Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +14 -0
- README.md +58 -0
- chat_template.jinja +89 -0
- checkpoint-1000/chat_template.jinja +89 -0
- checkpoint-1000/config.json +63 -0
- checkpoint-1000/generation_config.json +12 -0
- checkpoint-1000/model.safetensors +3 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/tokenizer.json +3 -0
- checkpoint-1000/tokenizer_config.json +30 -0
- checkpoint-1000/trainer_state.json +234 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-1500/chat_template.jinja +89 -0
- checkpoint-1500/config.json +63 -0
- checkpoint-1500/generation_config.json +12 -0
- checkpoint-1500/model.safetensors +3 -0
- checkpoint-1500/optimizer.pt +3 -0
- checkpoint-1500/rng_state.pth +3 -0
- checkpoint-1500/scheduler.pt +3 -0
- checkpoint-1500/tokenizer.json +3 -0
- checkpoint-1500/tokenizer_config.json +30 -0
- checkpoint-1500/trainer_state.json +334 -0
- checkpoint-1500/training_args.bin +3 -0
- checkpoint-2000/chat_template.jinja +89 -0
- checkpoint-2000/config.json +63 -0
- checkpoint-2000/generation_config.json +12 -0
- checkpoint-2000/model.safetensors +3 -0
- checkpoint-2000/optimizer.pt +3 -0
- checkpoint-2000/rng_state.pth +3 -0
- checkpoint-2000/scheduler.pt +3 -0
- checkpoint-2000/tokenizer.json +3 -0
- checkpoint-2000/tokenizer_config.json +30 -0
- checkpoint-2000/trainer_state.json +434 -0
- checkpoint-2000/training_args.bin +3 -0
- checkpoint-2500/chat_template.jinja +89 -0
- checkpoint-2500/config.json +63 -0
- checkpoint-2500/generation_config.json +12 -0
- checkpoint-2500/model.safetensors +3 -0
- checkpoint-2500/optimizer.pt +3 -0
- checkpoint-2500/rng_state.pth +3 -0
- checkpoint-2500/scheduler.pt +3 -0
- checkpoint-2500/tokenizer.json +3 -0
- checkpoint-2500/tokenizer_config.json +30 -0
- checkpoint-2500/trainer_state.json +534 -0
- checkpoint-2500/training_args.bin +3 -0
- checkpoint-3000/chat_template.jinja +89 -0
- checkpoint-3000/config.json +63 -0
- checkpoint-3000/generation_config.json +12 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,17 @@ 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-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-1500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
checkpoint-2000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
checkpoint-2500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
checkpoint-3000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
checkpoint-3500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
checkpoint-4000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
checkpoint-4500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 44 |
+
checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
checkpoint-5000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
checkpoint-5500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
checkpoint-6000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
checkpoint-6241/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen3-1.7B
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: general_knowledge_model
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- sft
|
| 8 |
+
- trl
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for general_knowledge_model
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B).
|
| 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 |
+
|
| 34 |
+
This model was trained with SFT.
|
| 35 |
+
|
| 36 |
+
### Framework versions
|
| 37 |
+
|
| 38 |
+
- TRL: 1.3.0
|
| 39 |
+
- Transformers: 5.7.0
|
| 40 |
+
- Pytorch: 2.10.0+cu128
|
| 41 |
+
- Datasets: 4.8.5
|
| 42 |
+
- Tokenizers: 0.22.2
|
| 43 |
+
|
| 44 |
+
## Citations
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
Cite TRL as:
|
| 49 |
+
|
| 50 |
+
```bibtex
|
| 51 |
+
@software{vonwerra2020trl,
|
| 52 |
+
title = {{TRL: Transformers Reinforcement Learning}},
|
| 53 |
+
author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
|
| 54 |
+
license = {Apache-2.0},
|
| 55 |
+
url = {https://github.com/huggingface/trl},
|
| 56 |
+
year = {2020}
|
| 57 |
+
}
|
| 58 |
+
```
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-1000/chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 6144,
|
| 15 |
+
"layer_types": [
|
| 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 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention"
|
| 44 |
+
],
|
| 45 |
+
"max_position_embeddings": 40960,
|
| 46 |
+
"max_window_layers": 28,
|
| 47 |
+
"model_type": "qwen3",
|
| 48 |
+
"num_attention_heads": 16,
|
| 49 |
+
"num_hidden_layers": 28,
|
| 50 |
+
"num_key_value_heads": 8,
|
| 51 |
+
"pad_token_id": 151643,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"sliding_window": null,
|
| 58 |
+
"tie_word_embeddings": true,
|
| 59 |
+
"transformers_version": "5.7.0",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-1000/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"temperature": 0.6,
|
| 9 |
+
"top_k": 20,
|
| 10 |
+
"top_p": 0.95,
|
| 11 |
+
"transformers_version": "5.7.0"
|
| 12 |
+
}
|
checkpoint-1000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa6a5ffa325a42521faee189eed67b3922773291e206e209e90b3c770b4d4539
|
| 3 |
+
size 3441185608
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c51b9e4e783c1596ac7d185b6f7589123a938675998d0d32a5f077dfa2dfefd
|
| 3 |
+
size 6882572207
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61c19bab1174704a4a4441475683bf1270277af15d2e2c95e964789128e482c4
|
| 3 |
+
size 14645
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:03766410d0ee20761a88c2327709b696e608f6be1dc35ba4d82dcb2eea2325f7
|
| 3 |
+
size 1465
|
checkpoint-1000/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
checkpoint-1000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": false,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 131072,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null
|
| 30 |
+
}
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,234 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.16024998998437562,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 2.223189059495926,
|
| 14 |
+
"epoch": 0.008012499499218782,
|
| 15 |
+
"grad_norm": 2.875,
|
| 16 |
+
"learning_rate": 1.9842973882390643e-05,
|
| 17 |
+
"loss": 2.3428109741210936,
|
| 18 |
+
"mean_token_accuracy": 0.4963963569700718,
|
| 19 |
+
"num_tokens": 305366.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 2.188130117058754,
|
| 24 |
+
"epoch": 0.016024998998437564,
|
| 25 |
+
"grad_norm": 3.328125,
|
| 26 |
+
"learning_rate": 1.96827431501362e-05,
|
| 27 |
+
"loss": 2.225358734130859,
|
| 28 |
+
"mean_token_accuracy": 0.5110921229422093,
|
| 29 |
+
"num_tokens": 605042.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 2.197347580194473,
|
| 34 |
+
"epoch": 0.024037498497656344,
|
| 35 |
+
"grad_norm": 3.453125,
|
| 36 |
+
"learning_rate": 1.952251241788175e-05,
|
| 37 |
+
"loss": 2.2299420166015627,
|
| 38 |
+
"mean_token_accuracy": 0.5093490639328957,
|
| 39 |
+
"num_tokens": 906888.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 2.1913950395584108,
|
| 44 |
+
"epoch": 0.03204999799687513,
|
| 45 |
+
"grad_norm": 2.90625,
|
| 46 |
+
"learning_rate": 1.9362281685627303e-05,
|
| 47 |
+
"loss": 2.210995941162109,
|
| 48 |
+
"mean_token_accuracy": 0.5097909501194954,
|
| 49 |
+
"num_tokens": 1213125.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 2.1611998677253723,
|
| 54 |
+
"epoch": 0.040062497496093905,
|
| 55 |
+
"grad_norm": 3.25,
|
| 56 |
+
"learning_rate": 1.920205095337286e-05,
|
| 57 |
+
"loss": 2.1867474365234374,
|
| 58 |
+
"mean_token_accuracy": 0.5154410660266876,
|
| 59 |
+
"num_tokens": 1509026.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 2.168051454424858,
|
| 64 |
+
"epoch": 0.04807499699531269,
|
| 65 |
+
"grad_norm": 2.9375,
|
| 66 |
+
"learning_rate": 1.904182022111841e-05,
|
| 67 |
+
"loss": 2.190071716308594,
|
| 68 |
+
"mean_token_accuracy": 0.5156884534657001,
|
| 69 |
+
"num_tokens": 1806850.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 2.1737154543399813,
|
| 74 |
+
"epoch": 0.05608749649453147,
|
| 75 |
+
"grad_norm": 3.421875,
|
| 76 |
+
"learning_rate": 1.8881589488863967e-05,
|
| 77 |
+
"loss": 2.199147186279297,
|
| 78 |
+
"mean_token_accuracy": 0.514045044630766,
|
| 79 |
+
"num_tokens": 2104505.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 2.148039159178734,
|
| 84 |
+
"epoch": 0.06409999599375026,
|
| 85 |
+
"grad_norm": 3.109375,
|
| 86 |
+
"learning_rate": 1.872135875660952e-05,
|
| 87 |
+
"loss": 2.176946105957031,
|
| 88 |
+
"mean_token_accuracy": 0.5177112428843975,
|
| 89 |
+
"num_tokens": 2391088.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 2.194890711903572,
|
| 94 |
+
"epoch": 0.07211249549296903,
|
| 95 |
+
"grad_norm": 3.28125,
|
| 96 |
+
"learning_rate": 1.856112802435507e-05,
|
| 97 |
+
"loss": 2.216288757324219,
|
| 98 |
+
"mean_token_accuracy": 0.5110335703194141,
|
| 99 |
+
"num_tokens": 2691272.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 2.1557786524295808,
|
| 104 |
+
"epoch": 0.08012499499218781,
|
| 105 |
+
"grad_norm": 4.15625,
|
| 106 |
+
"learning_rate": 1.8400897292100627e-05,
|
| 107 |
+
"loss": 2.1812680053710936,
|
| 108 |
+
"mean_token_accuracy": 0.5176697050035,
|
| 109 |
+
"num_tokens": 2991855.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"entropy": 2.1708167713880537,
|
| 114 |
+
"epoch": 0.08813749449140659,
|
| 115 |
+
"grad_norm": 4.34375,
|
| 116 |
+
"learning_rate": 1.824066655984618e-05,
|
| 117 |
+
"loss": 2.1921002197265627,
|
| 118 |
+
"mean_token_accuracy": 0.5156507430970669,
|
| 119 |
+
"num_tokens": 3295049.0,
|
| 120 |
+
"step": 550
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"entropy": 2.1686314886808393,
|
| 124 |
+
"epoch": 0.09614999399062538,
|
| 125 |
+
"grad_norm": 2.875,
|
| 126 |
+
"learning_rate": 1.8080435827591735e-05,
|
| 127 |
+
"loss": 2.1914617919921877,
|
| 128 |
+
"mean_token_accuracy": 0.5139613692462445,
|
| 129 |
+
"num_tokens": 3597206.0,
|
| 130 |
+
"step": 600
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"entropy": 2.15742906332016,
|
| 134 |
+
"epoch": 0.10416249348984416,
|
| 135 |
+
"grad_norm": 3.109375,
|
| 136 |
+
"learning_rate": 1.7920205095337287e-05,
|
| 137 |
+
"loss": 2.174853973388672,
|
| 138 |
+
"mean_token_accuracy": 0.5159518288075924,
|
| 139 |
+
"num_tokens": 3905268.0,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"entropy": 2.162363375425339,
|
| 144 |
+
"epoch": 0.11217499298906294,
|
| 145 |
+
"grad_norm": 3.171875,
|
| 146 |
+
"learning_rate": 1.7759974363082843e-05,
|
| 147 |
+
"loss": 2.184231872558594,
|
| 148 |
+
"mean_token_accuracy": 0.516631413847208,
|
| 149 |
+
"num_tokens": 4207934.0,
|
| 150 |
+
"step": 700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"entropy": 2.1516289287805557,
|
| 154 |
+
"epoch": 0.12018749248828171,
|
| 155 |
+
"grad_norm": 3.609375,
|
| 156 |
+
"learning_rate": 1.7599743630828395e-05,
|
| 157 |
+
"loss": 2.167020263671875,
|
| 158 |
+
"mean_token_accuracy": 0.516358617246151,
|
| 159 |
+
"num_tokens": 4515774.0,
|
| 160 |
+
"step": 750
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"entropy": 2.1549544781446457,
|
| 164 |
+
"epoch": 0.1281999919875005,
|
| 165 |
+
"grad_norm": 3.28125,
|
| 166 |
+
"learning_rate": 1.7439512898573947e-05,
|
| 167 |
+
"loss": 2.181917877197266,
|
| 168 |
+
"mean_token_accuracy": 0.5182220858335495,
|
| 169 |
+
"num_tokens": 4814709.0,
|
| 170 |
+
"step": 800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"entropy": 2.1532615756988527,
|
| 174 |
+
"epoch": 0.1362124914867193,
|
| 175 |
+
"grad_norm": 3.078125,
|
| 176 |
+
"learning_rate": 1.7279282166319503e-05,
|
| 177 |
+
"loss": 2.1659007263183594,
|
| 178 |
+
"mean_token_accuracy": 0.5180425970256328,
|
| 179 |
+
"num_tokens": 5112585.0,
|
| 180 |
+
"step": 850
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"entropy": 2.1704017293453215,
|
| 184 |
+
"epoch": 0.14422499098593805,
|
| 185 |
+
"grad_norm": 2.921875,
|
| 186 |
+
"learning_rate": 1.7119051434065055e-05,
|
| 187 |
+
"loss": 2.1934996032714844,
|
| 188 |
+
"mean_token_accuracy": 0.5154964691400528,
|
| 189 |
+
"num_tokens": 5411369.0,
|
| 190 |
+
"step": 900
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"entropy": 2.1349107217788696,
|
| 194 |
+
"epoch": 0.15223749048515683,
|
| 195 |
+
"grad_norm": 2.984375,
|
| 196 |
+
"learning_rate": 1.6958820701810608e-05,
|
| 197 |
+
"loss": 2.1598545837402345,
|
| 198 |
+
"mean_token_accuracy": 0.5208920259773732,
|
| 199 |
+
"num_tokens": 5707060.0,
|
| 200 |
+
"step": 950
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"entropy": 2.1541705918312073,
|
| 204 |
+
"epoch": 0.16024998998437562,
|
| 205 |
+
"grad_norm": 3.203125,
|
| 206 |
+
"learning_rate": 1.6798589969556163e-05,
|
| 207 |
+
"loss": 2.170956573486328,
|
| 208 |
+
"mean_token_accuracy": 0.5204938448965549,
|
| 209 |
+
"num_tokens": 6004079.0,
|
| 210 |
+
"step": 1000
|
| 211 |
+
}
|
| 212 |
+
],
|
| 213 |
+
"logging_steps": 50,
|
| 214 |
+
"max_steps": 6241,
|
| 215 |
+
"num_input_tokens_seen": 0,
|
| 216 |
+
"num_train_epochs": 1,
|
| 217 |
+
"save_steps": 500,
|
| 218 |
+
"stateful_callbacks": {
|
| 219 |
+
"TrainerControl": {
|
| 220 |
+
"args": {
|
| 221 |
+
"should_epoch_stop": false,
|
| 222 |
+
"should_evaluate": false,
|
| 223 |
+
"should_log": false,
|
| 224 |
+
"should_save": true,
|
| 225 |
+
"should_training_stop": false
|
| 226 |
+
},
|
| 227 |
+
"attributes": {}
|
| 228 |
+
}
|
| 229 |
+
},
|
| 230 |
+
"total_flos": 7.095551986099814e+16,
|
| 231 |
+
"train_batch_size": 4,
|
| 232 |
+
"trial_name": null,
|
| 233 |
+
"trial_params": null
|
| 234 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78622f63304ec0913ed0b4ea7a1b26c08a5621b3661a380d798c4ea426235671
|
| 3 |
+
size 5713
|
checkpoint-1500/chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-1500/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 6144,
|
| 15 |
+
"layer_types": [
|
| 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 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention"
|
| 44 |
+
],
|
| 45 |
+
"max_position_embeddings": 40960,
|
| 46 |
+
"max_window_layers": 28,
|
| 47 |
+
"model_type": "qwen3",
|
| 48 |
+
"num_attention_heads": 16,
|
| 49 |
+
"num_hidden_layers": 28,
|
| 50 |
+
"num_key_value_heads": 8,
|
| 51 |
+
"pad_token_id": 151643,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"sliding_window": null,
|
| 58 |
+
"tie_word_embeddings": true,
|
| 59 |
+
"transformers_version": "5.7.0",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-1500/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"temperature": 0.6,
|
| 9 |
+
"top_k": 20,
|
| 10 |
+
"top_p": 0.95,
|
| 11 |
+
"transformers_version": "5.7.0"
|
| 12 |
+
}
|
checkpoint-1500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:438e34183f5e94f608c7669e0b2cb247069e6a47ebcded96167ec3afb7af8101
|
| 3 |
+
size 3441185608
|
checkpoint-1500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11e561d4c48540c45fb0712b82d5ffdd92c408f08ae4ed12ecbced4e409b1208
|
| 3 |
+
size 6882572207
|
checkpoint-1500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61c19bab1174704a4a4441475683bf1270277af15d2e2c95e964789128e482c4
|
| 3 |
+
size 14645
|
checkpoint-1500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b666bb023db840fb022cfa0357e113ba9d208c4303ad969fafd6098567ba3038
|
| 3 |
+
size 1465
|
checkpoint-1500/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
checkpoint-1500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": false,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 131072,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null
|
| 30 |
+
}
|
checkpoint-1500/trainer_state.json
ADDED
|
@@ -0,0 +1,334 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.24037498497656343,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 2.223189059495926,
|
| 14 |
+
"epoch": 0.008012499499218782,
|
| 15 |
+
"grad_norm": 2.875,
|
| 16 |
+
"learning_rate": 1.9842973882390643e-05,
|
| 17 |
+
"loss": 2.3428109741210936,
|
| 18 |
+
"mean_token_accuracy": 0.4963963569700718,
|
| 19 |
+
"num_tokens": 305366.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 2.188130117058754,
|
| 24 |
+
"epoch": 0.016024998998437564,
|
| 25 |
+
"grad_norm": 3.328125,
|
| 26 |
+
"learning_rate": 1.96827431501362e-05,
|
| 27 |
+
"loss": 2.225358734130859,
|
| 28 |
+
"mean_token_accuracy": 0.5110921229422093,
|
| 29 |
+
"num_tokens": 605042.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 2.197347580194473,
|
| 34 |
+
"epoch": 0.024037498497656344,
|
| 35 |
+
"grad_norm": 3.453125,
|
| 36 |
+
"learning_rate": 1.952251241788175e-05,
|
| 37 |
+
"loss": 2.2299420166015627,
|
| 38 |
+
"mean_token_accuracy": 0.5093490639328957,
|
| 39 |
+
"num_tokens": 906888.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 2.1913950395584108,
|
| 44 |
+
"epoch": 0.03204999799687513,
|
| 45 |
+
"grad_norm": 2.90625,
|
| 46 |
+
"learning_rate": 1.9362281685627303e-05,
|
| 47 |
+
"loss": 2.210995941162109,
|
| 48 |
+
"mean_token_accuracy": 0.5097909501194954,
|
| 49 |
+
"num_tokens": 1213125.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 2.1611998677253723,
|
| 54 |
+
"epoch": 0.040062497496093905,
|
| 55 |
+
"grad_norm": 3.25,
|
| 56 |
+
"learning_rate": 1.920205095337286e-05,
|
| 57 |
+
"loss": 2.1867474365234374,
|
| 58 |
+
"mean_token_accuracy": 0.5154410660266876,
|
| 59 |
+
"num_tokens": 1509026.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 2.168051454424858,
|
| 64 |
+
"epoch": 0.04807499699531269,
|
| 65 |
+
"grad_norm": 2.9375,
|
| 66 |
+
"learning_rate": 1.904182022111841e-05,
|
| 67 |
+
"loss": 2.190071716308594,
|
| 68 |
+
"mean_token_accuracy": 0.5156884534657001,
|
| 69 |
+
"num_tokens": 1806850.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 2.1737154543399813,
|
| 74 |
+
"epoch": 0.05608749649453147,
|
| 75 |
+
"grad_norm": 3.421875,
|
| 76 |
+
"learning_rate": 1.8881589488863967e-05,
|
| 77 |
+
"loss": 2.199147186279297,
|
| 78 |
+
"mean_token_accuracy": 0.514045044630766,
|
| 79 |
+
"num_tokens": 2104505.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 2.148039159178734,
|
| 84 |
+
"epoch": 0.06409999599375026,
|
| 85 |
+
"grad_norm": 3.109375,
|
| 86 |
+
"learning_rate": 1.872135875660952e-05,
|
| 87 |
+
"loss": 2.176946105957031,
|
| 88 |
+
"mean_token_accuracy": 0.5177112428843975,
|
| 89 |
+
"num_tokens": 2391088.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 2.194890711903572,
|
| 94 |
+
"epoch": 0.07211249549296903,
|
| 95 |
+
"grad_norm": 3.28125,
|
| 96 |
+
"learning_rate": 1.856112802435507e-05,
|
| 97 |
+
"loss": 2.216288757324219,
|
| 98 |
+
"mean_token_accuracy": 0.5110335703194141,
|
| 99 |
+
"num_tokens": 2691272.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 2.1557786524295808,
|
| 104 |
+
"epoch": 0.08012499499218781,
|
| 105 |
+
"grad_norm": 4.15625,
|
| 106 |
+
"learning_rate": 1.8400897292100627e-05,
|
| 107 |
+
"loss": 2.1812680053710936,
|
| 108 |
+
"mean_token_accuracy": 0.5176697050035,
|
| 109 |
+
"num_tokens": 2991855.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"entropy": 2.1708167713880537,
|
| 114 |
+
"epoch": 0.08813749449140659,
|
| 115 |
+
"grad_norm": 4.34375,
|
| 116 |
+
"learning_rate": 1.824066655984618e-05,
|
| 117 |
+
"loss": 2.1921002197265627,
|
| 118 |
+
"mean_token_accuracy": 0.5156507430970669,
|
| 119 |
+
"num_tokens": 3295049.0,
|
| 120 |
+
"step": 550
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"entropy": 2.1686314886808393,
|
| 124 |
+
"epoch": 0.09614999399062538,
|
| 125 |
+
"grad_norm": 2.875,
|
| 126 |
+
"learning_rate": 1.8080435827591735e-05,
|
| 127 |
+
"loss": 2.1914617919921877,
|
| 128 |
+
"mean_token_accuracy": 0.5139613692462445,
|
| 129 |
+
"num_tokens": 3597206.0,
|
| 130 |
+
"step": 600
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"entropy": 2.15742906332016,
|
| 134 |
+
"epoch": 0.10416249348984416,
|
| 135 |
+
"grad_norm": 3.109375,
|
| 136 |
+
"learning_rate": 1.7920205095337287e-05,
|
| 137 |
+
"loss": 2.174853973388672,
|
| 138 |
+
"mean_token_accuracy": 0.5159518288075924,
|
| 139 |
+
"num_tokens": 3905268.0,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"entropy": 2.162363375425339,
|
| 144 |
+
"epoch": 0.11217499298906294,
|
| 145 |
+
"grad_norm": 3.171875,
|
| 146 |
+
"learning_rate": 1.7759974363082843e-05,
|
| 147 |
+
"loss": 2.184231872558594,
|
| 148 |
+
"mean_token_accuracy": 0.516631413847208,
|
| 149 |
+
"num_tokens": 4207934.0,
|
| 150 |
+
"step": 700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"entropy": 2.1516289287805557,
|
| 154 |
+
"epoch": 0.12018749248828171,
|
| 155 |
+
"grad_norm": 3.609375,
|
| 156 |
+
"learning_rate": 1.7599743630828395e-05,
|
| 157 |
+
"loss": 2.167020263671875,
|
| 158 |
+
"mean_token_accuracy": 0.516358617246151,
|
| 159 |
+
"num_tokens": 4515774.0,
|
| 160 |
+
"step": 750
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"entropy": 2.1549544781446457,
|
| 164 |
+
"epoch": 0.1281999919875005,
|
| 165 |
+
"grad_norm": 3.28125,
|
| 166 |
+
"learning_rate": 1.7439512898573947e-05,
|
| 167 |
+
"loss": 2.181917877197266,
|
| 168 |
+
"mean_token_accuracy": 0.5182220858335495,
|
| 169 |
+
"num_tokens": 4814709.0,
|
| 170 |
+
"step": 800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"entropy": 2.1532615756988527,
|
| 174 |
+
"epoch": 0.1362124914867193,
|
| 175 |
+
"grad_norm": 3.078125,
|
| 176 |
+
"learning_rate": 1.7279282166319503e-05,
|
| 177 |
+
"loss": 2.1659007263183594,
|
| 178 |
+
"mean_token_accuracy": 0.5180425970256328,
|
| 179 |
+
"num_tokens": 5112585.0,
|
| 180 |
+
"step": 850
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"entropy": 2.1704017293453215,
|
| 184 |
+
"epoch": 0.14422499098593805,
|
| 185 |
+
"grad_norm": 2.921875,
|
| 186 |
+
"learning_rate": 1.7119051434065055e-05,
|
| 187 |
+
"loss": 2.1934996032714844,
|
| 188 |
+
"mean_token_accuracy": 0.5154964691400528,
|
| 189 |
+
"num_tokens": 5411369.0,
|
| 190 |
+
"step": 900
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"entropy": 2.1349107217788696,
|
| 194 |
+
"epoch": 0.15223749048515683,
|
| 195 |
+
"grad_norm": 2.984375,
|
| 196 |
+
"learning_rate": 1.6958820701810608e-05,
|
| 197 |
+
"loss": 2.1598545837402345,
|
| 198 |
+
"mean_token_accuracy": 0.5208920259773732,
|
| 199 |
+
"num_tokens": 5707060.0,
|
| 200 |
+
"step": 950
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"entropy": 2.1541705918312073,
|
| 204 |
+
"epoch": 0.16024998998437562,
|
| 205 |
+
"grad_norm": 3.203125,
|
| 206 |
+
"learning_rate": 1.6798589969556163e-05,
|
| 207 |
+
"loss": 2.170956573486328,
|
| 208 |
+
"mean_token_accuracy": 0.5204938448965549,
|
| 209 |
+
"num_tokens": 6004079.0,
|
| 210 |
+
"step": 1000
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"entropy": 2.1523867017030716,
|
| 214 |
+
"epoch": 0.1682624894835944,
|
| 215 |
+
"grad_norm": 3.09375,
|
| 216 |
+
"learning_rate": 1.6638359237301716e-05,
|
| 217 |
+
"loss": 2.1695118713378907,
|
| 218 |
+
"mean_token_accuracy": 0.5177376234531402,
|
| 219 |
+
"num_tokens": 6315462.0,
|
| 220 |
+
"step": 1050
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"entropy": 2.1356022965908052,
|
| 224 |
+
"epoch": 0.17627498898281319,
|
| 225 |
+
"grad_norm": 3.21875,
|
| 226 |
+
"learning_rate": 1.6478128505047268e-05,
|
| 227 |
+
"loss": 2.1513107299804686,
|
| 228 |
+
"mean_token_accuracy": 0.5217180585861206,
|
| 229 |
+
"num_tokens": 6613060.0,
|
| 230 |
+
"step": 1100
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"entropy": 2.1483460652828215,
|
| 234 |
+
"epoch": 0.18428748848203197,
|
| 235 |
+
"grad_norm": 3.21875,
|
| 236 |
+
"learning_rate": 1.6317897772792824e-05,
|
| 237 |
+
"loss": 2.1671595764160156,
|
| 238 |
+
"mean_token_accuracy": 0.5199222292006016,
|
| 239 |
+
"num_tokens": 6909655.0,
|
| 240 |
+
"step": 1150
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"entropy": 2.1074040442705155,
|
| 244 |
+
"epoch": 0.19229998798125075,
|
| 245 |
+
"grad_norm": 3.828125,
|
| 246 |
+
"learning_rate": 1.6157667040538376e-05,
|
| 247 |
+
"loss": 2.12770751953125,
|
| 248 |
+
"mean_token_accuracy": 0.5252153894305229,
|
| 249 |
+
"num_tokens": 7203263.0,
|
| 250 |
+
"step": 1200
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"entropy": 2.1354027593135836,
|
| 254 |
+
"epoch": 0.20031248748046954,
|
| 255 |
+
"grad_norm": 2.921875,
|
| 256 |
+
"learning_rate": 1.5997436308283932e-05,
|
| 257 |
+
"loss": 2.156183166503906,
|
| 258 |
+
"mean_token_accuracy": 0.518907299786806,
|
| 259 |
+
"num_tokens": 7505411.0,
|
| 260 |
+
"step": 1250
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"entropy": 2.142924894094467,
|
| 264 |
+
"epoch": 0.20832498697968832,
|
| 265 |
+
"grad_norm": 3.5625,
|
| 266 |
+
"learning_rate": 1.5837205576029484e-05,
|
| 267 |
+
"loss": 2.158486633300781,
|
| 268 |
+
"mean_token_accuracy": 0.5205909128487111,
|
| 269 |
+
"num_tokens": 7803562.0,
|
| 270 |
+
"step": 1300
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"entropy": 2.1305291467905043,
|
| 274 |
+
"epoch": 0.2163374864789071,
|
| 275 |
+
"grad_norm": 3.34375,
|
| 276 |
+
"learning_rate": 1.5676974843775036e-05,
|
| 277 |
+
"loss": 2.151043701171875,
|
| 278 |
+
"mean_token_accuracy": 0.5211617235839366,
|
| 279 |
+
"num_tokens": 8108037.0,
|
| 280 |
+
"step": 1350
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"entropy": 2.160947166085243,
|
| 284 |
+
"epoch": 0.2243499859781259,
|
| 285 |
+
"grad_norm": 3.234375,
|
| 286 |
+
"learning_rate": 1.5516744111520592e-05,
|
| 287 |
+
"loss": 2.177792663574219,
|
| 288 |
+
"mean_token_accuracy": 0.5167004990577698,
|
| 289 |
+
"num_tokens": 8408674.0,
|
| 290 |
+
"step": 1400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"entropy": 2.120403967499733,
|
| 294 |
+
"epoch": 0.23236248547734467,
|
| 295 |
+
"grad_norm": 3.375,
|
| 296 |
+
"learning_rate": 1.5356513379266144e-05,
|
| 297 |
+
"loss": 2.1462562561035154,
|
| 298 |
+
"mean_token_accuracy": 0.5225985661149025,
|
| 299 |
+
"num_tokens": 8707650.0,
|
| 300 |
+
"step": 1450
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"entropy": 2.138882914185524,
|
| 304 |
+
"epoch": 0.24037498497656343,
|
| 305 |
+
"grad_norm": 3.171875,
|
| 306 |
+
"learning_rate": 1.5196282647011698e-05,
|
| 307 |
+
"loss": 2.1499905395507812,
|
| 308 |
+
"mean_token_accuracy": 0.5204668046534061,
|
| 309 |
+
"num_tokens": 9006296.0,
|
| 310 |
+
"step": 1500
|
| 311 |
+
}
|
| 312 |
+
],
|
| 313 |
+
"logging_steps": 50,
|
| 314 |
+
"max_steps": 6241,
|
| 315 |
+
"num_input_tokens_seen": 0,
|
| 316 |
+
"num_train_epochs": 1,
|
| 317 |
+
"save_steps": 500,
|
| 318 |
+
"stateful_callbacks": {
|
| 319 |
+
"TrainerControl": {
|
| 320 |
+
"args": {
|
| 321 |
+
"should_epoch_stop": false,
|
| 322 |
+
"should_evaluate": false,
|
| 323 |
+
"should_log": false,
|
| 324 |
+
"should_save": true,
|
| 325 |
+
"should_training_stop": false
|
| 326 |
+
},
|
| 327 |
+
"attributes": {}
|
| 328 |
+
}
|
| 329 |
+
},
|
| 330 |
+
"total_flos": 1.064051705174999e+17,
|
| 331 |
+
"train_batch_size": 4,
|
| 332 |
+
"trial_name": null,
|
| 333 |
+
"trial_params": null
|
| 334 |
+
}
|
checkpoint-1500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78622f63304ec0913ed0b4ea7a1b26c08a5621b3661a380d798c4ea426235671
|
| 3 |
+
size 5713
|
checkpoint-2000/chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-2000/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 6144,
|
| 15 |
+
"layer_types": [
|
| 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 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention"
|
| 44 |
+
],
|
| 45 |
+
"max_position_embeddings": 40960,
|
| 46 |
+
"max_window_layers": 28,
|
| 47 |
+
"model_type": "qwen3",
|
| 48 |
+
"num_attention_heads": 16,
|
| 49 |
+
"num_hidden_layers": 28,
|
| 50 |
+
"num_key_value_heads": 8,
|
| 51 |
+
"pad_token_id": 151643,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"sliding_window": null,
|
| 58 |
+
"tie_word_embeddings": true,
|
| 59 |
+
"transformers_version": "5.7.0",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-2000/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"temperature": 0.6,
|
| 9 |
+
"top_k": 20,
|
| 10 |
+
"top_p": 0.95,
|
| 11 |
+
"transformers_version": "5.7.0"
|
| 12 |
+
}
|
checkpoint-2000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec88ee383178456b8f8ab7ee942be2785b0a32656809f229cae94c76dff1b810
|
| 3 |
+
size 3441185608
|
checkpoint-2000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea2dc818dc38c52000eeebc5b101562caf8fb11d4790f976bb066ee636e4f14d
|
| 3 |
+
size 6882572207
|
checkpoint-2000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61c19bab1174704a4a4441475683bf1270277af15d2e2c95e964789128e482c4
|
| 3 |
+
size 14645
|
checkpoint-2000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11b128c7c709b8b18f78ca21ff87ee25957a3215823c4227b4ba298165c8d222
|
| 3 |
+
size 1465
|
checkpoint-2000/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
checkpoint-2000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": false,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 131072,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null
|
| 30 |
+
}
|
checkpoint-2000/trainer_state.json
ADDED
|
@@ -0,0 +1,434 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.32049997996875124,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 2.223189059495926,
|
| 14 |
+
"epoch": 0.008012499499218782,
|
| 15 |
+
"grad_norm": 2.875,
|
| 16 |
+
"learning_rate": 1.9842973882390643e-05,
|
| 17 |
+
"loss": 2.3428109741210936,
|
| 18 |
+
"mean_token_accuracy": 0.4963963569700718,
|
| 19 |
+
"num_tokens": 305366.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 2.188130117058754,
|
| 24 |
+
"epoch": 0.016024998998437564,
|
| 25 |
+
"grad_norm": 3.328125,
|
| 26 |
+
"learning_rate": 1.96827431501362e-05,
|
| 27 |
+
"loss": 2.225358734130859,
|
| 28 |
+
"mean_token_accuracy": 0.5110921229422093,
|
| 29 |
+
"num_tokens": 605042.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 2.197347580194473,
|
| 34 |
+
"epoch": 0.024037498497656344,
|
| 35 |
+
"grad_norm": 3.453125,
|
| 36 |
+
"learning_rate": 1.952251241788175e-05,
|
| 37 |
+
"loss": 2.2299420166015627,
|
| 38 |
+
"mean_token_accuracy": 0.5093490639328957,
|
| 39 |
+
"num_tokens": 906888.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 2.1913950395584108,
|
| 44 |
+
"epoch": 0.03204999799687513,
|
| 45 |
+
"grad_norm": 2.90625,
|
| 46 |
+
"learning_rate": 1.9362281685627303e-05,
|
| 47 |
+
"loss": 2.210995941162109,
|
| 48 |
+
"mean_token_accuracy": 0.5097909501194954,
|
| 49 |
+
"num_tokens": 1213125.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 2.1611998677253723,
|
| 54 |
+
"epoch": 0.040062497496093905,
|
| 55 |
+
"grad_norm": 3.25,
|
| 56 |
+
"learning_rate": 1.920205095337286e-05,
|
| 57 |
+
"loss": 2.1867474365234374,
|
| 58 |
+
"mean_token_accuracy": 0.5154410660266876,
|
| 59 |
+
"num_tokens": 1509026.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 2.168051454424858,
|
| 64 |
+
"epoch": 0.04807499699531269,
|
| 65 |
+
"grad_norm": 2.9375,
|
| 66 |
+
"learning_rate": 1.904182022111841e-05,
|
| 67 |
+
"loss": 2.190071716308594,
|
| 68 |
+
"mean_token_accuracy": 0.5156884534657001,
|
| 69 |
+
"num_tokens": 1806850.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 2.1737154543399813,
|
| 74 |
+
"epoch": 0.05608749649453147,
|
| 75 |
+
"grad_norm": 3.421875,
|
| 76 |
+
"learning_rate": 1.8881589488863967e-05,
|
| 77 |
+
"loss": 2.199147186279297,
|
| 78 |
+
"mean_token_accuracy": 0.514045044630766,
|
| 79 |
+
"num_tokens": 2104505.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 2.148039159178734,
|
| 84 |
+
"epoch": 0.06409999599375026,
|
| 85 |
+
"grad_norm": 3.109375,
|
| 86 |
+
"learning_rate": 1.872135875660952e-05,
|
| 87 |
+
"loss": 2.176946105957031,
|
| 88 |
+
"mean_token_accuracy": 0.5177112428843975,
|
| 89 |
+
"num_tokens": 2391088.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 2.194890711903572,
|
| 94 |
+
"epoch": 0.07211249549296903,
|
| 95 |
+
"grad_norm": 3.28125,
|
| 96 |
+
"learning_rate": 1.856112802435507e-05,
|
| 97 |
+
"loss": 2.216288757324219,
|
| 98 |
+
"mean_token_accuracy": 0.5110335703194141,
|
| 99 |
+
"num_tokens": 2691272.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 2.1557786524295808,
|
| 104 |
+
"epoch": 0.08012499499218781,
|
| 105 |
+
"grad_norm": 4.15625,
|
| 106 |
+
"learning_rate": 1.8400897292100627e-05,
|
| 107 |
+
"loss": 2.1812680053710936,
|
| 108 |
+
"mean_token_accuracy": 0.5176697050035,
|
| 109 |
+
"num_tokens": 2991855.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"entropy": 2.1708167713880537,
|
| 114 |
+
"epoch": 0.08813749449140659,
|
| 115 |
+
"grad_norm": 4.34375,
|
| 116 |
+
"learning_rate": 1.824066655984618e-05,
|
| 117 |
+
"loss": 2.1921002197265627,
|
| 118 |
+
"mean_token_accuracy": 0.5156507430970669,
|
| 119 |
+
"num_tokens": 3295049.0,
|
| 120 |
+
"step": 550
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"entropy": 2.1686314886808393,
|
| 124 |
+
"epoch": 0.09614999399062538,
|
| 125 |
+
"grad_norm": 2.875,
|
| 126 |
+
"learning_rate": 1.8080435827591735e-05,
|
| 127 |
+
"loss": 2.1914617919921877,
|
| 128 |
+
"mean_token_accuracy": 0.5139613692462445,
|
| 129 |
+
"num_tokens": 3597206.0,
|
| 130 |
+
"step": 600
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"entropy": 2.15742906332016,
|
| 134 |
+
"epoch": 0.10416249348984416,
|
| 135 |
+
"grad_norm": 3.109375,
|
| 136 |
+
"learning_rate": 1.7920205095337287e-05,
|
| 137 |
+
"loss": 2.174853973388672,
|
| 138 |
+
"mean_token_accuracy": 0.5159518288075924,
|
| 139 |
+
"num_tokens": 3905268.0,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"entropy": 2.162363375425339,
|
| 144 |
+
"epoch": 0.11217499298906294,
|
| 145 |
+
"grad_norm": 3.171875,
|
| 146 |
+
"learning_rate": 1.7759974363082843e-05,
|
| 147 |
+
"loss": 2.184231872558594,
|
| 148 |
+
"mean_token_accuracy": 0.516631413847208,
|
| 149 |
+
"num_tokens": 4207934.0,
|
| 150 |
+
"step": 700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"entropy": 2.1516289287805557,
|
| 154 |
+
"epoch": 0.12018749248828171,
|
| 155 |
+
"grad_norm": 3.609375,
|
| 156 |
+
"learning_rate": 1.7599743630828395e-05,
|
| 157 |
+
"loss": 2.167020263671875,
|
| 158 |
+
"mean_token_accuracy": 0.516358617246151,
|
| 159 |
+
"num_tokens": 4515774.0,
|
| 160 |
+
"step": 750
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"entropy": 2.1549544781446457,
|
| 164 |
+
"epoch": 0.1281999919875005,
|
| 165 |
+
"grad_norm": 3.28125,
|
| 166 |
+
"learning_rate": 1.7439512898573947e-05,
|
| 167 |
+
"loss": 2.181917877197266,
|
| 168 |
+
"mean_token_accuracy": 0.5182220858335495,
|
| 169 |
+
"num_tokens": 4814709.0,
|
| 170 |
+
"step": 800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"entropy": 2.1532615756988527,
|
| 174 |
+
"epoch": 0.1362124914867193,
|
| 175 |
+
"grad_norm": 3.078125,
|
| 176 |
+
"learning_rate": 1.7279282166319503e-05,
|
| 177 |
+
"loss": 2.1659007263183594,
|
| 178 |
+
"mean_token_accuracy": 0.5180425970256328,
|
| 179 |
+
"num_tokens": 5112585.0,
|
| 180 |
+
"step": 850
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"entropy": 2.1704017293453215,
|
| 184 |
+
"epoch": 0.14422499098593805,
|
| 185 |
+
"grad_norm": 2.921875,
|
| 186 |
+
"learning_rate": 1.7119051434065055e-05,
|
| 187 |
+
"loss": 2.1934996032714844,
|
| 188 |
+
"mean_token_accuracy": 0.5154964691400528,
|
| 189 |
+
"num_tokens": 5411369.0,
|
| 190 |
+
"step": 900
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"entropy": 2.1349107217788696,
|
| 194 |
+
"epoch": 0.15223749048515683,
|
| 195 |
+
"grad_norm": 2.984375,
|
| 196 |
+
"learning_rate": 1.6958820701810608e-05,
|
| 197 |
+
"loss": 2.1598545837402345,
|
| 198 |
+
"mean_token_accuracy": 0.5208920259773732,
|
| 199 |
+
"num_tokens": 5707060.0,
|
| 200 |
+
"step": 950
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"entropy": 2.1541705918312073,
|
| 204 |
+
"epoch": 0.16024998998437562,
|
| 205 |
+
"grad_norm": 3.203125,
|
| 206 |
+
"learning_rate": 1.6798589969556163e-05,
|
| 207 |
+
"loss": 2.170956573486328,
|
| 208 |
+
"mean_token_accuracy": 0.5204938448965549,
|
| 209 |
+
"num_tokens": 6004079.0,
|
| 210 |
+
"step": 1000
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"entropy": 2.1523867017030716,
|
| 214 |
+
"epoch": 0.1682624894835944,
|
| 215 |
+
"grad_norm": 3.09375,
|
| 216 |
+
"learning_rate": 1.6638359237301716e-05,
|
| 217 |
+
"loss": 2.1695118713378907,
|
| 218 |
+
"mean_token_accuracy": 0.5177376234531402,
|
| 219 |
+
"num_tokens": 6315462.0,
|
| 220 |
+
"step": 1050
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"entropy": 2.1356022965908052,
|
| 224 |
+
"epoch": 0.17627498898281319,
|
| 225 |
+
"grad_norm": 3.21875,
|
| 226 |
+
"learning_rate": 1.6478128505047268e-05,
|
| 227 |
+
"loss": 2.1513107299804686,
|
| 228 |
+
"mean_token_accuracy": 0.5217180585861206,
|
| 229 |
+
"num_tokens": 6613060.0,
|
| 230 |
+
"step": 1100
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"entropy": 2.1483460652828215,
|
| 234 |
+
"epoch": 0.18428748848203197,
|
| 235 |
+
"grad_norm": 3.21875,
|
| 236 |
+
"learning_rate": 1.6317897772792824e-05,
|
| 237 |
+
"loss": 2.1671595764160156,
|
| 238 |
+
"mean_token_accuracy": 0.5199222292006016,
|
| 239 |
+
"num_tokens": 6909655.0,
|
| 240 |
+
"step": 1150
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"entropy": 2.1074040442705155,
|
| 244 |
+
"epoch": 0.19229998798125075,
|
| 245 |
+
"grad_norm": 3.828125,
|
| 246 |
+
"learning_rate": 1.6157667040538376e-05,
|
| 247 |
+
"loss": 2.12770751953125,
|
| 248 |
+
"mean_token_accuracy": 0.5252153894305229,
|
| 249 |
+
"num_tokens": 7203263.0,
|
| 250 |
+
"step": 1200
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"entropy": 2.1354027593135836,
|
| 254 |
+
"epoch": 0.20031248748046954,
|
| 255 |
+
"grad_norm": 2.921875,
|
| 256 |
+
"learning_rate": 1.5997436308283932e-05,
|
| 257 |
+
"loss": 2.156183166503906,
|
| 258 |
+
"mean_token_accuracy": 0.518907299786806,
|
| 259 |
+
"num_tokens": 7505411.0,
|
| 260 |
+
"step": 1250
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"entropy": 2.142924894094467,
|
| 264 |
+
"epoch": 0.20832498697968832,
|
| 265 |
+
"grad_norm": 3.5625,
|
| 266 |
+
"learning_rate": 1.5837205576029484e-05,
|
| 267 |
+
"loss": 2.158486633300781,
|
| 268 |
+
"mean_token_accuracy": 0.5205909128487111,
|
| 269 |
+
"num_tokens": 7803562.0,
|
| 270 |
+
"step": 1300
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"entropy": 2.1305291467905043,
|
| 274 |
+
"epoch": 0.2163374864789071,
|
| 275 |
+
"grad_norm": 3.34375,
|
| 276 |
+
"learning_rate": 1.5676974843775036e-05,
|
| 277 |
+
"loss": 2.151043701171875,
|
| 278 |
+
"mean_token_accuracy": 0.5211617235839366,
|
| 279 |
+
"num_tokens": 8108037.0,
|
| 280 |
+
"step": 1350
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"entropy": 2.160947166085243,
|
| 284 |
+
"epoch": 0.2243499859781259,
|
| 285 |
+
"grad_norm": 3.234375,
|
| 286 |
+
"learning_rate": 1.5516744111520592e-05,
|
| 287 |
+
"loss": 2.177792663574219,
|
| 288 |
+
"mean_token_accuracy": 0.5167004990577698,
|
| 289 |
+
"num_tokens": 8408674.0,
|
| 290 |
+
"step": 1400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"entropy": 2.120403967499733,
|
| 294 |
+
"epoch": 0.23236248547734467,
|
| 295 |
+
"grad_norm": 3.375,
|
| 296 |
+
"learning_rate": 1.5356513379266144e-05,
|
| 297 |
+
"loss": 2.1462562561035154,
|
| 298 |
+
"mean_token_accuracy": 0.5225985661149025,
|
| 299 |
+
"num_tokens": 8707650.0,
|
| 300 |
+
"step": 1450
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"entropy": 2.138882914185524,
|
| 304 |
+
"epoch": 0.24037498497656343,
|
| 305 |
+
"grad_norm": 3.171875,
|
| 306 |
+
"learning_rate": 1.5196282647011698e-05,
|
| 307 |
+
"loss": 2.1499905395507812,
|
| 308 |
+
"mean_token_accuracy": 0.5204668046534061,
|
| 309 |
+
"num_tokens": 9006296.0,
|
| 310 |
+
"step": 1500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"entropy": 2.131927245259285,
|
| 314 |
+
"epoch": 0.2483874844757822,
|
| 315 |
+
"grad_norm": 3.546875,
|
| 316 |
+
"learning_rate": 1.5036051914757252e-05,
|
| 317 |
+
"loss": 2.157076416015625,
|
| 318 |
+
"mean_token_accuracy": 0.5220408822596073,
|
| 319 |
+
"num_tokens": 9307203.0,
|
| 320 |
+
"step": 1550
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"entropy": 2.1308567851781843,
|
| 324 |
+
"epoch": 0.256399983975001,
|
| 325 |
+
"grad_norm": 3.546875,
|
| 326 |
+
"learning_rate": 1.4875821182502805e-05,
|
| 327 |
+
"loss": 2.146601104736328,
|
| 328 |
+
"mean_token_accuracy": 0.5219539681077003,
|
| 329 |
+
"num_tokens": 9610070.0,
|
| 330 |
+
"step": 1600
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"entropy": 2.1097783839702604,
|
| 334 |
+
"epoch": 0.2644124834742198,
|
| 335 |
+
"grad_norm": 3.84375,
|
| 336 |
+
"learning_rate": 1.4715590450248359e-05,
|
| 337 |
+
"loss": 2.1268209838867187,
|
| 338 |
+
"mean_token_accuracy": 0.5236652019619942,
|
| 339 |
+
"num_tokens": 9909117.0,
|
| 340 |
+
"step": 1650
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"entropy": 2.123499180674553,
|
| 344 |
+
"epoch": 0.2724249829734386,
|
| 345 |
+
"grad_norm": 3.46875,
|
| 346 |
+
"learning_rate": 1.4555359717993913e-05,
|
| 347 |
+
"loss": 2.1409934997558593,
|
| 348 |
+
"mean_token_accuracy": 0.5225465354323388,
|
| 349 |
+
"num_tokens": 10209461.0,
|
| 350 |
+
"step": 1700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"entropy": 2.13052960395813,
|
| 354 |
+
"epoch": 0.2804374824726573,
|
| 355 |
+
"grad_norm": 2.71875,
|
| 356 |
+
"learning_rate": 1.4395128985739467e-05,
|
| 357 |
+
"loss": 2.151373291015625,
|
| 358 |
+
"mean_token_accuracy": 0.5227928113937378,
|
| 359 |
+
"num_tokens": 10509282.0,
|
| 360 |
+
"step": 1750
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"entropy": 2.1226114094257356,
|
| 364 |
+
"epoch": 0.2884499819718761,
|
| 365 |
+
"grad_norm": 3.65625,
|
| 366 |
+
"learning_rate": 1.4234898253485019e-05,
|
| 367 |
+
"loss": 2.144451141357422,
|
| 368 |
+
"mean_token_accuracy": 0.5233827362954616,
|
| 369 |
+
"num_tokens": 10811056.0,
|
| 370 |
+
"step": 1800
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"entropy": 2.126143154501915,
|
| 374 |
+
"epoch": 0.2964624814710949,
|
| 375 |
+
"grad_norm": 3.359375,
|
| 376 |
+
"learning_rate": 1.4074667521230575e-05,
|
| 377 |
+
"loss": 2.1457598876953123,
|
| 378 |
+
"mean_token_accuracy": 0.5221967925131321,
|
| 379 |
+
"num_tokens": 11111153.0,
|
| 380 |
+
"step": 1850
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"entropy": 2.1048326581716537,
|
| 384 |
+
"epoch": 0.30447498097031367,
|
| 385 |
+
"grad_norm": 3.59375,
|
| 386 |
+
"learning_rate": 1.3914436788976127e-05,
|
| 387 |
+
"loss": 2.1260116577148436,
|
| 388 |
+
"mean_token_accuracy": 0.5272719909250736,
|
| 389 |
+
"num_tokens": 11405606.0,
|
| 390 |
+
"step": 1900
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"entropy": 2.1236617982387545,
|
| 394 |
+
"epoch": 0.31248748046953245,
|
| 395 |
+
"grad_norm": 3.0,
|
| 396 |
+
"learning_rate": 1.375420605672168e-05,
|
| 397 |
+
"loss": 2.151782684326172,
|
| 398 |
+
"mean_token_accuracy": 0.5227546173334122,
|
| 399 |
+
"num_tokens": 11703943.0,
|
| 400 |
+
"step": 1950
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"entropy": 2.121563429236412,
|
| 404 |
+
"epoch": 0.32049997996875124,
|
| 405 |
+
"grad_norm": 3.34375,
|
| 406 |
+
"learning_rate": 1.3593975324467235e-05,
|
| 407 |
+
"loss": 2.1370144653320313,
|
| 408 |
+
"mean_token_accuracy": 0.5235123547911644,
|
| 409 |
+
"num_tokens": 12002522.0,
|
| 410 |
+
"step": 2000
|
| 411 |
+
}
|
| 412 |
+
],
|
| 413 |
+
"logging_steps": 50,
|
| 414 |
+
"max_steps": 6241,
|
| 415 |
+
"num_input_tokens_seen": 0,
|
| 416 |
+
"num_train_epochs": 1,
|
| 417 |
+
"save_steps": 500,
|
| 418 |
+
"stateful_callbacks": {
|
| 419 |
+
"TrainerControl": {
|
| 420 |
+
"args": {
|
| 421 |
+
"should_epoch_stop": false,
|
| 422 |
+
"should_evaluate": false,
|
| 423 |
+
"should_log": false,
|
| 424 |
+
"should_save": true,
|
| 425 |
+
"should_training_stop": false
|
| 426 |
+
},
|
| 427 |
+
"attributes": {}
|
| 428 |
+
}
|
| 429 |
+
},
|
| 430 |
+
"total_flos": 1.4174735647666176e+17,
|
| 431 |
+
"train_batch_size": 4,
|
| 432 |
+
"trial_name": null,
|
| 433 |
+
"trial_params": null
|
| 434 |
+
}
|
checkpoint-2000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78622f63304ec0913ed0b4ea7a1b26c08a5621b3661a380d798c4ea426235671
|
| 3 |
+
size 5713
|
checkpoint-2500/chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-2500/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 6144,
|
| 15 |
+
"layer_types": [
|
| 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 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention"
|
| 44 |
+
],
|
| 45 |
+
"max_position_embeddings": 40960,
|
| 46 |
+
"max_window_layers": 28,
|
| 47 |
+
"model_type": "qwen3",
|
| 48 |
+
"num_attention_heads": 16,
|
| 49 |
+
"num_hidden_layers": 28,
|
| 50 |
+
"num_key_value_heads": 8,
|
| 51 |
+
"pad_token_id": 151643,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"sliding_window": null,
|
| 58 |
+
"tie_word_embeddings": true,
|
| 59 |
+
"transformers_version": "5.7.0",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-2500/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"temperature": 0.6,
|
| 9 |
+
"top_k": 20,
|
| 10 |
+
"top_p": 0.95,
|
| 11 |
+
"transformers_version": "5.7.0"
|
| 12 |
+
}
|
checkpoint-2500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98a04b91a11ec1c0b526f91d12bdfb28ff4c3c69e104271b9e28a2f75f44f891
|
| 3 |
+
size 3441185608
|
checkpoint-2500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7535b338ca59a9c376634a6fbbef04a836c5c21962e653fe7794832854643965
|
| 3 |
+
size 6882572207
|
checkpoint-2500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61c19bab1174704a4a4441475683bf1270277af15d2e2c95e964789128e482c4
|
| 3 |
+
size 14645
|
checkpoint-2500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c863cc66645dc4964c4c4f2cf30406fef13cbea85c7c7bd73c333d72670c821e
|
| 3 |
+
size 1465
|
checkpoint-2500/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
checkpoint-2500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": false,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 131072,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null
|
| 30 |
+
}
|
checkpoint-2500/trainer_state.json
ADDED
|
@@ -0,0 +1,534 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.4006249749609391,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 2.223189059495926,
|
| 14 |
+
"epoch": 0.008012499499218782,
|
| 15 |
+
"grad_norm": 2.875,
|
| 16 |
+
"learning_rate": 1.9842973882390643e-05,
|
| 17 |
+
"loss": 2.3428109741210936,
|
| 18 |
+
"mean_token_accuracy": 0.4963963569700718,
|
| 19 |
+
"num_tokens": 305366.0,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 2.188130117058754,
|
| 24 |
+
"epoch": 0.016024998998437564,
|
| 25 |
+
"grad_norm": 3.328125,
|
| 26 |
+
"learning_rate": 1.96827431501362e-05,
|
| 27 |
+
"loss": 2.225358734130859,
|
| 28 |
+
"mean_token_accuracy": 0.5110921229422093,
|
| 29 |
+
"num_tokens": 605042.0,
|
| 30 |
+
"step": 100
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 2.197347580194473,
|
| 34 |
+
"epoch": 0.024037498497656344,
|
| 35 |
+
"grad_norm": 3.453125,
|
| 36 |
+
"learning_rate": 1.952251241788175e-05,
|
| 37 |
+
"loss": 2.2299420166015627,
|
| 38 |
+
"mean_token_accuracy": 0.5093490639328957,
|
| 39 |
+
"num_tokens": 906888.0,
|
| 40 |
+
"step": 150
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 2.1913950395584108,
|
| 44 |
+
"epoch": 0.03204999799687513,
|
| 45 |
+
"grad_norm": 2.90625,
|
| 46 |
+
"learning_rate": 1.9362281685627303e-05,
|
| 47 |
+
"loss": 2.210995941162109,
|
| 48 |
+
"mean_token_accuracy": 0.5097909501194954,
|
| 49 |
+
"num_tokens": 1213125.0,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 2.1611998677253723,
|
| 54 |
+
"epoch": 0.040062497496093905,
|
| 55 |
+
"grad_norm": 3.25,
|
| 56 |
+
"learning_rate": 1.920205095337286e-05,
|
| 57 |
+
"loss": 2.1867474365234374,
|
| 58 |
+
"mean_token_accuracy": 0.5154410660266876,
|
| 59 |
+
"num_tokens": 1509026.0,
|
| 60 |
+
"step": 250
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 2.168051454424858,
|
| 64 |
+
"epoch": 0.04807499699531269,
|
| 65 |
+
"grad_norm": 2.9375,
|
| 66 |
+
"learning_rate": 1.904182022111841e-05,
|
| 67 |
+
"loss": 2.190071716308594,
|
| 68 |
+
"mean_token_accuracy": 0.5156884534657001,
|
| 69 |
+
"num_tokens": 1806850.0,
|
| 70 |
+
"step": 300
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 2.1737154543399813,
|
| 74 |
+
"epoch": 0.05608749649453147,
|
| 75 |
+
"grad_norm": 3.421875,
|
| 76 |
+
"learning_rate": 1.8881589488863967e-05,
|
| 77 |
+
"loss": 2.199147186279297,
|
| 78 |
+
"mean_token_accuracy": 0.514045044630766,
|
| 79 |
+
"num_tokens": 2104505.0,
|
| 80 |
+
"step": 350
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 2.148039159178734,
|
| 84 |
+
"epoch": 0.06409999599375026,
|
| 85 |
+
"grad_norm": 3.109375,
|
| 86 |
+
"learning_rate": 1.872135875660952e-05,
|
| 87 |
+
"loss": 2.176946105957031,
|
| 88 |
+
"mean_token_accuracy": 0.5177112428843975,
|
| 89 |
+
"num_tokens": 2391088.0,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 2.194890711903572,
|
| 94 |
+
"epoch": 0.07211249549296903,
|
| 95 |
+
"grad_norm": 3.28125,
|
| 96 |
+
"learning_rate": 1.856112802435507e-05,
|
| 97 |
+
"loss": 2.216288757324219,
|
| 98 |
+
"mean_token_accuracy": 0.5110335703194141,
|
| 99 |
+
"num_tokens": 2691272.0,
|
| 100 |
+
"step": 450
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 2.1557786524295808,
|
| 104 |
+
"epoch": 0.08012499499218781,
|
| 105 |
+
"grad_norm": 4.15625,
|
| 106 |
+
"learning_rate": 1.8400897292100627e-05,
|
| 107 |
+
"loss": 2.1812680053710936,
|
| 108 |
+
"mean_token_accuracy": 0.5176697050035,
|
| 109 |
+
"num_tokens": 2991855.0,
|
| 110 |
+
"step": 500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"entropy": 2.1708167713880537,
|
| 114 |
+
"epoch": 0.08813749449140659,
|
| 115 |
+
"grad_norm": 4.34375,
|
| 116 |
+
"learning_rate": 1.824066655984618e-05,
|
| 117 |
+
"loss": 2.1921002197265627,
|
| 118 |
+
"mean_token_accuracy": 0.5156507430970669,
|
| 119 |
+
"num_tokens": 3295049.0,
|
| 120 |
+
"step": 550
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"entropy": 2.1686314886808393,
|
| 124 |
+
"epoch": 0.09614999399062538,
|
| 125 |
+
"grad_norm": 2.875,
|
| 126 |
+
"learning_rate": 1.8080435827591735e-05,
|
| 127 |
+
"loss": 2.1914617919921877,
|
| 128 |
+
"mean_token_accuracy": 0.5139613692462445,
|
| 129 |
+
"num_tokens": 3597206.0,
|
| 130 |
+
"step": 600
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"entropy": 2.15742906332016,
|
| 134 |
+
"epoch": 0.10416249348984416,
|
| 135 |
+
"grad_norm": 3.109375,
|
| 136 |
+
"learning_rate": 1.7920205095337287e-05,
|
| 137 |
+
"loss": 2.174853973388672,
|
| 138 |
+
"mean_token_accuracy": 0.5159518288075924,
|
| 139 |
+
"num_tokens": 3905268.0,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"entropy": 2.162363375425339,
|
| 144 |
+
"epoch": 0.11217499298906294,
|
| 145 |
+
"grad_norm": 3.171875,
|
| 146 |
+
"learning_rate": 1.7759974363082843e-05,
|
| 147 |
+
"loss": 2.184231872558594,
|
| 148 |
+
"mean_token_accuracy": 0.516631413847208,
|
| 149 |
+
"num_tokens": 4207934.0,
|
| 150 |
+
"step": 700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"entropy": 2.1516289287805557,
|
| 154 |
+
"epoch": 0.12018749248828171,
|
| 155 |
+
"grad_norm": 3.609375,
|
| 156 |
+
"learning_rate": 1.7599743630828395e-05,
|
| 157 |
+
"loss": 2.167020263671875,
|
| 158 |
+
"mean_token_accuracy": 0.516358617246151,
|
| 159 |
+
"num_tokens": 4515774.0,
|
| 160 |
+
"step": 750
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"entropy": 2.1549544781446457,
|
| 164 |
+
"epoch": 0.1281999919875005,
|
| 165 |
+
"grad_norm": 3.28125,
|
| 166 |
+
"learning_rate": 1.7439512898573947e-05,
|
| 167 |
+
"loss": 2.181917877197266,
|
| 168 |
+
"mean_token_accuracy": 0.5182220858335495,
|
| 169 |
+
"num_tokens": 4814709.0,
|
| 170 |
+
"step": 800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"entropy": 2.1532615756988527,
|
| 174 |
+
"epoch": 0.1362124914867193,
|
| 175 |
+
"grad_norm": 3.078125,
|
| 176 |
+
"learning_rate": 1.7279282166319503e-05,
|
| 177 |
+
"loss": 2.1659007263183594,
|
| 178 |
+
"mean_token_accuracy": 0.5180425970256328,
|
| 179 |
+
"num_tokens": 5112585.0,
|
| 180 |
+
"step": 850
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"entropy": 2.1704017293453215,
|
| 184 |
+
"epoch": 0.14422499098593805,
|
| 185 |
+
"grad_norm": 2.921875,
|
| 186 |
+
"learning_rate": 1.7119051434065055e-05,
|
| 187 |
+
"loss": 2.1934996032714844,
|
| 188 |
+
"mean_token_accuracy": 0.5154964691400528,
|
| 189 |
+
"num_tokens": 5411369.0,
|
| 190 |
+
"step": 900
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"entropy": 2.1349107217788696,
|
| 194 |
+
"epoch": 0.15223749048515683,
|
| 195 |
+
"grad_norm": 2.984375,
|
| 196 |
+
"learning_rate": 1.6958820701810608e-05,
|
| 197 |
+
"loss": 2.1598545837402345,
|
| 198 |
+
"mean_token_accuracy": 0.5208920259773732,
|
| 199 |
+
"num_tokens": 5707060.0,
|
| 200 |
+
"step": 950
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"entropy": 2.1541705918312073,
|
| 204 |
+
"epoch": 0.16024998998437562,
|
| 205 |
+
"grad_norm": 3.203125,
|
| 206 |
+
"learning_rate": 1.6798589969556163e-05,
|
| 207 |
+
"loss": 2.170956573486328,
|
| 208 |
+
"mean_token_accuracy": 0.5204938448965549,
|
| 209 |
+
"num_tokens": 6004079.0,
|
| 210 |
+
"step": 1000
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"entropy": 2.1523867017030716,
|
| 214 |
+
"epoch": 0.1682624894835944,
|
| 215 |
+
"grad_norm": 3.09375,
|
| 216 |
+
"learning_rate": 1.6638359237301716e-05,
|
| 217 |
+
"loss": 2.1695118713378907,
|
| 218 |
+
"mean_token_accuracy": 0.5177376234531402,
|
| 219 |
+
"num_tokens": 6315462.0,
|
| 220 |
+
"step": 1050
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"entropy": 2.1356022965908052,
|
| 224 |
+
"epoch": 0.17627498898281319,
|
| 225 |
+
"grad_norm": 3.21875,
|
| 226 |
+
"learning_rate": 1.6478128505047268e-05,
|
| 227 |
+
"loss": 2.1513107299804686,
|
| 228 |
+
"mean_token_accuracy": 0.5217180585861206,
|
| 229 |
+
"num_tokens": 6613060.0,
|
| 230 |
+
"step": 1100
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"entropy": 2.1483460652828215,
|
| 234 |
+
"epoch": 0.18428748848203197,
|
| 235 |
+
"grad_norm": 3.21875,
|
| 236 |
+
"learning_rate": 1.6317897772792824e-05,
|
| 237 |
+
"loss": 2.1671595764160156,
|
| 238 |
+
"mean_token_accuracy": 0.5199222292006016,
|
| 239 |
+
"num_tokens": 6909655.0,
|
| 240 |
+
"step": 1150
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"entropy": 2.1074040442705155,
|
| 244 |
+
"epoch": 0.19229998798125075,
|
| 245 |
+
"grad_norm": 3.828125,
|
| 246 |
+
"learning_rate": 1.6157667040538376e-05,
|
| 247 |
+
"loss": 2.12770751953125,
|
| 248 |
+
"mean_token_accuracy": 0.5252153894305229,
|
| 249 |
+
"num_tokens": 7203263.0,
|
| 250 |
+
"step": 1200
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"entropy": 2.1354027593135836,
|
| 254 |
+
"epoch": 0.20031248748046954,
|
| 255 |
+
"grad_norm": 2.921875,
|
| 256 |
+
"learning_rate": 1.5997436308283932e-05,
|
| 257 |
+
"loss": 2.156183166503906,
|
| 258 |
+
"mean_token_accuracy": 0.518907299786806,
|
| 259 |
+
"num_tokens": 7505411.0,
|
| 260 |
+
"step": 1250
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"entropy": 2.142924894094467,
|
| 264 |
+
"epoch": 0.20832498697968832,
|
| 265 |
+
"grad_norm": 3.5625,
|
| 266 |
+
"learning_rate": 1.5837205576029484e-05,
|
| 267 |
+
"loss": 2.158486633300781,
|
| 268 |
+
"mean_token_accuracy": 0.5205909128487111,
|
| 269 |
+
"num_tokens": 7803562.0,
|
| 270 |
+
"step": 1300
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"entropy": 2.1305291467905043,
|
| 274 |
+
"epoch": 0.2163374864789071,
|
| 275 |
+
"grad_norm": 3.34375,
|
| 276 |
+
"learning_rate": 1.5676974843775036e-05,
|
| 277 |
+
"loss": 2.151043701171875,
|
| 278 |
+
"mean_token_accuracy": 0.5211617235839366,
|
| 279 |
+
"num_tokens": 8108037.0,
|
| 280 |
+
"step": 1350
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"entropy": 2.160947166085243,
|
| 284 |
+
"epoch": 0.2243499859781259,
|
| 285 |
+
"grad_norm": 3.234375,
|
| 286 |
+
"learning_rate": 1.5516744111520592e-05,
|
| 287 |
+
"loss": 2.177792663574219,
|
| 288 |
+
"mean_token_accuracy": 0.5167004990577698,
|
| 289 |
+
"num_tokens": 8408674.0,
|
| 290 |
+
"step": 1400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"entropy": 2.120403967499733,
|
| 294 |
+
"epoch": 0.23236248547734467,
|
| 295 |
+
"grad_norm": 3.375,
|
| 296 |
+
"learning_rate": 1.5356513379266144e-05,
|
| 297 |
+
"loss": 2.1462562561035154,
|
| 298 |
+
"mean_token_accuracy": 0.5225985661149025,
|
| 299 |
+
"num_tokens": 8707650.0,
|
| 300 |
+
"step": 1450
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"entropy": 2.138882914185524,
|
| 304 |
+
"epoch": 0.24037498497656343,
|
| 305 |
+
"grad_norm": 3.171875,
|
| 306 |
+
"learning_rate": 1.5196282647011698e-05,
|
| 307 |
+
"loss": 2.1499905395507812,
|
| 308 |
+
"mean_token_accuracy": 0.5204668046534061,
|
| 309 |
+
"num_tokens": 9006296.0,
|
| 310 |
+
"step": 1500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"entropy": 2.131927245259285,
|
| 314 |
+
"epoch": 0.2483874844757822,
|
| 315 |
+
"grad_norm": 3.546875,
|
| 316 |
+
"learning_rate": 1.5036051914757252e-05,
|
| 317 |
+
"loss": 2.157076416015625,
|
| 318 |
+
"mean_token_accuracy": 0.5220408822596073,
|
| 319 |
+
"num_tokens": 9307203.0,
|
| 320 |
+
"step": 1550
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"entropy": 2.1308567851781843,
|
| 324 |
+
"epoch": 0.256399983975001,
|
| 325 |
+
"grad_norm": 3.546875,
|
| 326 |
+
"learning_rate": 1.4875821182502805e-05,
|
| 327 |
+
"loss": 2.146601104736328,
|
| 328 |
+
"mean_token_accuracy": 0.5219539681077003,
|
| 329 |
+
"num_tokens": 9610070.0,
|
| 330 |
+
"step": 1600
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"entropy": 2.1097783839702604,
|
| 334 |
+
"epoch": 0.2644124834742198,
|
| 335 |
+
"grad_norm": 3.84375,
|
| 336 |
+
"learning_rate": 1.4715590450248359e-05,
|
| 337 |
+
"loss": 2.1268209838867187,
|
| 338 |
+
"mean_token_accuracy": 0.5236652019619942,
|
| 339 |
+
"num_tokens": 9909117.0,
|
| 340 |
+
"step": 1650
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"entropy": 2.123499180674553,
|
| 344 |
+
"epoch": 0.2724249829734386,
|
| 345 |
+
"grad_norm": 3.46875,
|
| 346 |
+
"learning_rate": 1.4555359717993913e-05,
|
| 347 |
+
"loss": 2.1409934997558593,
|
| 348 |
+
"mean_token_accuracy": 0.5225465354323388,
|
| 349 |
+
"num_tokens": 10209461.0,
|
| 350 |
+
"step": 1700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"entropy": 2.13052960395813,
|
| 354 |
+
"epoch": 0.2804374824726573,
|
| 355 |
+
"grad_norm": 2.71875,
|
| 356 |
+
"learning_rate": 1.4395128985739467e-05,
|
| 357 |
+
"loss": 2.151373291015625,
|
| 358 |
+
"mean_token_accuracy": 0.5227928113937378,
|
| 359 |
+
"num_tokens": 10509282.0,
|
| 360 |
+
"step": 1750
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"entropy": 2.1226114094257356,
|
| 364 |
+
"epoch": 0.2884499819718761,
|
| 365 |
+
"grad_norm": 3.65625,
|
| 366 |
+
"learning_rate": 1.4234898253485019e-05,
|
| 367 |
+
"loss": 2.144451141357422,
|
| 368 |
+
"mean_token_accuracy": 0.5233827362954616,
|
| 369 |
+
"num_tokens": 10811056.0,
|
| 370 |
+
"step": 1800
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"entropy": 2.126143154501915,
|
| 374 |
+
"epoch": 0.2964624814710949,
|
| 375 |
+
"grad_norm": 3.359375,
|
| 376 |
+
"learning_rate": 1.4074667521230575e-05,
|
| 377 |
+
"loss": 2.1457598876953123,
|
| 378 |
+
"mean_token_accuracy": 0.5221967925131321,
|
| 379 |
+
"num_tokens": 11111153.0,
|
| 380 |
+
"step": 1850
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"entropy": 2.1048326581716537,
|
| 384 |
+
"epoch": 0.30447498097031367,
|
| 385 |
+
"grad_norm": 3.59375,
|
| 386 |
+
"learning_rate": 1.3914436788976127e-05,
|
| 387 |
+
"loss": 2.1260116577148436,
|
| 388 |
+
"mean_token_accuracy": 0.5272719909250736,
|
| 389 |
+
"num_tokens": 11405606.0,
|
| 390 |
+
"step": 1900
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"entropy": 2.1236617982387545,
|
| 394 |
+
"epoch": 0.31248748046953245,
|
| 395 |
+
"grad_norm": 3.0,
|
| 396 |
+
"learning_rate": 1.375420605672168e-05,
|
| 397 |
+
"loss": 2.151782684326172,
|
| 398 |
+
"mean_token_accuracy": 0.5227546173334122,
|
| 399 |
+
"num_tokens": 11703943.0,
|
| 400 |
+
"step": 1950
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"entropy": 2.121563429236412,
|
| 404 |
+
"epoch": 0.32049997996875124,
|
| 405 |
+
"grad_norm": 3.34375,
|
| 406 |
+
"learning_rate": 1.3593975324467235e-05,
|
| 407 |
+
"loss": 2.1370144653320313,
|
| 408 |
+
"mean_token_accuracy": 0.5235123547911644,
|
| 409 |
+
"num_tokens": 12002522.0,
|
| 410 |
+
"step": 2000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"entropy": 2.1155461168289182,
|
| 414 |
+
"epoch": 0.32851247946797,
|
| 415 |
+
"grad_norm": 3.46875,
|
| 416 |
+
"learning_rate": 1.3433744592212787e-05,
|
| 417 |
+
"loss": 2.1364064025878906,
|
| 418 |
+
"mean_token_accuracy": 0.5237135806679726,
|
| 419 |
+
"num_tokens": 12303098.0,
|
| 420 |
+
"step": 2050
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"entropy": 2.1386052322387696,
|
| 424 |
+
"epoch": 0.3365249789671888,
|
| 425 |
+
"grad_norm": 3.484375,
|
| 426 |
+
"learning_rate": 1.327351385995834e-05,
|
| 427 |
+
"loss": 2.149888458251953,
|
| 428 |
+
"mean_token_accuracy": 0.522669050693512,
|
| 429 |
+
"num_tokens": 12605534.0,
|
| 430 |
+
"step": 2100
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"entropy": 2.102600795030594,
|
| 434 |
+
"epoch": 0.3445374784664076,
|
| 435 |
+
"grad_norm": 3.796875,
|
| 436 |
+
"learning_rate": 1.3113283127703895e-05,
|
| 437 |
+
"loss": 2.1236566162109374,
|
| 438 |
+
"mean_token_accuracy": 0.5279092155396938,
|
| 439 |
+
"num_tokens": 12901218.0,
|
| 440 |
+
"step": 2150
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"entropy": 2.132645187973976,
|
| 444 |
+
"epoch": 0.35254997796562637,
|
| 445 |
+
"grad_norm": 3.53125,
|
| 446 |
+
"learning_rate": 1.2953052395449447e-05,
|
| 447 |
+
"loss": 2.1523503112792968,
|
| 448 |
+
"mean_token_accuracy": 0.5198129612207413,
|
| 449 |
+
"num_tokens": 13213018.0,
|
| 450 |
+
"step": 2200
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"entropy": 2.1204321217536926,
|
| 454 |
+
"epoch": 0.36056247746484515,
|
| 455 |
+
"grad_norm": 3.25,
|
| 456 |
+
"learning_rate": 1.2792821663195001e-05,
|
| 457 |
+
"loss": 2.1325447082519533,
|
| 458 |
+
"mean_token_accuracy": 0.5243636940419674,
|
| 459 |
+
"num_tokens": 13514408.0,
|
| 460 |
+
"step": 2250
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"entropy": 2.102410731315613,
|
| 464 |
+
"epoch": 0.36857497696406394,
|
| 465 |
+
"grad_norm": 3.1875,
|
| 466 |
+
"learning_rate": 1.2632590930940555e-05,
|
| 467 |
+
"loss": 2.123565368652344,
|
| 468 |
+
"mean_token_accuracy": 0.5257614709436893,
|
| 469 |
+
"num_tokens": 13818272.0,
|
| 470 |
+
"step": 2300
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"entropy": 2.0876205080747603,
|
| 474 |
+
"epoch": 0.3765874764632827,
|
| 475 |
+
"grad_norm": 3.328125,
|
| 476 |
+
"learning_rate": 1.247236019868611e-05,
|
| 477 |
+
"loss": 2.116696472167969,
|
| 478 |
+
"mean_token_accuracy": 0.5275124031305313,
|
| 479 |
+
"num_tokens": 14110702.0,
|
| 480 |
+
"step": 2350
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"entropy": 2.1078518223762512,
|
| 484 |
+
"epoch": 0.3845999759625015,
|
| 485 |
+
"grad_norm": 3.453125,
|
| 486 |
+
"learning_rate": 1.2312129466431662e-05,
|
| 487 |
+
"loss": 2.1291993713378905,
|
| 488 |
+
"mean_token_accuracy": 0.524967464953661,
|
| 489 |
+
"num_tokens": 14407015.0,
|
| 490 |
+
"step": 2400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"entropy": 2.0930898267030718,
|
| 494 |
+
"epoch": 0.3926124754617203,
|
| 495 |
+
"grad_norm": 3.5,
|
| 496 |
+
"learning_rate": 1.2151898734177216e-05,
|
| 497 |
+
"loss": 2.1094123840332033,
|
| 498 |
+
"mean_token_accuracy": 0.5290587748587131,
|
| 499 |
+
"num_tokens": 14696565.0,
|
| 500 |
+
"step": 2450
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"entropy": 2.1109757220745085,
|
| 504 |
+
"epoch": 0.4006249749609391,
|
| 505 |
+
"grad_norm": 3.5,
|
| 506 |
+
"learning_rate": 1.199166800192277e-05,
|
| 507 |
+
"loss": 2.1246888732910154,
|
| 508 |
+
"mean_token_accuracy": 0.5258957357704639,
|
| 509 |
+
"num_tokens": 14994401.0,
|
| 510 |
+
"step": 2500
|
| 511 |
+
}
|
| 512 |
+
],
|
| 513 |
+
"logging_steps": 50,
|
| 514 |
+
"max_steps": 6241,
|
| 515 |
+
"num_input_tokens_seen": 0,
|
| 516 |
+
"num_train_epochs": 1,
|
| 517 |
+
"save_steps": 500,
|
| 518 |
+
"stateful_callbacks": {
|
| 519 |
+
"TrainerControl": {
|
| 520 |
+
"args": {
|
| 521 |
+
"should_epoch_stop": false,
|
| 522 |
+
"should_evaluate": false,
|
| 523 |
+
"should_log": false,
|
| 524 |
+
"should_save": true,
|
| 525 |
+
"should_training_stop": false
|
| 526 |
+
},
|
| 527 |
+
"attributes": {}
|
| 528 |
+
}
|
| 529 |
+
},
|
| 530 |
+
"total_flos": 1.773461452668027e+17,
|
| 531 |
+
"train_batch_size": 4,
|
| 532 |
+
"trial_name": null,
|
| 533 |
+
"trial_params": null
|
| 534 |
+
}
|
checkpoint-2500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78622f63304ec0913ed0b4ea7a1b26c08a5621b3661a380d798c4ea426235671
|
| 3 |
+
size 5713
|
checkpoint-3000/chat_template.jinja
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if message.content is string %}
|
| 27 |
+
{%- set content = message.content %}
|
| 28 |
+
{%- else %}
|
| 29 |
+
{%- set content = '' %}
|
| 30 |
+
{%- endif %}
|
| 31 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 32 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 33 |
+
{%- elif message.role == "assistant" %}
|
| 34 |
+
{%- set reasoning_content = '' %}
|
| 35 |
+
{%- if message.reasoning_content is string %}
|
| 36 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 37 |
+
{%- else %}
|
| 38 |
+
{%- if '</think>' in content %}
|
| 39 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 40 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 41 |
+
{%- endif %}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 44 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 45 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 46 |
+
{%- else %}
|
| 47 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 48 |
+
{%- endif %}
|
| 49 |
+
{%- else %}
|
| 50 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 51 |
+
{%- endif %}
|
| 52 |
+
{%- if message.tool_calls %}
|
| 53 |
+
{%- for tool_call in message.tool_calls %}
|
| 54 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 55 |
+
{{- '\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tool_call.function %}
|
| 58 |
+
{%- set tool_call = tool_call.function %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 61 |
+
{{- tool_call.name }}
|
| 62 |
+
{{- '", "arguments": ' }}
|
| 63 |
+
{%- if tool_call.arguments is string %}
|
| 64 |
+
{{- tool_call.arguments }}
|
| 65 |
+
{%- else %}
|
| 66 |
+
{{- tool_call.arguments | tojson }}
|
| 67 |
+
{%- endif %}
|
| 68 |
+
{{- '}\n</tool_call>' }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{%- endif %}
|
| 71 |
+
{{- '<|im_end|>\n' }}
|
| 72 |
+
{%- elif message.role == "tool" %}
|
| 73 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 74 |
+
{{- '<|im_start|>user' }}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{{- '\n<tool_response>\n' }}
|
| 77 |
+
{{- content }}
|
| 78 |
+
{{- '\n</tool_response>' }}
|
| 79 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 80 |
+
{{- '<|im_end|>\n' }}
|
| 81 |
+
{%- endif %}
|
| 82 |
+
{%- endif %}
|
| 83 |
+
{%- endfor %}
|
| 84 |
+
{%- if add_generation_prompt %}
|
| 85 |
+
{{- '<|im_start|>assistant\n' }}
|
| 86 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 87 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 88 |
+
{%- endif %}
|
| 89 |
+
{%- endif %}
|
checkpoint-3000/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 6144,
|
| 15 |
+
"layer_types": [
|
| 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 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention"
|
| 44 |
+
],
|
| 45 |
+
"max_position_embeddings": 40960,
|
| 46 |
+
"max_window_layers": 28,
|
| 47 |
+
"model_type": "qwen3",
|
| 48 |
+
"num_attention_heads": 16,
|
| 49 |
+
"num_hidden_layers": 28,
|
| 50 |
+
"num_key_value_heads": 8,
|
| 51 |
+
"pad_token_id": 151643,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_theta": 1000000,
|
| 55 |
+
"rope_type": "default"
|
| 56 |
+
},
|
| 57 |
+
"sliding_window": null,
|
| 58 |
+
"tie_word_embeddings": true,
|
| 59 |
+
"transformers_version": "5.7.0",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-3000/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"temperature": 0.6,
|
| 9 |
+
"top_k": 20,
|
| 10 |
+
"top_p": 0.95,
|
| 11 |
+
"transformers_version": "5.7.0"
|
| 12 |
+
}
|