 |
+
"val/ratio": 0.9175019264221191,
|
| 1208 |
+
"val/ratio_var": 2.5528926926199347e-05
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"episode": 928,
|
| 1212 |
+
"epoch": 1.6280701754385964,
|
| 1213 |
+
"eps": 3,
|
| 1214 |
+
"loss/policy_avg": -0.004241641610860825,
|
| 1215 |
+
"loss/value_avg": 0.6380342245101929,
|
| 1216 |
+
"lr": 6.588785046728972e-06,
|
| 1217 |
+
"objective/entropy": 10.172576904296875,
|
| 1218 |
+
"objective/kl": 172.64210510253906,
|
| 1219 |
+
"objective/non_score_reward": -8.632105827331543,
|
| 1220 |
+
"objective/rlhf_reward": -6.085230827331543,
|
| 1221 |
+
"objective/scores": 2.546875,
|
| 1222 |
+
"policy/approxkl_avg": 5.1512861251831055,
|
| 1223 |
+
"policy/clipfrac_avg": 0.09669811278581619,
|
| 1224 |
+
"policy/entropy_avg": 0.44444799423217773,
|
| 1225 |
+
"step": 58,
|
| 1226 |
+
"val/clipfrac_avg": 0.00294811325147748,
|
| 1227 |
+
"val/num_eos_tokens": 0,
|
| 1228 |
+
"val/ratio": 0.9051207304000854,
|
| 1229 |
+
"val/ratio_var": 9.685073746368289e-05
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"episode": 944,
|
| 1233 |
+
"epoch": 1.656140350877193,
|
| 1234 |
+
"eps": 3,
|
| 1235 |
+
"loss/policy_avg": 0.005844447761774063,
|
| 1236 |
+
"loss/value_avg": 0.46530038118362427,
|
| 1237 |
+
"lr": 6.457009345794393e-06,
|
| 1238 |
+
"objective/entropy": 11.34018611907959,
|
| 1239 |
+
"objective/kl": 167.05087280273438,
|
| 1240 |
+
"objective/non_score_reward": -8.352543830871582,
|
| 1241 |
+
"objective/rlhf_reward": -5.368168830871582,
|
| 1242 |
+
"objective/scores": 2.984375,
|
| 1243 |
+
"policy/approxkl_avg": 4.73173713684082,
|
| 1244 |
+
"policy/clipfrac_avg": 0.06898584961891174,
|
| 1245 |
+
"policy/entropy_avg": 0.4987587630748749,
|
| 1246 |
+
"step": 59,
|
| 1247 |
+
"val/clipfrac_avg": 0.003537735901772976,
|
| 1248 |
+
"val/num_eos_tokens": 0,
|
| 1249 |
+
"val/ratio": 0.9096221327781677,
|
| 1250 |
+
"val/ratio_var": 0.0002903940330725163
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"episode": 960,
|
| 1254 |
+
"epoch": 1.6842105263157894,
|
| 1255 |
+
"eps": 3,
|
| 1256 |
+
"loss/policy_avg": 0.0015796682564541698,
|
| 1257 |
+
"loss/value_avg": 0.5465973615646362,
|
| 1258 |
+
"lr": 6.3252336448598135e-06,
|
| 1259 |
+
"objective/entropy": 10.832345962524414,
|
| 1260 |
+
"objective/kl": 166.35125732421875,
|
| 1261 |
+
"objective/non_score_reward": -8.317562103271484,
|
| 1262 |
+
"objective/rlhf_reward": -5.114437103271484,
|
| 1263 |
+
"objective/scores": 3.203125,
|
| 1264 |
+
"policy/approxkl_avg": 4.080867767333984,
|
| 1265 |
+
"policy/clipfrac_avg": 0.08726415038108826,
|
| 1266 |
+
"policy/entropy_avg": 0.46615108847618103,
|
| 1267 |
+
"step": 60,
|
| 1268 |
+
"val/clipfrac_avg": 0.018278302624821663,
|
| 1269 |
+
"val/num_eos_tokens": 0,
|
| 1270 |
+
"val/ratio": 0.9084208011627197,
|
| 1271 |
+
"val/ratio_var": 1.8292890672455542e-05
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"episode": 976,
|
| 1275 |
+
"epoch": 1.712280701754386,
|
| 1276 |
+
"eps": 3,
|
| 1277 |
+
"loss/policy_avg": -0.0016184533014893532,
|
| 1278 |
+
"loss/value_avg": 0.6316072344779968,
|
| 1279 |
+
"lr": 6.193457943925234e-06,
|
| 1280 |
+
"objective/entropy": 9.0885648727417,
|
| 1281 |
+
"objective/kl": 172.646240234375,
|
| 1282 |
+
"objective/non_score_reward": -8.632311820983887,
|
| 1283 |
+
"objective/rlhf_reward": -5.194811820983887,
|
| 1284 |
+
"objective/scores": 3.4375,
|
| 1285 |
+
"policy/approxkl_avg": 4.502593994140625,
|
| 1286 |
+
"policy/clipfrac_avg": 0.06603773683309555,
|
| 1287 |
+
"policy/entropy_avg": 0.41100969910621643,
|
| 1288 |
+
"step": 61,
|
| 1289 |
+
"val/clipfrac_avg": 0.044811319559812546,
|
| 1290 |
+
"val/num_eos_tokens": 0,
|
| 1291 |
+
"val/ratio": 0.9256702661514282,
|
| 1292 |
+
"val/ratio_var": 7.894221198512241e-05
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"episode": 992,
|
| 1296 |
+
"epoch": 1.7403508771929825,
|
| 1297 |
+
"eps": 3,
|
| 1298 |
+
"loss/policy_avg": -0.0019415542483329773,
|
| 1299 |
+
"loss/value_avg": 0.6046911478042603,
|
| 1300 |
+
"lr": 6.061682242990655e-06,
|
| 1301 |
+
"objective/entropy": 9.12926197052002,
|
| 1302 |
+
"objective/kl": 169.4315185546875,
|
| 1303 |
+
"objective/non_score_reward": -8.471575736999512,
|
| 1304 |
+
"objective/rlhf_reward": -5.424700736999512,
|
| 1305 |
+
"objective/scores": 3.046875,
|
| 1306 |
+
"policy/approxkl_avg": 5.609973907470703,
|
| 1307 |
+
"policy/clipfrac_avg": 0.09198112785816193,
|
| 1308 |
+
"policy/entropy_avg": 0.4236205816268921,
|
| 1309 |
+
"step": 62,
|
| 1310 |
+
"val/clipfrac_avg": 0.001768867950886488,
|
| 1311 |
+
"val/num_eos_tokens": 0,
|
| 1312 |
+
"val/ratio": 0.9198966026306152,
|
| 1313 |
+
"val/ratio_var": 6.228529673535377e-05
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"episode": 1008,
|
| 1317 |
+
"epoch": 1.768421052631579,
|
| 1318 |
+
"eps": 3,
|
| 1319 |
+
"loss/policy_avg": -0.007835395634174347,
|
| 1320 |
+
"loss/value_avg": 0.6853305697441101,
|
| 1321 |
+
"lr": 5.929906542056075e-06,
|
| 1322 |
+
"objective/entropy": 8.566083908081055,
|
| 1323 |
+
"objective/kl": 163.68191528320312,
|
| 1324 |
+
"objective/non_score_reward": -8.18409538269043,
|
| 1325 |
+
"objective/rlhf_reward": -4.09034538269043,
|
| 1326 |
+
"objective/scores": 4.09375,
|
| 1327 |
+
"policy/approxkl_avg": 3.7664973735809326,
|
| 1328 |
+
"policy/clipfrac_avg": 0.07429245114326477,
|
| 1329 |
+
"policy/entropy_avg": 0.41426771879196167,
|
| 1330 |
+
"step": 63,
|
| 1331 |
+
"val/clipfrac_avg": 0.007665094453841448,
|
| 1332 |
+
"val/num_eos_tokens": 0,
|
| 1333 |
+
"val/ratio": 0.9395467042922974,
|
| 1334 |
+
"val/ratio_var": 0.00018259203352499753
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"episode": 1024,
|
| 1338 |
+
"epoch": 1.7964912280701755,
|
| 1339 |
+
"eps": 3,
|
| 1340 |
+
"loss/policy_avg": 0.0056846365332603455,
|
| 1341 |
+
"loss/value_avg": 0.8050791621208191,
|
| 1342 |
+
"lr": 5.798130841121495e-06,
|
| 1343 |
+
"objective/entropy": 7.867904186248779,
|
| 1344 |
+
"objective/kl": 176.44961547851562,
|
| 1345 |
+
"objective/non_score_reward": -8.822481155395508,
|
| 1346 |
+
"objective/rlhf_reward": -4.931856155395508,
|
| 1347 |
+
"objective/scores": 3.890625,
|
| 1348 |
+
"policy/approxkl_avg": 4.615470886230469,
|
| 1349 |
+
"policy/clipfrac_avg": 0.07016509771347046,
|
| 1350 |
+
"policy/entropy_avg": 0.40076911449432373,
|
| 1351 |
+
"step": 64,
|
| 1352 |
+
"val/clipfrac_avg": 0.1179245263338089,
|
| 1353 |
+
"val/num_eos_tokens": 0,
|
| 1354 |
+
"val/ratio": 0.9168256521224976,
|
| 1355 |
+
"val/ratio_var": 1.1071170774812344e-05
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"episode": 1040,
|
| 1359 |
+
"epoch": 1.8245614035087718,
|
| 1360 |
+
"eps": 3,
|
| 1361 |
+
"loss/policy_avg": -0.004829235374927521,
|
| 1362 |
+
"loss/value_avg": 0.7683409452438354,
|
| 1363 |
+
"lr": 5.666355140186916e-06,
|
| 1364 |
+
"objective/entropy": 8.73065185546875,
|
| 1365 |
+
"objective/kl": 165.93441772460938,
|
| 1366 |
+
"objective/non_score_reward": -8.296720504760742,
|
| 1367 |
+
"objective/rlhf_reward": -4.531095504760742,
|
| 1368 |
+
"objective/scores": 3.765625,
|
| 1369 |
+
"policy/approxkl_avg": 4.037623882293701,
|
| 1370 |
+
"policy/clipfrac_avg": 0.0625,
|
| 1371 |
+
"policy/entropy_avg": 0.38483142852783203,
|
| 1372 |
+
"step": 65,
|
| 1373 |
+
"val/clipfrac_avg": 0.0,
|
| 1374 |
+
"val/num_eos_tokens": 0,
|
| 1375 |
+
"val/ratio": 0.9364030361175537,
|
| 1376 |
+
"val/ratio_var": 0.0001283105229958892
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"episode": 1056,
|
| 1380 |
+
"epoch": 1.8526315789473684,
|
| 1381 |
+
"eps": 3,
|
| 1382 |
+
"loss/policy_avg": -0.002082128543406725,
|
| 1383 |
+
"loss/value_avg": 0.8781827688217163,
|
| 1384 |
+
"lr": 5.534579439252336e-06,
|
| 1385 |
+
"objective/entropy": 6.81689977645874,
|
| 1386 |
+
"objective/kl": 173.76760864257812,
|
| 1387 |
+
"objective/non_score_reward": -8.688380241394043,
|
| 1388 |
+
"objective/rlhf_reward": -5.454005241394043,
|
| 1389 |
+
"objective/scores": 3.234375,
|
| 1390 |
+
"policy/approxkl_avg": 5.1825032234191895,
|
| 1391 |
+
"policy/clipfrac_avg": 0.07488207519054413,
|
| 1392 |
+
"policy/entropy_avg": 0.3771995007991791,
|
| 1393 |
+
"step": 66,
|
| 1394 |
+
"val/clipfrac_avg": 0.0,
|
| 1395 |
+
"val/num_eos_tokens": 0,
|
| 1396 |
+
"val/ratio": 0.9298049211502075,
|
| 1397 |
+
"val/ratio_var": 0.00015954735863488168
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"episode": 1072,
|
| 1401 |
+
"epoch": 1.880701754385965,
|
| 1402 |
+
"eps": 3,
|
| 1403 |
+
"loss/policy_avg": 0.005034355446696281,
|
| 1404 |
+
"loss/value_avg": 1.0226874351501465,
|
| 1405 |
+
"lr": 5.402803738317757e-06,
|
| 1406 |
+
"objective/entropy": 5.308557510375977,
|
| 1407 |
+
"objective/kl": 171.6015167236328,
|
| 1408 |
+
"objective/non_score_reward": -8.580076217651367,
|
| 1409 |
+
"objective/rlhf_reward": -4.236326217651367,
|
| 1410 |
+
"objective/scores": 4.34375,
|
| 1411 |
+
"policy/approxkl_avg": 5.336367607116699,
|
| 1412 |
+
"policy/clipfrac_avg": 0.04599056765437126,
|
| 1413 |
+
"policy/entropy_avg": 0.34400177001953125,
|
| 1414 |
+
"step": 67,
|
| 1415 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1416 |
+
"val/num_eos_tokens": 0,
|
| 1417 |
+
"val/ratio": 0.9246993064880371,
|
| 1418 |
+
"val/ratio_var": 9.672338592281449e-07
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"episode": 1088,
|
| 1422 |
+
"epoch": 1.9087719298245613,
|
| 1423 |
+
"eps": 3,
|
| 1424 |
+
"loss/policy_avg": 0.023275576531887054,
|
| 1425 |
+
"loss/value_avg": 0.6750494241714478,
|
| 1426 |
+
"lr": 5.271028037383177e-06,
|
| 1427 |
+
"objective/entropy": 7.23941707611084,
|
| 1428 |
+
"objective/kl": 166.45547485351562,
|
| 1429 |
+
"objective/non_score_reward": -8.322773933410645,
|
| 1430 |
+
"objective/rlhf_reward": -4.6352739334106445,
|
| 1431 |
+
"objective/scores": 3.6875,
|
| 1432 |
+
"policy/approxkl_avg": 3.1369752883911133,
|
| 1433 |
+
"policy/clipfrac_avg": 0.05837263911962509,
|
| 1434 |
+
"policy/entropy_avg": 0.3996211886405945,
|
| 1435 |
+
"step": 68,
|
| 1436 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1437 |
+
"val/num_eos_tokens": 0,
|
| 1438 |
+
"val/ratio": 0.9343756437301636,
|
| 1439 |
+
"val/ratio_var": 0.00011849942529806867
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"episode": 1104,
|
| 1443 |
+
"epoch": 1.936842105263158,
|
| 1444 |
+
"eps": 3,
|
| 1445 |
+
"loss/policy_avg": 0.001583978533744812,
|
| 1446 |
+
"loss/value_avg": 0.7364473342895508,
|
| 1447 |
+
"lr": 5.139252336448598e-06,
|
| 1448 |
+
"objective/entropy": 8.292254447937012,
|
| 1449 |
+
"objective/kl": 174.10446166992188,
|
| 1450 |
+
"objective/non_score_reward": -8.705223083496094,
|
| 1451 |
+
"objective/rlhf_reward": -4.517723083496094,
|
| 1452 |
+
"objective/scores": 4.1875,
|
| 1453 |
+
"policy/approxkl_avg": 5.407079696655273,
|
| 1454 |
+
"policy/clipfrac_avg": 0.06780660152435303,
|
| 1455 |
+
"policy/entropy_avg": 0.3910168409347534,
|
| 1456 |
+
"step": 69,
|
| 1457 |
+
"val/clipfrac_avg": 0.001179245300590992,
|
| 1458 |
+
"val/num_eos_tokens": 0,
|
| 1459 |
+
"val/ratio": 0.9262620210647583,
|
| 1460 |
+
"val/ratio_var": 8.509035978931934e-06
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"episode": 1120,
|
| 1464 |
+
"epoch": 1.9649122807017543,
|
| 1465 |
+
"eps": 3,
|
| 1466 |
+
"loss/policy_avg": 0.014011572115123272,
|
| 1467 |
+
"loss/value_avg": 0.49188750982284546,
|
| 1468 |
+
"lr": 5.0074766355140185e-06,
|
| 1469 |
+
"objective/entropy": 4.73923397064209,
|
| 1470 |
+
"objective/kl": 170.3909912109375,
|
| 1471 |
+
"objective/non_score_reward": -8.519549369812012,
|
| 1472 |
+
"objective/rlhf_reward": -4.535174369812012,
|
| 1473 |
+
"objective/scores": 3.984375,
|
| 1474 |
+
"policy/approxkl_avg": 4.553505897521973,
|
| 1475 |
+
"policy/clipfrac_avg": 0.04658018797636032,
|
| 1476 |
+
"policy/entropy_avg": 0.314146488904953,
|
| 1477 |
+
"step": 70,
|
| 1478 |
+
"val/clipfrac_avg": 0.001768867950886488,
|
| 1479 |
+
"val/num_eos_tokens": 0,
|
| 1480 |
+
"val/ratio": 0.9345220327377319,
|
| 1481 |
+
"val/ratio_var": 0.00011590120993787423
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"episode": 1136,
|
| 1485 |
+
"epoch": 1.9929824561403509,
|
| 1486 |
+
"eps": 3,
|
| 1487 |
+
"loss/policy_avg": 0.014443885535001755,
|
| 1488 |
+
"loss/value_avg": 0.8583539724349976,
|
| 1489 |
+
"lr": 4.875700934579439e-06,
|
| 1490 |
+
"objective/entropy": 6.110556602478027,
|
| 1491 |
+
"objective/kl": 168.1246337890625,
|
| 1492 |
+
"objective/non_score_reward": -8.406231880187988,
|
| 1493 |
+
"objective/rlhf_reward": -4.781231880187988,
|
| 1494 |
+
"objective/scores": 3.625,
|
| 1495 |
+
"policy/approxkl_avg": 3.3112387657165527,
|
| 1496 |
+
"policy/clipfrac_avg": 0.04716981202363968,
|
| 1497 |
+
"policy/entropy_avg": 0.3741912841796875,
|
| 1498 |
+
"step": 71,
|
| 1499 |
+
"val/clipfrac_avg": 0.0,
|
| 1500 |
+
"val/num_eos_tokens": 0,
|
| 1501 |
+
"val/ratio": 0.9311293363571167,
|
| 1502 |
+
"val/ratio_var": 1.7129657862824388e-05
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"episode": 1152,
|
| 1506 |
+
"epoch": 2.0210526315789474,
|
| 1507 |
+
"eps": 3,
|
| 1508 |
+
"loss/policy_avg": 0.0069357771426439285,
|
| 1509 |
+
"loss/value_avg": 0.6024092435836792,
|
| 1510 |
+
"lr": 4.74392523364486e-06,
|
| 1511 |
+
"objective/entropy": 1.9080017805099487,
|
| 1512 |
+
"objective/kl": 175.54367065429688,
|
| 1513 |
+
"objective/non_score_reward": -8.777183532714844,
|
| 1514 |
+
"objective/rlhf_reward": -3.7771835327148438,
|
| 1515 |
+
"objective/scores": 5.0,
|
| 1516 |
+
"policy/approxkl_avg": 6.433887004852295,
|
| 1517 |
+
"policy/clipfrac_avg": 0.03655660152435303,
|
| 1518 |
+
"policy/entropy_avg": 0.2685927748680115,
|
| 1519 |
+
"step": 72,
|
| 1520 |
+
"val/clipfrac_avg": 0.0,
|
| 1521 |
+
"val/num_eos_tokens": 0,
|
| 1522 |
+
"val/ratio": 0.9363144040107727,
|
| 1523 |
+
"val/ratio_var": 2.7767631763708778e-05
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"episode": 1168,
|
| 1527 |
+
"epoch": 2.049122807017544,
|
| 1528 |
+
"eps": 3,
|
| 1529 |
+
"loss/policy_avg": 0.004744451027363539,
|
| 1530 |
+
"loss/value_avg": 0.6521505117416382,
|
| 1531 |
+
"lr": 4.61214953271028e-06,
|
| 1532 |
+
"objective/entropy": 2.584568500518799,
|
| 1533 |
+
"objective/kl": 171.61709594726562,
|
| 1534 |
+
"objective/non_score_reward": -8.580854415893555,
|
| 1535 |
+
"objective/rlhf_reward": -3.1121044158935547,
|
| 1536 |
+
"objective/scores": 5.46875,
|
| 1537 |
+
"policy/approxkl_avg": 4.509120941162109,
|
| 1538 |
+
"policy/clipfrac_avg": 0.03478773683309555,
|
| 1539 |
+
"policy/entropy_avg": 0.2757822573184967,
|
| 1540 |
+
"step": 73,
|
| 1541 |
+
"val/clipfrac_avg": 0.2900943160057068,
|
| 1542 |
+
"val/num_eos_tokens": 0,
|
| 1543 |
+
"val/ratio": 0.9448769092559814,
|
| 1544 |
+
"val/ratio_var": 1.143478584708646e-05
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"episode": 1184,
|
| 1548 |
+
"epoch": 2.07719298245614,
|
| 1549 |
+
"eps": 3,
|
| 1550 |
+
"loss/policy_avg": 0.004101068712770939,
|
| 1551 |
+
"loss/value_avg": 0.44417738914489746,
|
| 1552 |
+
"lr": 4.480373831775701e-06,
|
| 1553 |
+
"objective/entropy": 3.265643835067749,
|
| 1554 |
+
"objective/kl": 179.89352416992188,
|
| 1555 |
+
"objective/non_score_reward": -8.99467658996582,
|
| 1556 |
+
"objective/rlhf_reward": -4.11967658996582,
|
| 1557 |
+
"objective/scores": 4.875,
|
| 1558 |
+
"policy/approxkl_avg": 5.798920154571533,
|
| 1559 |
+
"policy/clipfrac_avg": 0.028891509398818016,
|
| 1560 |
+
"policy/entropy_avg": 0.29206639528274536,
|
| 1561 |
+
"step": 74,
|
| 1562 |
+
"val/clipfrac_avg": 0.05188679322600365,
|
| 1563 |
+
"val/num_eos_tokens": 0,
|
| 1564 |
+
"val/ratio": 0.9267533421516418,
|
| 1565 |
+
"val/ratio_var": 0.0001095838742912747
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"episode": 1200,
|
| 1569 |
+
"epoch": 2.1052631578947367,
|
| 1570 |
+
"eps": 3,
|
| 1571 |
+
"loss/policy_avg": 0.004760343115776777,
|
| 1572 |
+
"loss/value_avg": 0.3549901843070984,
|
| 1573 |
+
"lr": 4.3485981308411215e-06,
|
| 1574 |
+
"objective/entropy": 2.9447989463806152,
|
| 1575 |
+
"objective/kl": 175.41961669921875,
|
| 1576 |
+
"objective/non_score_reward": -8.770980834960938,
|
| 1577 |
+
"objective/rlhf_reward": -3.2084808349609375,
|
| 1578 |
+
"objective/scores": 5.5625,
|
| 1579 |
+
"policy/approxkl_avg": 4.80606746673584,
|
| 1580 |
+
"policy/clipfrac_avg": 0.03419811278581619,
|
| 1581 |
+
"policy/entropy_avg": 0.30916815996170044,
|
| 1582 |
+
"step": 75,
|
| 1583 |
+
"val/clipfrac_avg": 0.004127358552068472,
|
| 1584 |
+
"val/num_eos_tokens": 0,
|
| 1585 |
+
"val/ratio": 0.9180092215538025,
|
| 1586 |
+
"val/ratio_var": 2.3069829694577493e-05
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"episode": 1216,
|
| 1590 |
+
"epoch": 2.1333333333333333,
|
| 1591 |
+
"eps": 3,
|
| 1592 |
+
"loss/policy_avg": 0.010298425331711769,
|
| 1593 |
+
"loss/value_avg": 0.15927723050117493,
|
| 1594 |
+
"lr": 4.216822429906542e-06,
|
| 1595 |
+
"objective/entropy": 1.4227180480957031,
|
| 1596 |
+
"objective/kl": 176.0067138671875,
|
| 1597 |
+
"objective/non_score_reward": -8.800336837768555,
|
| 1598 |
+
"objective/rlhf_reward": -2.6753368377685547,
|
| 1599 |
+
"objective/scores": 6.125,
|
| 1600 |
+
"policy/approxkl_avg": 4.99057149887085,
|
| 1601 |
+
"policy/clipfrac_avg": 0.028301887214183807,
|
| 1602 |
+
"policy/entropy_avg": 0.2751670479774475,
|
| 1603 |
+
"step": 76,
|
| 1604 |
+
"val/clipfrac_avg": 0.10495282709598541,
|
| 1605 |
+
"val/num_eos_tokens": 0,
|
| 1606 |
+
"val/ratio": 0.9201045036315918,
|
| 1607 |
+
"val/ratio_var": 5.366753157431958e-06
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"episode": 1232,
|
| 1611 |
+
"epoch": 2.16140350877193,
|
| 1612 |
+
"eps": 3,
|
| 1613 |
+
"loss/policy_avg": -0.005462624132633209,
|
| 1614 |
+
"loss/value_avg": 0.28704196214675903,
|
| 1615 |
+
"lr": 4.085046728971963e-06,
|
| 1616 |
+
"objective/entropy": 1.6171071529388428,
|
| 1617 |
+
"objective/kl": 176.83685302734375,
|
| 1618 |
+
"objective/non_score_reward": -8.841842651367188,
|
| 1619 |
+
"objective/rlhf_reward": -3.1543426513671875,
|
| 1620 |
+
"objective/scores": 5.6875,
|
| 1621 |
+
"policy/approxkl_avg": 5.847208023071289,
|
| 1622 |
+
"policy/clipfrac_avg": 0.028891509398818016,
|
| 1623 |
+
"policy/entropy_avg": 0.286138117313385,
|
| 1624 |
+
"step": 77,
|
| 1625 |
+
"val/clipfrac_avg": 0.07075471431016922,
|
| 1626 |
+
"val/num_eos_tokens": 0,
|
| 1627 |
+
"val/ratio": 0.9195102453231812,
|
| 1628 |
+
"val/ratio_var": 9.577343917044345e-06
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"episode": 1248,
|
| 1632 |
+
"epoch": 2.1894736842105265,
|
| 1633 |
+
"eps": 3,
|
| 1634 |
+
"loss/policy_avg": 0.0010141655802726746,
|
| 1635 |
+
"loss/value_avg": 0.8408201932907104,
|
| 1636 |
+
"lr": 3.953271028037383e-06,
|
| 1637 |
+
"objective/entropy": 7.40260124206543,
|
| 1638 |
+
"objective/kl": 177.4427490234375,
|
| 1639 |
+
"objective/non_score_reward": -8.872137069702148,
|
| 1640 |
+
"objective/rlhf_reward": -4.903387069702148,
|
| 1641 |
+
"objective/scores": 3.96875,
|
| 1642 |
+
"policy/approxkl_avg": 5.285105228424072,
|
| 1643 |
+
"policy/clipfrac_avg": 0.04658018797636032,
|
| 1644 |
+
"policy/entropy_avg": 0.4253733158111572,
|
| 1645 |
+
"step": 78,
|
| 1646 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1647 |
+
"val/num_eos_tokens": 0,
|
| 1648 |
+
"val/ratio": 0.9203585982322693,
|
| 1649 |
+
"val/ratio_var": 1.65566543728346e-05
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"episode": 1264,
|
| 1653 |
+
"epoch": 2.2175438596491226,
|
| 1654 |
+
"eps": 3,
|
| 1655 |
+
"loss/policy_avg": -0.004624534398317337,
|
| 1656 |
+
"loss/value_avg": 0.7719740271568298,
|
| 1657 |
+
"lr": 3.821495327102804e-06,
|
| 1658 |
+
"objective/entropy": 4.648886203765869,
|
| 1659 |
+
"objective/kl": 181.51986694335938,
|
| 1660 |
+
"objective/non_score_reward": -9.075994491577148,
|
| 1661 |
+
"objective/rlhf_reward": -4.700994491577148,
|
| 1662 |
+
"objective/scores": 4.375,
|
| 1663 |
+
"policy/approxkl_avg": 4.547338485717773,
|
| 1664 |
+
"policy/clipfrac_avg": 0.0383254736661911,
|
| 1665 |
+
"policy/entropy_avg": 0.3513880968093872,
|
| 1666 |
+
"step": 79,
|
| 1667 |
+
"val/clipfrac_avg": 0.1291273534297943,
|
| 1668 |
+
"val/num_eos_tokens": 0,
|
| 1669 |
+
"val/ratio": 0.9286638498306274,
|
| 1670 |
+
"val/ratio_var": 6.422147998819128e-05
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"episode": 1280,
|
| 1674 |
+
"epoch": 2.245614035087719,
|
| 1675 |
+
"eps": 3,
|
| 1676 |
+
"loss/policy_avg": 0.012380128726363182,
|
| 1677 |
+
"loss/value_avg": 0.40563684701919556,
|
| 1678 |
+
"lr": 3.689719626168224e-06,
|
| 1679 |
+
"objective/entropy": 7.685408115386963,
|
| 1680 |
+
"objective/kl": 168.90484619140625,
|
| 1681 |
+
"objective/non_score_reward": -8.445242881774902,
|
| 1682 |
+
"objective/rlhf_reward": -3.4452428817749023,
|
| 1683 |
+
"objective/scores": 5.0,
|
| 1684 |
+
"policy/approxkl_avg": 3.4143970012664795,
|
| 1685 |
+
"policy/clipfrac_avg": 0.041273586452007294,
|
| 1686 |
+
"policy/entropy_avg": 0.4100227355957031,
|
| 1687 |
+
"step": 80,
|
| 1688 |
+
"val/clipfrac_avg": 0.015330187976360321,
|
| 1689 |
+
"val/num_eos_tokens": 0,
|
| 1690 |
+
"val/ratio": 0.9217012524604797,
|
| 1691 |
+
"val/ratio_var": 7.061174983391538e-05
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"episode": 1296,
|
| 1695 |
+
"epoch": 2.2736842105263158,
|
| 1696 |
+
"eps": 3,
|
| 1697 |
+
"loss/policy_avg": 0.011339722201228142,
|
| 1698 |
+
"loss/value_avg": 0.3490160405635834,
|
| 1699 |
+
"lr": 3.5579439252336446e-06,
|
| 1700 |
+
"objective/entropy": 4.046834945678711,
|
| 1701 |
+
"objective/kl": 177.92718505859375,
|
| 1702 |
+
"objective/non_score_reward": -8.89635944366455,
|
| 1703 |
+
"objective/rlhf_reward": -3.958859443664551,
|
| 1704 |
+
"objective/scores": 4.9375,
|
| 1705 |
+
"policy/approxkl_avg": 5.583766460418701,
|
| 1706 |
+
"policy/clipfrac_avg": 0.03478773683309555,
|
| 1707 |
+
"policy/entropy_avg": 0.3193933963775635,
|
| 1708 |
+
"step": 81,
|
| 1709 |
+
"val/clipfrac_avg": 0.0,
|
| 1710 |
+
"val/num_eos_tokens": 0,
|
| 1711 |
+
"val/ratio": 0.9363265037536621,
|
| 1712 |
+
"val/ratio_var": 5.8069101214641705e-05
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"episode": 1312,
|
| 1716 |
+
"epoch": 2.3017543859649123,
|
| 1717 |
+
"eps": 3,
|
| 1718 |
+
"loss/policy_avg": 0.007465606089681387,
|
| 1719 |
+
"loss/value_avg": 0.3137081563472748,
|
| 1720 |
+
"lr": 3.4261682242990656e-06,
|
| 1721 |
+
"objective/entropy": 3.293423652648926,
|
| 1722 |
+
"objective/kl": 173.18377685546875,
|
| 1723 |
+
"objective/non_score_reward": -8.659189224243164,
|
| 1724 |
+
"objective/rlhf_reward": -3.065439224243164,
|
| 1725 |
+
"objective/scores": 5.59375,
|
| 1726 |
+
"policy/approxkl_avg": 4.794089317321777,
|
| 1727 |
+
"policy/clipfrac_avg": 0.0383254736661911,
|
| 1728 |
+
"policy/entropy_avg": 0.29685914516448975,
|
| 1729 |
+
"step": 82,
|
| 1730 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1731 |
+
"val/num_eos_tokens": 0,
|
| 1732 |
+
"val/ratio": 0.9458816647529602,
|
| 1733 |
+
"val/ratio_var": 0.00023432180751115084
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"episode": 1328,
|
| 1737 |
+
"epoch": 2.329824561403509,
|
| 1738 |
+
"eps": 3,
|
| 1739 |
+
"loss/policy_avg": -0.00093865767121315,
|
| 1740 |
+
"loss/value_avg": 0.9402576684951782,
|
| 1741 |
+
"lr": 3.294392523364486e-06,
|
| 1742 |
+
"objective/entropy": 5.09280252456665,
|
| 1743 |
+
"objective/kl": 173.88351440429688,
|
| 1744 |
+
"objective/non_score_reward": -8.694175720214844,
|
| 1745 |
+
"objective/rlhf_reward": -4.866050720214844,
|
| 1746 |
+
"objective/scores": 3.828125,
|
| 1747 |
+
"policy/approxkl_avg": 3.8168904781341553,
|
| 1748 |
+
"policy/clipfrac_avg": 0.03891509398818016,
|
| 1749 |
+
"policy/entropy_avg": 0.35750845074653625,
|
| 1750 |
+
"step": 83,
|
| 1751 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1752 |
+
"val/num_eos_tokens": 0,
|
| 1753 |
+
"val/ratio": 0.9232673645019531,
|
| 1754 |
+
"val/ratio_var": 4.1539384255884215e-05
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"episode": 1344,
|
| 1758 |
+
"epoch": 2.357894736842105,
|
| 1759 |
+
"eps": 3,
|
| 1760 |
+
"loss/policy_avg": -0.007357731461524963,
|
| 1761 |
+
"loss/value_avg": 0.36178284883499146,
|
| 1762 |
+
"lr": 3.1626168224299067e-06,
|
| 1763 |
+
"objective/entropy": 5.281716346740723,
|
| 1764 |
+
"objective/kl": 179.2125701904297,
|
| 1765 |
+
"objective/non_score_reward": -8.960628509521484,
|
| 1766 |
+
"objective/rlhf_reward": -3.9918785095214844,
|
| 1767 |
+
"objective/scores": 4.96875,
|
| 1768 |
+
"policy/approxkl_avg": 4.461269378662109,
|
| 1769 |
+
"policy/clipfrac_avg": 0.05365566164255142,
|
| 1770 |
+
"policy/entropy_avg": 0.35694169998168945,
|
| 1771 |
+
"step": 84,
|
| 1772 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1773 |
+
"val/num_eos_tokens": 0,
|
| 1774 |
+
"val/ratio": 0.9178085923194885,
|
| 1775 |
+
"val/ratio_var": 4.949720823788084e-05
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"episode": 1360,
|
| 1779 |
+
"epoch": 2.3859649122807016,
|
| 1780 |
+
"eps": 3,
|
| 1781 |
+
"loss/policy_avg": 0.0004696398973464966,
|
| 1782 |
+
"loss/value_avg": 0.30143094062805176,
|
| 1783 |
+
"lr": 3.0308411214953273e-06,
|
| 1784 |
+
"objective/entropy": 2.755769729614258,
|
| 1785 |
+
"objective/kl": 173.08140563964844,
|
| 1786 |
+
"objective/non_score_reward": -8.654069900512695,
|
| 1787 |
+
"objective/rlhf_reward": -2.9665699005126953,
|
| 1788 |
+
"objective/scores": 5.6875,
|
| 1789 |
+
"policy/approxkl_avg": 5.356992721557617,
|
| 1790 |
+
"policy/clipfrac_avg": 0.03242924436926842,
|
| 1791 |
+
"policy/entropy_avg": 0.282896488904953,
|
| 1792 |
+
"step": 85,
|
| 1793 |
+
"val/clipfrac_avg": 0.0,
|
| 1794 |
+
"val/num_eos_tokens": 0,
|
| 1795 |
+
"val/ratio": 0.9419326782226562,
|
| 1796 |
+
"val/ratio_var": 3.9359238144243136e-05
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"episode": 1376,
|
| 1800 |
+
"epoch": 2.414035087719298,
|
| 1801 |
+
"eps": 3,
|
| 1802 |
+
"loss/policy_avg": 0.0008706599473953247,
|
| 1803 |
+
"loss/value_avg": 0.5158276557922363,
|
| 1804 |
+
"lr": 2.8990654205607475e-06,
|
| 1805 |
+
"objective/entropy": 4.149250507354736,
|
| 1806 |
+
"objective/kl": 173.71505737304688,
|
| 1807 |
+
"objective/non_score_reward": -8.685752868652344,
|
| 1808 |
+
"objective/rlhf_reward": -3.6857528686523438,
|
| 1809 |
+
"objective/scores": 5.0,
|
| 1810 |
+
"policy/approxkl_avg": 5.095344066619873,
|
| 1811 |
+
"policy/clipfrac_avg": 0.030070755630731583,
|
| 1812 |
+
"policy/entropy_avg": 0.3076534867286682,
|
| 1813 |
+
"step": 86,
|
| 1814 |
+
"val/clipfrac_avg": 0.0,
|
| 1815 |
+
"val/num_eos_tokens": 0,
|
| 1816 |
+
"val/ratio": 0.93389892578125,
|
| 1817 |
+
"val/ratio_var": 9.108168342208955e-06
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"episode": 1392,
|
| 1821 |
+
"epoch": 2.442105263157895,
|
| 1822 |
+
"eps": 3,
|
| 1823 |
+
"loss/policy_avg": -0.0013641200494021177,
|
| 1824 |
+
"loss/value_avg": 0.46665364503860474,
|
| 1825 |
+
"lr": 2.767289719626168e-06,
|
| 1826 |
+
"objective/entropy": 3.9847404956817627,
|
| 1827 |
+
"objective/kl": 171.97903442382812,
|
| 1828 |
+
"objective/non_score_reward": -8.59895133972168,
|
| 1829 |
+
"objective/rlhf_reward": -3.4114513397216797,
|
| 1830 |
+
"objective/scores": 5.1875,
|
| 1831 |
+
"policy/approxkl_avg": 4.758839130401611,
|
| 1832 |
+
"policy/clipfrac_avg": 0.02771226316690445,
|
| 1833 |
+
"policy/entropy_avg": 0.3068329691886902,
|
| 1834 |
+
"step": 87,
|
| 1835 |
+
"val/clipfrac_avg": 0.002358490601181984,
|
| 1836 |
+
"val/num_eos_tokens": 0,
|
| 1837 |
+
"val/ratio": 0.9303795099258423,
|
| 1838 |
+
"val/ratio_var": 2.7705931643140502e-05
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"episode": 1408,
|
| 1842 |
+
"epoch": 2.4701754385964914,
|
| 1843 |
+
"eps": 3,
|
| 1844 |
+
"loss/policy_avg": -0.009293105453252792,
|
| 1845 |
+
"loss/value_avg": 0.1374308168888092,
|
| 1846 |
+
"lr": 2.6355140186915887e-06,
|
| 1847 |
+
"objective/entropy": 2.8504319190979004,
|
| 1848 |
+
"objective/kl": 178.8887176513672,
|
| 1849 |
+
"objective/non_score_reward": -8.944437026977539,
|
| 1850 |
+
"objective/rlhf_reward": -2.975687026977539,
|
| 1851 |
+
"objective/scores": 5.96875,
|
| 1852 |
+
"policy/approxkl_avg": 5.254701614379883,
|
| 1853 |
+
"policy/clipfrac_avg": 0.026533018797636032,
|
| 1854 |
+
"policy/entropy_avg": 0.2935040593147278,
|
| 1855 |
+
"step": 88,
|
| 1856 |
+
"val/clipfrac_avg": 0.002358490601181984,
|
| 1857 |
+
"val/num_eos_tokens": 0,
|
| 1858 |
+
"val/ratio": 0.9373711943626404,
|
| 1859 |
+
"val/ratio_var": 1.766591776686255e-05
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"episode": 1424,
|
| 1863 |
+
"epoch": 2.498245614035088,
|
| 1864 |
+
"eps": 3,
|
| 1865 |
+
"loss/policy_avg": 0.00495288148522377,
|
| 1866 |
+
"loss/value_avg": 0.20061969757080078,
|
| 1867 |
+
"lr": 2.5037383177570093e-06,
|
| 1868 |
+
"objective/entropy": 4.51104211807251,
|
| 1869 |
+
"objective/kl": 169.7410125732422,
|
| 1870 |
+
"objective/non_score_reward": -8.487051010131836,
|
| 1871 |
+
"objective/rlhf_reward": -2.768301010131836,
|
| 1872 |
+
"objective/scores": 5.71875,
|
| 1873 |
+
"policy/approxkl_avg": 4.481791019439697,
|
| 1874 |
+
"policy/clipfrac_avg": 0.03478773683309555,
|
| 1875 |
+
"policy/entropy_avg": 0.3265501856803894,
|
| 1876 |
+
"step": 89,
|
| 1877 |
+
"val/clipfrac_avg": 0.0,
|
| 1878 |
+
"val/num_eos_tokens": 0,
|
| 1879 |
+
"val/ratio": 0.9414163827896118,
|
| 1880 |
+
"val/ratio_var": 0.00011632378300419077
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"episode": 1440,
|
| 1884 |
+
"epoch": 2.526315789473684,
|
| 1885 |
+
"eps": 3,
|
| 1886 |
+
"loss/policy_avg": 0.00646105594933033,
|
| 1887 |
+
"loss/value_avg": 0.42740941047668457,
|
| 1888 |
+
"lr": 2.37196261682243e-06,
|
| 1889 |
+
"objective/entropy": 3.2403650283813477,
|
| 1890 |
+
"objective/kl": 176.238037109375,
|
| 1891 |
+
"objective/non_score_reward": -8.811902046203613,
|
| 1892 |
+
"objective/rlhf_reward": -3.9994020462036133,
|
| 1893 |
+
"objective/scores": 4.8125,
|
| 1894 |
+
"policy/approxkl_avg": 5.44842529296875,
|
| 1895 |
+
"policy/clipfrac_avg": 0.01591981202363968,
|
| 1896 |
+
"policy/entropy_avg": 0.2933845520019531,
|
| 1897 |
+
"step": 90,
|
| 1898 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1899 |
+
"val/num_eos_tokens": 0,
|
| 1900 |
+
"val/ratio": 0.9381667971611023,
|
| 1901 |
+
"val/ratio_var": 9.369335020892322e-05
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"episode": 1456,
|
| 1905 |
+
"epoch": 2.5543859649122806,
|
| 1906 |
+
"eps": 3,
|
| 1907 |
+
"loss/policy_avg": -0.005389541387557983,
|
| 1908 |
+
"loss/value_avg": 0.4948211908340454,
|
| 1909 |
+
"lr": 2.2401869158878504e-06,
|
| 1910 |
+
"objective/entropy": 2.898387908935547,
|
| 1911 |
+
"objective/kl": 173.48486328125,
|
| 1912 |
+
"objective/non_score_reward": -8.674242973327637,
|
| 1913 |
+
"objective/rlhf_reward": -3.5179929733276367,
|
| 1914 |
+
"objective/scores": 5.15625,
|
| 1915 |
+
"policy/approxkl_avg": 4.66801643371582,
|
| 1916 |
+
"policy/clipfrac_avg": 0.020636793226003647,
|
| 1917 |
+
"policy/entropy_avg": 0.2913670837879181,
|
| 1918 |
+
"step": 91,
|
| 1919 |
+
"val/clipfrac_avg": 0.001179245300590992,
|
| 1920 |
+
"val/num_eos_tokens": 0,
|
| 1921 |
+
"val/ratio": 0.9378049373626709,
|
| 1922 |
+
"val/ratio_var": 1.2327662261668593e-05
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"episode": 1472,
|
| 1926 |
+
"epoch": 2.5824561403508772,
|
| 1927 |
+
"eps": 3,
|
| 1928 |
+
"loss/policy_avg": -0.010267895646393299,
|
| 1929 |
+
"loss/value_avg": 0.26834648847579956,
|
| 1930 |
+
"lr": 2.108411214953271e-06,
|
| 1931 |
+
"objective/entropy": 4.616816997528076,
|
| 1932 |
+
"objective/kl": 171.12762451171875,
|
| 1933 |
+
"objective/non_score_reward": -8.556382179260254,
|
| 1934 |
+
"objective/rlhf_reward": -3.587632179260254,
|
| 1935 |
+
"objective/scores": 4.96875,
|
| 1936 |
+
"policy/approxkl_avg": 4.146580219268799,
|
| 1937 |
+
"policy/clipfrac_avg": 0.041273586452007294,
|
| 1938 |
+
"policy/entropy_avg": 0.34417253732681274,
|
| 1939 |
+
"step": 92,
|
| 1940 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 1941 |
+
"val/num_eos_tokens": 0,
|
| 1942 |
+
"val/ratio": 0.9241670370101929,
|
| 1943 |
+
"val/ratio_var": 4.9057460273616016e-05
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"episode": 1488,
|
| 1947 |
+
"epoch": 2.610526315789474,
|
| 1948 |
+
"eps": 3,
|
| 1949 |
+
"loss/policy_avg": 0.0006395354866981506,
|
| 1950 |
+
"loss/value_avg": 0.7872554063796997,
|
| 1951 |
+
"lr": 1.9766355140186916e-06,
|
| 1952 |
+
"objective/entropy": 5.483046531677246,
|
| 1953 |
+
"objective/kl": 169.1695098876953,
|
| 1954 |
+
"objective/non_score_reward": -8.458476066589355,
|
| 1955 |
+
"objective/rlhf_reward": -4.4741010665893555,
|
| 1956 |
+
"objective/scores": 3.984375,
|
| 1957 |
+
"policy/approxkl_avg": 2.8852078914642334,
|
| 1958 |
+
"policy/clipfrac_avg": 0.032429248094558716,
|
| 1959 |
+
"policy/entropy_avg": 0.35312554240226746,
|
| 1960 |
+
"step": 93,
|
| 1961 |
+
"val/clipfrac_avg": 0.001768867950886488,
|
| 1962 |
+
"val/num_eos_tokens": 0,
|
| 1963 |
+
"val/ratio": 0.9286659955978394,
|
| 1964 |
+
"val/ratio_var": 2.8370095606078394e-05
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"episode": 1504,
|
| 1968 |
+
"epoch": 2.6385964912280704,
|
| 1969 |
+
"eps": 3,
|
| 1970 |
+
"loss/policy_avg": -0.00652042031288147,
|
| 1971 |
+
"loss/value_avg": 0.17014235258102417,
|
| 1972 |
+
"lr": 1.844859813084112e-06,
|
| 1973 |
+
"objective/entropy": 2.737617015838623,
|
| 1974 |
+
"objective/kl": 178.22747802734375,
|
| 1975 |
+
"objective/non_score_reward": -8.911375045776367,
|
| 1976 |
+
"objective/rlhf_reward": -3.286375045776367,
|
| 1977 |
+
"objective/scores": 5.625,
|
| 1978 |
+
"policy/approxkl_avg": 5.498225688934326,
|
| 1979 |
+
"policy/clipfrac_avg": 0.028891511261463165,
|
| 1980 |
+
"policy/entropy_avg": 0.28995558619499207,
|
| 1981 |
+
"step": 94,
|
| 1982 |
+
"val/clipfrac_avg": 0.001768867950886488,
|
| 1983 |
+
"val/num_eos_tokens": 0,
|
| 1984 |
+
"val/ratio": 0.9345540404319763,
|
| 1985 |
+
"val/ratio_var": 1.2709216434814152e-06
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"episode": 1520,
|
| 1989 |
+
"epoch": 2.6666666666666665,
|
| 1990 |
+
"eps": 3,
|
| 1991 |
+
"loss/policy_avg": 0.007195580750703812,
|
| 1992 |
+
"loss/value_avg": 0.40153437852859497,
|
| 1993 |
+
"lr": 1.7130841121495328e-06,
|
| 1994 |
+
"objective/entropy": 4.30942440032959,
|
| 1995 |
+
"objective/kl": 176.95938110351562,
|
| 1996 |
+
"objective/non_score_reward": -8.847970008850098,
|
| 1997 |
+
"objective/rlhf_reward": -3.8479700088500977,
|
| 1998 |
+
"objective/scores": 5.0,
|
| 1999 |
+
"policy/approxkl_avg": 5.267421245574951,
|
| 2000 |
+
"policy/clipfrac_avg": 0.03125,
|
| 2001 |
+
"policy/entropy_avg": 0.31008654832839966,
|
| 2002 |
+
"step": 95,
|
| 2003 |
+
"val/clipfrac_avg": 0.0,
|
| 2004 |
+
"val/num_eos_tokens": 0,
|
| 2005 |
+
"val/ratio": 0.9336869716644287,
|
| 2006 |
+
"val/ratio_var": 4.966981941834092e-05
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"episode": 1536,
|
| 2010 |
+
"epoch": 2.694736842105263,
|
| 2011 |
+
"eps": 3,
|
| 2012 |
+
"loss/policy_avg": 0.012591801583766937,
|
| 2013 |
+
"loss/value_avg": 0.3597390055656433,
|
| 2014 |
+
"lr": 1.5813084112149534e-06,
|
| 2015 |
+
"objective/entropy": 5.459916591644287,
|
| 2016 |
+
"objective/kl": 172.44110107421875,
|
| 2017 |
+
"objective/non_score_reward": -8.622055053710938,
|
| 2018 |
+
"objective/rlhf_reward": -3.6220550537109375,
|
| 2019 |
+
"objective/scores": 5.0,
|
| 2020 |
+
"policy/approxkl_avg": 4.5339765548706055,
|
| 2021 |
+
"policy/clipfrac_avg": 0.03537736088037491,
|
| 2022 |
+
"policy/entropy_avg": 0.3411254286766052,
|
| 2023 |
+
"step": 96,
|
| 2024 |
+
"val/clipfrac_avg": 0.0,
|
| 2025 |
+
"val/num_eos_tokens": 0,
|
| 2026 |
+
"val/ratio": 0.9462149739265442,
|
| 2027 |
+
"val/ratio_var": 3.8459929783130065e-05
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"episode": 1552,
|
| 2031 |
+
"epoch": 2.7228070175438597,
|
| 2032 |
+
"eps": 3,
|
| 2033 |
+
"loss/policy_avg": -0.003356472123414278,
|
| 2034 |
+
"loss/value_avg": 0.6434417963027954,
|
| 2035 |
+
"lr": 1.4495327102803737e-06,
|
| 2036 |
+
"objective/entropy": 5.633913516998291,
|
| 2037 |
+
"objective/kl": 172.26502990722656,
|
| 2038 |
+
"objective/non_score_reward": -8.613250732421875,
|
| 2039 |
+
"objective/rlhf_reward": -4.363250732421875,
|
| 2040 |
+
"objective/scores": 4.25,
|
| 2041 |
+
"policy/approxkl_avg": 3.585165500640869,
|
| 2042 |
+
"policy/clipfrac_avg": 0.03537735715508461,
|
| 2043 |
+
"policy/entropy_avg": 0.34199586510658264,
|
| 2044 |
+
"step": 97,
|
| 2045 |
+
"val/clipfrac_avg": 0.001179245300590992,
|
| 2046 |
+
"val/num_eos_tokens": 0,
|
| 2047 |
+
"val/ratio": 0.9311625957489014,
|
| 2048 |
+
"val/ratio_var": 6.935850979061797e-05
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"episode": 1568,
|
| 2052 |
+
"epoch": 2.7508771929824563,
|
| 2053 |
+
"eps": 3,
|
| 2054 |
+
"loss/policy_avg": -0.003898909315466881,
|
| 2055 |
+
"loss/value_avg": 0.36550819873809814,
|
| 2056 |
+
"lr": 1.3177570093457943e-06,
|
| 2057 |
+
"objective/entropy": 4.281040191650391,
|
| 2058 |
+
"objective/kl": 174.15972900390625,
|
| 2059 |
+
"objective/non_score_reward": -8.707986831665039,
|
| 2060 |
+
"objective/rlhf_reward": -3.614236831665039,
|
| 2061 |
+
"objective/scores": 5.09375,
|
| 2062 |
+
"policy/approxkl_avg": 5.1715850830078125,
|
| 2063 |
+
"policy/clipfrac_avg": 0.02712264284491539,
|
| 2064 |
+
"policy/entropy_avg": 0.3137935698032379,
|
| 2065 |
+
"step": 98,
|
| 2066 |
+
"val/clipfrac_avg": 0.0,
|
| 2067 |
+
"val/num_eos_tokens": 0,
|
| 2068 |
+
"val/ratio": 0.9350335597991943,
|
| 2069 |
+
"val/ratio_var": 1.248767239303561e-05
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"episode": 1584,
|
| 2073 |
+
"epoch": 2.778947368421053,
|
| 2074 |
+
"eps": 3,
|
| 2075 |
+
"loss/policy_avg": 0.0017306804656982422,
|
| 2076 |
+
"loss/value_avg": 0.2737918496131897,
|
| 2077 |
+
"lr": 1.185981308411215e-06,
|
| 2078 |
+
"objective/entropy": 5.210065841674805,
|
| 2079 |
+
"objective/kl": 173.97068786621094,
|
| 2080 |
+
"objective/non_score_reward": -8.69853401184082,
|
| 2081 |
+
"objective/rlhf_reward": -3.8860340118408203,
|
| 2082 |
+
"objective/scores": 4.8125,
|
| 2083 |
+
"policy/approxkl_avg": 5.011469841003418,
|
| 2084 |
+
"policy/clipfrac_avg": 0.04304245114326477,
|
| 2085 |
+
"policy/entropy_avg": 0.3419041931629181,
|
| 2086 |
+
"step": 99,
|
| 2087 |
+
"val/clipfrac_avg": 0.0,
|
| 2088 |
+
"val/num_eos_tokens": 0,
|
| 2089 |
+
"val/ratio": 0.9419320225715637,
|
| 2090 |
+
"val/ratio_var": 8.726042324269656e-06
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"episode": 1600,
|
| 2094 |
+
"epoch": 2.807017543859649,
|
| 2095 |
+
"eps": 3,
|
| 2096 |
+
"loss/policy_avg": -0.006221463903784752,
|
| 2097 |
+
"loss/value_avg": 0.3625496029853821,
|
| 2098 |
+
"lr": 1.0542056074766355e-06,
|
| 2099 |
+
"objective/entropy": 3.721562623977661,
|
| 2100 |
+
"objective/kl": 175.773193359375,
|
| 2101 |
+
"objective/non_score_reward": -8.78865909576416,
|
| 2102 |
+
"objective/rlhf_reward": -3.47615909576416,
|
| 2103 |
+
"objective/scores": 5.3125,
|
| 2104 |
+
"policy/approxkl_avg": 5.388751029968262,
|
| 2105 |
+
"policy/clipfrac_avg": 0.03419811278581619,
|
| 2106 |
+
"policy/entropy_avg": 0.29315799474716187,
|
| 2107 |
+
"step": 100,
|
| 2108 |
+
"val/clipfrac_avg": 0.0,
|
| 2109 |
+
"val/num_eos_tokens": 0,
|
| 2110 |
+
"val/ratio": 0.949840784072876,
|
| 2111 |
+
"val/ratio_var": 3.881535303662531e-05
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"episode": 1616,
|
| 2115 |
+
"epoch": 2.8350877192982455,
|
| 2116 |
+
"eps": 3,
|
| 2117 |
+
"loss/policy_avg": -0.005170758813619614,
|
| 2118 |
+
"loss/value_avg": 0.4136154055595398,
|
| 2119 |
+
"lr": 9.22429906542056e-07,
|
| 2120 |
+
"objective/entropy": 5.907715320587158,
|
| 2121 |
+
"objective/kl": 171.47390747070312,
|
| 2122 |
+
"objective/non_score_reward": -8.573695182800293,
|
| 2123 |
+
"objective/rlhf_reward": -4.167445182800293,
|
| 2124 |
+
"objective/scores": 4.40625,
|
| 2125 |
+
"policy/approxkl_avg": 5.403087615966797,
|
| 2126 |
+
"policy/clipfrac_avg": 0.03478773683309555,
|
| 2127 |
+
"policy/entropy_avg": 0.33564049005508423,
|
| 2128 |
+
"step": 101,
|
| 2129 |
+
"val/clipfrac_avg": 0.000589622650295496,
|
| 2130 |
+
"val/num_eos_tokens": 0,
|
| 2131 |
+
"val/ratio": 0.9340347051620483,
|
| 2132 |
+
"val/ratio_var": 3.0960076401242986e-05
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"episode": 1632,
|
| 2136 |
+
"epoch": 2.863157894736842,
|
| 2137 |
+
"eps": 3,
|
| 2138 |
+
"loss/policy_avg": 0.002457182854413986,
|
| 2139 |
+
"loss/value_avg": 0.27742013335227966,
|
| 2140 |
+
"lr": 7.906542056074767e-07,
|
| 2141 |
+
"objective/entropy": 5.222499370574951,
|
| 2142 |
+
"objective/kl": 176.05380249023438,
|
| 2143 |
+
"objective/non_score_reward": -8.802690505981445,
|
| 2144 |
+
"objective/rlhf_reward": -3.6151905059814453,
|
| 2145 |
+
"objective/scores": 5.1875,
|
| 2146 |
+
"policy/approxkl_avg": 4.675132751464844,
|
| 2147 |
+
"policy/clipfrac_avg": 0.04304245114326477,
|
| 2148 |
+
"policy/entropy_avg": 0.3403066396713257,
|
| 2149 |
+
"step": 102,
|
| 2150 |
+
"val/clipfrac_avg": 0.0,
|
| 2151 |
+
"val/num_eos_tokens": 0,
|
| 2152 |
+
"val/ratio": 0.9404515027999878,
|
| 2153 |
+
"val/ratio_var": 7.289678615052253e-05
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"episode": 1648,
|
| 2157 |
+
"epoch": 2.8912280701754387,
|
| 2158 |
+
"eps": 3,
|
| 2159 |
+
"loss/policy_avg": -0.006544323638081551,
|
| 2160 |
+
"loss/value_avg": 0.29731422662734985,
|
| 2161 |
+
"lr": 6.588785046728972e-07,
|
| 2162 |
+
"objective/entropy": 4.219725608825684,
|
| 2163 |
+
"objective/kl": 178.1557159423828,
|
| 2164 |
+
"objective/non_score_reward": -8.90778636932373,
|
| 2165 |
+
"objective/rlhf_reward": -3.8452863693237305,
|
| 2166 |
+
"objective/scores": 5.0625,
|
| 2167 |
+
"policy/approxkl_avg": 5.953890800476074,
|
| 2168 |
+
"policy/clipfrac_avg": 0.0383254699409008,
|
| 2169 |
+
"policy/entropy_avg": 0.31199511885643005,
|
| 2170 |
+
"step": 103,
|
| 2171 |
+
"val/clipfrac_avg": 0.0,
|
| 2172 |
+
"val/num_eos_tokens": 0,
|
| 2173 |
+
"val/ratio": 0.9407525658607483,
|
| 2174 |
+
"val/ratio_var": 5.8047560742124915e-05
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"episode": 1664,
|
| 2178 |
+
"epoch": 2.9192982456140353,
|
| 2179 |
+
"eps": 3,
|
| 2180 |
+
"loss/policy_avg": -0.0011880630627274513,
|
| 2181 |
+
"loss/value_avg": 0.21903052926063538,
|
| 2182 |
+
"lr": 5.271028037383178e-07,
|
| 2183 |
+
"objective/entropy": 5.557653427124023,
|
| 2184 |
+
"objective/kl": 170.518798828125,
|
| 2185 |
+
"objective/non_score_reward": -8.52593994140625,
|
| 2186 |
+
"objective/rlhf_reward": -3.05718994140625,
|
| 2187 |
+
"objective/scores": 5.46875,
|
| 2188 |
+
"policy/approxkl_avg": 4.447786331176758,
|
| 2189 |
+
"policy/clipfrac_avg": 0.0383254699409008,
|
| 2190 |
+
"policy/entropy_avg": 0.33431151509284973,
|
| 2191 |
+
"step": 104,
|
| 2192 |
+
"val/clipfrac_avg": 0.0,
|
| 2193 |
+
"val/num_eos_tokens": 0,
|
| 2194 |
+
"val/ratio": 0.9298925399780273,
|
| 2195 |
+
"val/ratio_var": 9.144405339611694e-06
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"episode": 1680,
|
| 2199 |
+
"epoch": 2.9473684210526314,
|
| 2200 |
+
"eps": 3,
|
| 2201 |
+
"loss/policy_avg": -0.0007513905875384808,
|
| 2202 |
+
"loss/value_avg": 0.21037007868289948,
|
| 2203 |
+
"lr": 3.9532710280373834e-07,
|
| 2204 |
+
"objective/entropy": 3.900575876235962,
|
| 2205 |
+
"objective/kl": 174.99456787109375,
|
| 2206 |
+
"objective/non_score_reward": -8.74972915649414,
|
| 2207 |
+
"objective/rlhf_reward": -3.5309791564941406,
|
| 2208 |
+
"objective/scores": 5.21875,
|
| 2209 |
+
"policy/approxkl_avg": 5.165627479553223,
|
| 2210 |
+
"policy/clipfrac_avg": 0.028301887214183807,
|
| 2211 |
+
"policy/entropy_avg": 0.2935909032821655,
|
| 2212 |
+
"step": 105,
|
| 2213 |
+
"val/clipfrac_avg": 0.001179245300590992,
|
| 2214 |
+
"val/num_eos_tokens": 0,
|
| 2215 |
+
"val/ratio": 0.9422957897186279,
|
| 2216 |
+
"val/ratio_var": 2.6614558009896427e-05
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"episode": 1696,
|
| 2220 |
+
"epoch": 2.975438596491228,
|
| 2221 |
+
"eps": 3,
|
| 2222 |
+
"loss/policy_avg": -0.005426734685897827,
|
| 2223 |
+
"loss/value_avg": 0.21496959030628204,
|
| 2224 |
+
"lr": 2.635514018691589e-07,
|
| 2225 |
+
"objective/entropy": 4.556634902954102,
|
| 2226 |
+
"objective/kl": 173.05136108398438,
|
| 2227 |
+
"objective/non_score_reward": -8.652568817138672,
|
| 2228 |
+
"objective/rlhf_reward": -2.933818817138672,
|
| 2229 |
+
"objective/scores": 5.71875,
|
| 2230 |
+
"policy/approxkl_avg": 4.820314884185791,
|
| 2231 |
+
"policy/clipfrac_avg": 0.04304245486855507,
|
| 2232 |
+
"policy/entropy_avg": 0.31966692209243774,
|
| 2233 |
+
"step": 106,
|
| 2234 |
+
"val/clipfrac_avg": 0.0,
|
| 2235 |
+
"val/num_eos_tokens": 0,
|
| 2236 |
+
"val/ratio": 0.9327390193939209,
|
| 2237 |
+
"val/ratio_var": 8.508096652803943e-05
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"episode": 1712,
|
| 2241 |
+
"epoch": 3.0035087719298246,
|
| 2242 |
+
"eps": 3,
|
| 2243 |
+
"loss/policy_avg": 0.0001406269147992134,
|
| 2244 |
+
"loss/value_avg": 0.186610609292984,
|
| 2245 |
+
"lr": 1.3177570093457944e-07,
|
| 2246 |
+
"objective/entropy": 4.999897003173828,
|
| 2247 |
+
"objective/kl": 173.25045776367188,
|
| 2248 |
+
"objective/non_score_reward": -8.66252326965332,
|
| 2249 |
+
"objective/rlhf_reward": -2.9437732696533203,
|
| 2250 |
+
"objective/scores": 5.71875,
|
| 2251 |
+
"policy/approxkl_avg": 4.337066173553467,
|
| 2252 |
+
"policy/clipfrac_avg": 0.04716981202363968,
|
| 2253 |
+
"policy/entropy_avg": 0.3558363914489746,
|
| 2254 |
+
"step": 107,
|
| 2255 |
+
"val/clipfrac_avg": 0.0,
|
| 2256 |
+
"val/num_eos_tokens": 0,
|
| 2257 |
+
"val/ratio": 0.9243938326835632,
|
| 2258 |
+
"val/ratio_var": 0.000128799001686275
|
| 2259 |
+
}
|
| 2260 |
+
],
|
| 2261 |
+
"logging_steps": 10,
|
| 2262 |
+
"max_steps": 107,
|
| 2263 |
+
"num_input_tokens_seen": 0,
|
| 2264 |
+
"num_train_epochs": 3.0,
|
| 2265 |
+
"save_steps": 500,
|
| 2266 |
+
"stateful_callbacks": {
|
| 2267 |
+
"TrainerControl": {
|
| 2268 |
+
"args": {
|
| 2269 |
+
"should_epoch_stop": false,
|
| 2270 |
+
"should_evaluate": false,
|
| 2271 |
+
"should_log": false,
|
| 2272 |
+
"should_save": true,
|
| 2273 |
+
"should_training_stop": true
|
| 2274 |
+
},
|
| 2275 |
+
"attributes": {}
|
| 2276 |
+
}
|
| 2277 |
+
},
|
| 2278 |
+
"total_flos": 0,
|
| 2279 |
+
"train_batch_size": null,
|
| 2280 |
+
"trial_name": null,
|
| 2281 |
+
"trial_params": null
|
| 2282 |
+
}
|
checkpoint-107/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ce71d5d5d23d373f1100beface744ea0e78db036a0645797d9ced3b0c1d8365
|
| 3 |
+
size 6673
|
checkpoint-107/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen2ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"dtype": "bfloat16",
|
| 7 |
+
"eos_token_id": 151645,
|
| 8 |
+
"hidden_act": "silu",
|
| 9 |
+
"hidden_size": 1536,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 8960,
|
| 12 |
+
"layer_types": [
|
| 13 |
+
"full_attention",
|
| 14 |
+
"full_attention",
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention"
|
| 41 |
+
],
|
| 42 |
+
"max_position_embeddings": 32768,
|
| 43 |
+
"max_window_layers": 21,
|
| 44 |
+
"model_type": "qwen2",
|
| 45 |
+
"num_attention_heads": 12,
|
| 46 |
+
"num_hidden_layers": 28,
|
| 47 |
+
"num_key_value_heads": 2,
|
| 48 |
+
"pad_token_id": 151643,
|
| 49 |
+
"rms_norm_eps": 1e-06,
|
| 50 |
+
"rope_scaling": null,
|
| 51 |
+
"rope_theta": 1000000.0,
|
| 52 |
+
"sliding_window": null,
|
| 53 |
+
"tie_word_embeddings": true,
|
| 54 |
+
"transformers_version": "4.57.1",
|
| 55 |
+
"use_cache": false,
|
| 56 |
+
"use_sliding_window": false,
|
| 57 |
+
"vocab_size": 151936
|
| 58 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"pad_token_id": 151643,
|
| 4 |
+
"repetition_penalty": 1.1,
|
| 5 |
+
"temperature": 0.7,
|
| 6 |
+
"top_k": 20,
|
| 7 |
+
"top_p": 0.8,
|
| 8 |
+
"transformers_version": "4.57.1"
|
| 9 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c93779623565bec515d60d0d5ba3ac8bb8d891825cd93bd6ffb072dbf265550
|
| 3 |
+
size 3087467144
|
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:f650899cdee1a74abb093b96e45774316750373414f9e11647b8480290d3937f
|
| 3 |
+
size 11421988
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,207 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
},
|
| 182 |
+
"additional_special_tokens": [
|
| 183 |
+
"<|im_start|>",
|
| 184 |
+
"<|im_end|>",
|
| 185 |
+
"<|object_ref_start|>",
|
| 186 |
+
"<|object_ref_end|>",
|
| 187 |
+
"<|box_start|>",
|
| 188 |
+
"<|box_end|>",
|
| 189 |
+
"<|quad_start|>",
|
| 190 |
+
"<|quad_end|>",
|
| 191 |
+
"<|vision_start|>",
|
| 192 |
+
"<|vision_end|>",
|
| 193 |
+
"<|vision_pad|>",
|
| 194 |
+
"<|image_pad|>",
|
| 195 |
+
"<|video_pad|>"
|
| 196 |
+
],
|
| 197 |
+
"bos_token": null,
|
| 198 |
+
"clean_up_tokenization_spaces": false,
|
| 199 |
+
"eos_token": "<|im_end|>",
|
| 200 |
+
"errors": "replace",
|
| 201 |
+
"extra_special_tokens": {},
|
| 202 |
+
"model_max_length": 131072,
|
| 203 |
+
"pad_token": "<|endoftext|>",
|
| 204 |
+
"split_special_tokens": false,
|
| 205 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 206 |
+
"unk_token": null
|
| 207 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ce71d5d5d23d373f1100beface744ea0e78db036a0645797d9ced3b0c1d8365
|
| 3 |
+
size 6673
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|