Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- README.md +2 -2
- checkpoint-64/chat_template.jinja +89 -0
- checkpoint-64/config.json +63 -0
- checkpoint-64/generation_config.json +12 -0
- checkpoint-64/model.safetensors +3 -0
- checkpoint-64/optimizer.pt +3 -0
- checkpoint-64/rng_state.pth +3 -0
- checkpoint-64/scheduler.pt +3 -0
- checkpoint-64/tokenizer.json +3 -0
- checkpoint-64/tokenizer_config.json +75 -0
- checkpoint-64/trainer_state.json +208 -0
- checkpoint-64/training_args.bin +3 -0
- completions/completions_00010.parquet +2 -2
- completions/completions_00020.parquet +3 -0
- completions/completions_00030.parquet +3 -0
- completions/completions_00040.parquet +3 -0
- completions/completions_00050.parquet +3 -0
- completions/completions_00060.parquet +3 -0
- completions/completions_00064.parquet +3 -0
- metrics/loss_curve.svg +4 -4
- metrics/loss_history.json +21 -1
- metrics/reward_curve.svg +4 -4
- metrics/reward_history.json +23 -3
- run_manifest.json +4 -4
- trainer_state.json +192 -47
.gitattributes
CHANGED
|
@@ -34,3 +34,4 @@ saved_model/**/* 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-12/tokenizer.json 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-12/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-64/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen3-0.6B
|
| 3 |
library_name: transformers
|
| 4 |
-
model_name:
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
- hf_jobs
|
|
@@ -10,7 +10,7 @@ tags:
|
|
| 10 |
licence: license
|
| 11 |
---
|
| 12 |
|
| 13 |
-
# Model Card for
|
| 14 |
|
| 15 |
This model is a fine-tuned version of [Qwen/Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B).
|
| 16 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
|
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen3-0.6B
|
| 3 |
library_name: transformers
|
| 4 |
+
model_name: pulse_er_grpo_curve_v2
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
- hf_jobs
|
|
|
|
| 10 |
licence: license
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# Model Card for pulse_er_grpo_curve_v2
|
| 14 |
|
| 15 |
This model is a fine-tuned version of [Qwen/Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B).
|
| 16 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
checkpoint-64/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-64/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": "float32",
|
| 9 |
+
"eos_token_id": 151645,
|
| 10 |
+
"head_dim": 128,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1024,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 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": "5.6.2",
|
| 60 |
+
"use_cache": false,
|
| 61 |
+
"use_sliding_window": false,
|
| 62 |
+
"vocab_size": 151936
|
| 63 |
+
}
|
checkpoint-64/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": "5.6.2"
|
| 12 |
+
}
|
checkpoint-64/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9daeb3bbd2ebb9b4a9910781ba136204aaeef6c683148c50637f29a347a9b51e
|
| 3 |
+
size 2384234968
|
checkpoint-64/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae22a58b918c4e55813693f56a60d74d3036e2924f45a5f20888c6e0c8074a2f
|
| 3 |
+
size 4768664614
|
checkpoint-64/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b95731ce312b6d3f623d553049d545337c5d22f01090116e149fd3cf089643e
|
| 3 |
+
size 14244
|
checkpoint-64/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:062d9669d1ff2893787af221dc1a10f378ff916c64725933b3d39e4de4dd1029
|
| 3 |
+
size 1064
|
checkpoint-64/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
checkpoint-64/tokenizer_config.json
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"padding_side": "left",
|
| 28 |
+
"response_schema": {
|
| 29 |
+
"properties": {
|
| 30 |
+
"content": {
|
| 31 |
+
"type": "string"
|
| 32 |
+
},
|
| 33 |
+
"reasoning_content": {
|
| 34 |
+
"type": "string"
|
| 35 |
+
},
|
| 36 |
+
"role": {
|
| 37 |
+
"const": "assistant"
|
| 38 |
+
},
|
| 39 |
+
"tool_calls": {
|
| 40 |
+
"items": {
|
| 41 |
+
"properties": {
|
| 42 |
+
"function": {
|
| 43 |
+
"properties": {
|
| 44 |
+
"arguments": {
|
| 45 |
+
"additionalProperties": {},
|
| 46 |
+
"type": "object"
|
| 47 |
+
},
|
| 48 |
+
"name": {
|
| 49 |
+
"type": "string"
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"type": "object"
|
| 53 |
+
},
|
| 54 |
+
"type": {
|
| 55 |
+
"const": "function"
|
| 56 |
+
}
|
| 57 |
+
},
|
| 58 |
+
"type": "object",
|
| 59 |
+
"x-parser": "json",
|
| 60 |
+
"x-parser-args": {
|
| 61 |
+
"transform": "{type: 'function', function: @}"
|
| 62 |
+
}
|
| 63 |
+
},
|
| 64 |
+
"type": "array",
|
| 65 |
+
"x-regex-iterator": "<tool_call>\\s*(.+?)\\s*</tool_call>"
|
| 66 |
+
}
|
| 67 |
+
},
|
| 68 |
+
"type": "object",
|
| 69 |
+
"x-regex": "^(?:<think>\\n?(?:(?P<reasoning_content>.*?\\S.*?)\\n?|[\\s]*)</think>\\s*)?(?P<content>.*?)(?:\\n(?=<tool_call>))?(?=(?:<tool_call>|<\\|im_end\\|>|$))(?P<tool_calls>(?:<tool_call>.+?</tool_call>\\s*)+)?\\s*(?:<\\|im_end\\|>|$)"
|
| 70 |
+
},
|
| 71 |
+
"split_special_tokens": false,
|
| 72 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 73 |
+
"truncation_side": "left",
|
| 74 |
+
"unk_token": null
|
| 75 |
+
}
|
checkpoint-64/trainer_state.json
ADDED
|
@@ -0,0 +1,208 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 64,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"clip_ratio/high_max": 0.0,
|
| 14 |
+
"clip_ratio/high_mean": 0.0,
|
| 15 |
+
"clip_ratio/low_mean": 0.0,
|
| 16 |
+
"clip_ratio/low_min": 0.0,
|
| 17 |
+
"clip_ratio/region_mean": 0.0,
|
| 18 |
+
"completions/clipped_ratio": 0.0,
|
| 19 |
+
"completions/max_length": 281.7,
|
| 20 |
+
"completions/max_terminated_length": 281.7,
|
| 21 |
+
"completions/mean_length": 173.90625,
|
| 22 |
+
"completions/mean_terminated_length": 173.90625,
|
| 23 |
+
"completions/min_length": 60.1,
|
| 24 |
+
"completions/min_terminated_length": 60.1,
|
| 25 |
+
"entropy": 0.34981602281332014,
|
| 26 |
+
"epoch": 0.3125,
|
| 27 |
+
"frac_reward_zero_std": 0.2,
|
| 28 |
+
"grad_norm": 3.5286595821380615,
|
| 29 |
+
"learning_rate": 8.593749999999999e-07,
|
| 30 |
+
"loss": -0.03454443216323853,
|
| 31 |
+
"num_tokens": 363490.0,
|
| 32 |
+
"reward": -1.9279687214642762,
|
| 33 |
+
"reward_std": 6.467041900753975,
|
| 34 |
+
"rewards/pulse_reward/mean": -1.9279687214642762,
|
| 35 |
+
"rewards/pulse_reward/std": 6.467041908204555,
|
| 36 |
+
"step": 10,
|
| 37 |
+
"step_time": 35.20241980789724,
|
| 38 |
+
"tools/call_frequency": 5.49375,
|
| 39 |
+
"tools/failure_frequency": 0.019680690905079245
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"clip_ratio/high_max": 0.0,
|
| 43 |
+
"clip_ratio/high_mean": 0.0,
|
| 44 |
+
"clip_ratio/low_mean": 0.0,
|
| 45 |
+
"clip_ratio/low_min": 0.0,
|
| 46 |
+
"clip_ratio/region_mean": 0.0,
|
| 47 |
+
"completions/clipped_ratio": 0.0,
|
| 48 |
+
"completions/max_length": 294.3,
|
| 49 |
+
"completions/max_terminated_length": 294.3,
|
| 50 |
+
"completions/mean_length": 172.56875,
|
| 51 |
+
"completions/mean_terminated_length": 172.56875,
|
| 52 |
+
"completions/min_length": 50.3,
|
| 53 |
+
"completions/min_terminated_length": 50.3,
|
| 54 |
+
"entropy": 0.23517516404390335,
|
| 55 |
+
"epoch": 0.625,
|
| 56 |
+
"frac_reward_zero_std": 0.2625,
|
| 57 |
+
"grad_norm": 3.8192903995513916,
|
| 58 |
+
"learning_rate": 7.031249999999999e-07,
|
| 59 |
+
"loss": 0.01356593519449234,
|
| 60 |
+
"num_tokens": 726552.0,
|
| 61 |
+
"reward": 0.09853125289082527,
|
| 62 |
+
"reward_std": 0.583341383934021,
|
| 63 |
+
"rewards/pulse_reward/mean": 0.09853125289082527,
|
| 64 |
+
"rewards/pulse_reward/std": 0.583341383934021,
|
| 65 |
+
"step": 20,
|
| 66 |
+
"step_time": 39.57589099079487,
|
| 67 |
+
"tools/call_frequency": 6.35,
|
| 68 |
+
"tools/failure_frequency": 0.005996426660567522
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"clip_ratio/high_max": 0.0,
|
| 72 |
+
"clip_ratio/high_mean": 0.0,
|
| 73 |
+
"clip_ratio/low_mean": 0.0,
|
| 74 |
+
"clip_ratio/low_min": 0.0,
|
| 75 |
+
"clip_ratio/region_mean": 0.0,
|
| 76 |
+
"completions/clipped_ratio": 0.0,
|
| 77 |
+
"completions/max_length": 273.0,
|
| 78 |
+
"completions/max_terminated_length": 273.0,
|
| 79 |
+
"completions/mean_length": 149.115625,
|
| 80 |
+
"completions/mean_terminated_length": 149.115625,
|
| 81 |
+
"completions/min_length": 50.9,
|
| 82 |
+
"completions/min_terminated_length": 50.9,
|
| 83 |
+
"entropy": 0.1797773003578186,
|
| 84 |
+
"epoch": 0.9375,
|
| 85 |
+
"frac_reward_zero_std": 0.0875,
|
| 86 |
+
"grad_norm": 5.692322254180908,
|
| 87 |
+
"learning_rate": 5.46875e-07,
|
| 88 |
+
"loss": 0.06958286762237549,
|
| 89 |
+
"num_tokens": 1082109.0,
|
| 90 |
+
"reward": 0.3879156395792961,
|
| 91 |
+
"reward_std": 0.6661841243505477,
|
| 92 |
+
"rewards/pulse_reward/mean": 0.3879156395792961,
|
| 93 |
+
"rewards/pulse_reward/std": 0.6661841526627541,
|
| 94 |
+
"step": 30,
|
| 95 |
+
"step_time": 30.443527194001945,
|
| 96 |
+
"tools/call_frequency": 5.58125,
|
| 97 |
+
"tools/failure_frequency": 0.005842319130897522
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"clip_ratio/high_max": 0.0,
|
| 101 |
+
"clip_ratio/high_mean": 0.0,
|
| 102 |
+
"clip_ratio/low_mean": 0.0,
|
| 103 |
+
"clip_ratio/low_min": 0.0,
|
| 104 |
+
"clip_ratio/region_mean": 0.0,
|
| 105 |
+
"completions/clipped_ratio": 0.0,
|
| 106 |
+
"completions/max_length": 115.3,
|
| 107 |
+
"completions/max_terminated_length": 115.3,
|
| 108 |
+
"completions/mean_length": 70.528125,
|
| 109 |
+
"completions/mean_terminated_length": 70.528125,
|
| 110 |
+
"completions/min_length": 38.1,
|
| 111 |
+
"completions/min_terminated_length": 38.1,
|
| 112 |
+
"entropy": 0.1547975329682231,
|
| 113 |
+
"epoch": 1.25,
|
| 114 |
+
"frac_reward_zero_std": 0.1625,
|
| 115 |
+
"grad_norm": 3.677471399307251,
|
| 116 |
+
"learning_rate": 3.9062499999999997e-07,
|
| 117 |
+
"loss": 0.02682075500488281,
|
| 118 |
+
"num_tokens": 1412518.0,
|
| 119 |
+
"reward": 1.0730250239372254,
|
| 120 |
+
"reward_std": 0.7086882412433624,
|
| 121 |
+
"rewards/pulse_reward/mean": 1.0730250239372254,
|
| 122 |
+
"rewards/pulse_reward/std": 0.7086882352828979,
|
| 123 |
+
"step": 40,
|
| 124 |
+
"step_time": 11.361506971600466,
|
| 125 |
+
"tools/call_frequency": 2.215625,
|
| 126 |
+
"tools/failure_frequency": 0.0015384615398943424
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"clip_ratio/high_max": 0.0,
|
| 130 |
+
"clip_ratio/high_mean": 0.0,
|
| 131 |
+
"clip_ratio/low_mean": 0.0,
|
| 132 |
+
"clip_ratio/low_min": 0.0,
|
| 133 |
+
"clip_ratio/region_mean": 0.0,
|
| 134 |
+
"completions/clipped_ratio": 0.0,
|
| 135 |
+
"completions/max_length": 83.1,
|
| 136 |
+
"completions/max_terminated_length": 83.1,
|
| 137 |
+
"completions/mean_length": 65.421875,
|
| 138 |
+
"completions/mean_terminated_length": 65.421875,
|
| 139 |
+
"completions/min_length": 48.0,
|
| 140 |
+
"completions/min_terminated_length": 48.0,
|
| 141 |
+
"entropy": 0.12452723067253828,
|
| 142 |
+
"epoch": 1.5625,
|
| 143 |
+
"frac_reward_zero_std": 0.6625,
|
| 144 |
+
"grad_norm": 3.609057664871216,
|
| 145 |
+
"learning_rate": 2.3437499999999998e-07,
|
| 146 |
+
"loss": -3.346521407365799e-05,
|
| 147 |
+
"num_tokens": 1741293.0,
|
| 148 |
+
"reward": 1.4852312803268433,
|
| 149 |
+
"reward_std": 0.3715872406959534,
|
| 150 |
+
"rewards/pulse_reward/mean": 1.4852312803268433,
|
| 151 |
+
"rewards/pulse_reward/std": 0.3715872406959534,
|
| 152 |
+
"step": 50,
|
| 153 |
+
"step_time": 8.49522676919878,
|
| 154 |
+
"tools/call_frequency": 2.0,
|
| 155 |
+
"tools/failure_frequency": 0.0
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"clip_ratio/high_max": 0.0,
|
| 159 |
+
"clip_ratio/high_mean": 0.0,
|
| 160 |
+
"clip_ratio/low_mean": 0.0,
|
| 161 |
+
"clip_ratio/low_min": 0.0,
|
| 162 |
+
"clip_ratio/region_mean": 0.0,
|
| 163 |
+
"completions/clipped_ratio": 0.0,
|
| 164 |
+
"completions/max_length": 86.6,
|
| 165 |
+
"completions/max_terminated_length": 86.6,
|
| 166 |
+
"completions/mean_length": 65.471875,
|
| 167 |
+
"completions/mean_terminated_length": 65.471875,
|
| 168 |
+
"completions/min_length": 53.6,
|
| 169 |
+
"completions/min_terminated_length": 53.6,
|
| 170 |
+
"entropy": 0.11676975060254335,
|
| 171 |
+
"epoch": 1.875,
|
| 172 |
+
"frac_reward_zero_std": 0.8125,
|
| 173 |
+
"grad_norm": 2.142204761505127,
|
| 174 |
+
"learning_rate": 7.812499999999999e-08,
|
| 175 |
+
"loss": -0.00026753861457109453,
|
| 176 |
+
"num_tokens": 2070084.0,
|
| 177 |
+
"reward": 1.538143789768219,
|
| 178 |
+
"reward_std": 0.2553505107760429,
|
| 179 |
+
"rewards/pulse_reward/mean": 1.538143789768219,
|
| 180 |
+
"rewards/pulse_reward/std": 0.25535051375627515,
|
| 181 |
+
"step": 60,
|
| 182 |
+
"step_time": 8.590878555196104,
|
| 183 |
+
"tools/call_frequency": 1.996875,
|
| 184 |
+
"tools/failure_frequency": 0.0
|
| 185 |
+
}
|
| 186 |
+
],
|
| 187 |
+
"logging_steps": 10,
|
| 188 |
+
"max_steps": 64,
|
| 189 |
+
"num_input_tokens_seen": 2201581,
|
| 190 |
+
"num_train_epochs": 2,
|
| 191 |
+
"save_steps": 500,
|
| 192 |
+
"stateful_callbacks": {
|
| 193 |
+
"TrainerControl": {
|
| 194 |
+
"args": {
|
| 195 |
+
"should_epoch_stop": false,
|
| 196 |
+
"should_evaluate": false,
|
| 197 |
+
"should_log": false,
|
| 198 |
+
"should_save": true,
|
| 199 |
+
"should_training_stop": true
|
| 200 |
+
},
|
| 201 |
+
"attributes": {}
|
| 202 |
+
}
|
| 203 |
+
},
|
| 204 |
+
"total_flos": 0.0,
|
| 205 |
+
"train_batch_size": 8,
|
| 206 |
+
"trial_name": null,
|
| 207 |
+
"trial_params": null
|
| 208 |
+
}
|
checkpoint-64/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:beef4bb9265600265ee78a20bc82f4606af28d145e10bb1b99f8c1c25b2dfef3
|
| 3 |
+
size 6776
|
completions/completions_00010.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3ed1e88b120607f7b8cbc21e6d009da8f473f089b81b63a4cc6d2e30f33eda3
|
| 3 |
+
size 31937
|
completions/completions_00020.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67b0276af96688b86a5d64aa8e3321028609a4197484f003ed3731c25c153538
|
| 3 |
+
size 33667
|
completions/completions_00030.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1499c8f1a195fe8930fd7338a09d90d9f5551f604fcfdab147f034f0ecc71827
|
| 3 |
+
size 28958
|
completions/completions_00040.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:646c066b43013c281723f0d4c62facd9a369b7025636ba5658a350bab1a5180e
|
| 3 |
+
size 23850
|
completions/completions_00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d48aad7bd937639532ebe9dba92c318f0ad86db035761861204516149084a29a
|
| 3 |
+
size 23846
|
completions/completions_00060.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be67ebb6ca105cca7c9f08e11e8259c50ea4bed9b1a3a00a70f000e14ed38c6d
|
| 3 |
+
size 23589
|
completions/completions_00064.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad2757babd27c9411393c4809b69ca6ee606012f352645a97006503671aa6bcb
|
| 3 |
+
size 23756
|
metrics/loss_curve.svg
CHANGED
|
|
metrics/loss_history.json
CHANGED
|
@@ -1,6 +1,26 @@
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 10.0,
|
| 4 |
-
"value": -0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
}
|
| 6 |
]
|
|
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 10.0,
|
| 4 |
+
"value": -0.03454443216323853
|
| 5 |
+
},
|
| 6 |
+
{
|
| 7 |
+
"step": 20.0,
|
| 8 |
+
"value": 0.01356593519449234
|
| 9 |
+
},
|
| 10 |
+
{
|
| 11 |
+
"step": 30.0,
|
| 12 |
+
"value": 0.06958286762237549
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"step": 40.0,
|
| 16 |
+
"value": 0.02682075500488281
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"step": 50.0,
|
| 20 |
+
"value": -3.346521407365799e-05
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"step": 60.0,
|
| 24 |
+
"value": -0.00026753861457109453
|
| 25 |
}
|
| 26 |
]
|
metrics/reward_curve.svg
CHANGED
|
|
metrics/reward_history.json
CHANGED
|
@@ -1,10 +1,30 @@
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 10.0,
|
| 4 |
-
"value": -1.
|
| 5 |
},
|
| 6 |
{
|
| 7 |
-
"step":
|
| 8 |
-
"value":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
}
|
| 10 |
]
|
|
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 10.0,
|
| 4 |
+
"value": -1.9279687214642762
|
| 5 |
},
|
| 6 |
{
|
| 7 |
+
"step": 20.0,
|
| 8 |
+
"value": 0.09853125289082527
|
| 9 |
+
},
|
| 10 |
+
{
|
| 11 |
+
"step": 30.0,
|
| 12 |
+
"value": 0.3879156395792961
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"step": 40.0,
|
| 16 |
+
"value": 1.0730250239372254
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"step": 50.0,
|
| 20 |
+
"value": 1.4852312803268433
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"step": 60.0,
|
| 24 |
+
"value": 1.538143789768219
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"step": 64.0,
|
| 28 |
+
"value": 1.57924222946167
|
| 29 |
}
|
| 30 |
]
|
run_manifest.json
CHANGED
|
@@ -4,14 +4,14 @@
|
|
| 4 |
"env_url": "http://127.0.0.1:8000",
|
| 5 |
"fp16": false,
|
| 6 |
"git_commit": "3b3d65241147f5ac61616a235e860943f956b9af",
|
| 7 |
-
"gradient_accumulation_steps":
|
| 8 |
"learning_rate": 1e-06,
|
| 9 |
"max_steps": 1024,
|
| 10 |
"model": "Qwen/Qwen3-0.6B",
|
| 11 |
"num_generations": 4,
|
| 12 |
-
"num_samples":
|
| 13 |
-
"num_train_epochs":
|
| 14 |
-
"per_device_train_batch_size":
|
| 15 |
"scenario_id": "respiratory_distress",
|
| 16 |
"seed": 42,
|
| 17 |
"use_cpu": false
|
|
|
|
| 4 |
"env_url": "http://127.0.0.1:8000",
|
| 5 |
"fp16": false,
|
| 6 |
"git_commit": "3b3d65241147f5ac61616a235e860943f956b9af",
|
| 7 |
+
"gradient_accumulation_steps": 4,
|
| 8 |
"learning_rate": 1e-06,
|
| 9 |
"max_steps": 1024,
|
| 10 |
"model": "Qwen/Qwen3-0.6B",
|
| 11 |
"num_generations": 4,
|
| 12 |
+
"num_samples": 256,
|
| 13 |
+
"num_train_epochs": 2.0,
|
| 14 |
+
"per_device_train_batch_size": 8,
|
| 15 |
"scenario_id": "respiratory_distress",
|
| 16 |
"seed": 42,
|
| 17 |
"use_cpu": false
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch":
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -16,27 +16,27 @@
|
|
| 16 |
"clip_ratio/low_min": 0.0,
|
| 17 |
"clip_ratio/region_mean": 0.0,
|
| 18 |
"completions/clipped_ratio": 0.0,
|
| 19 |
-
"completions/max_length":
|
| 20 |
-
"completions/max_terminated_length":
|
| 21 |
-
"completions/mean_length":
|
| 22 |
-
"completions/mean_terminated_length":
|
| 23 |
-
"completions/min_length":
|
| 24 |
-
"completions/min_terminated_length":
|
| 25 |
-
"entropy": 0.
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"frac_reward_zero_std": 0.
|
| 28 |
-
"grad_norm":
|
| 29 |
-
"learning_rate":
|
| 30 |
-
"loss": -0.
|
| 31 |
-
"num_tokens":
|
| 32 |
-
"reward": -1.
|
| 33 |
-
"reward_std":
|
| 34 |
-
"rewards/pulse_reward/mean": -1.
|
| 35 |
-
"rewards/pulse_reward/std":
|
| 36 |
"step": 10,
|
| 37 |
-
"step_time":
|
| 38 |
-
"tools/call_frequency":
|
| 39 |
-
"tools/failure_frequency": 0.
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"clip_ratio/high_max": 0.0,
|
|
@@ -45,35 +45,180 @@
|
|
| 45 |
"clip_ratio/low_min": 0.0,
|
| 46 |
"clip_ratio/region_mean": 0.0,
|
| 47 |
"completions/clipped_ratio": 0.0,
|
| 48 |
-
"completions/max_length":
|
| 49 |
-
"completions/max_terminated_length":
|
| 50 |
-
"completions/mean_length":
|
| 51 |
-
"completions/mean_terminated_length":
|
| 52 |
-
"completions/min_length":
|
| 53 |
-
"completions/min_terminated_length":
|
| 54 |
-
"entropy": 0.
|
| 55 |
-
"epoch":
|
| 56 |
-
"frac_reward_zero_std": 0.
|
| 57 |
-
"
|
| 58 |
-
"
|
| 59 |
-
"
|
| 60 |
-
"
|
| 61 |
-
"
|
| 62 |
-
"
|
| 63 |
-
"
|
| 64 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
"tools/failure_frequency": 0.0,
|
| 66 |
"total_flos": 0.0,
|
| 67 |
-
"train_loss":
|
| 68 |
-
"train_runtime":
|
| 69 |
-
"train_samples_per_second": 0.
|
| 70 |
-
"train_steps_per_second": 0.
|
| 71 |
}
|
| 72 |
],
|
| 73 |
"logging_steps": 10,
|
| 74 |
-
"max_steps":
|
| 75 |
-
"num_input_tokens_seen":
|
| 76 |
-
"num_train_epochs":
|
| 77 |
"save_steps": 500,
|
| 78 |
"stateful_callbacks": {
|
| 79 |
"TrainerControl": {
|
|
@@ -88,7 +233,7 @@
|
|
| 88 |
}
|
| 89 |
},
|
| 90 |
"total_flos": 0.0,
|
| 91 |
-
"train_batch_size":
|
| 92 |
"trial_name": null,
|
| 93 |
"trial_params": null
|
| 94 |
}
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 64,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 16 |
"clip_ratio/low_min": 0.0,
|
| 17 |
"clip_ratio/region_mean": 0.0,
|
| 18 |
"completions/clipped_ratio": 0.0,
|
| 19 |
+
"completions/max_length": 281.7,
|
| 20 |
+
"completions/max_terminated_length": 281.7,
|
| 21 |
+
"completions/mean_length": 173.90625,
|
| 22 |
+
"completions/mean_terminated_length": 173.90625,
|
| 23 |
+
"completions/min_length": 60.1,
|
| 24 |
+
"completions/min_terminated_length": 60.1,
|
| 25 |
+
"entropy": 0.34981602281332014,
|
| 26 |
+
"epoch": 0.3125,
|
| 27 |
+
"frac_reward_zero_std": 0.2,
|
| 28 |
+
"grad_norm": 3.5286595821380615,
|
| 29 |
+
"learning_rate": 8.593749999999999e-07,
|
| 30 |
+
"loss": -0.03454443216323853,
|
| 31 |
+
"num_tokens": 363490.0,
|
| 32 |
+
"reward": -1.9279687214642762,
|
| 33 |
+
"reward_std": 6.467041900753975,
|
| 34 |
+
"rewards/pulse_reward/mean": -1.9279687214642762,
|
| 35 |
+
"rewards/pulse_reward/std": 6.467041908204555,
|
| 36 |
"step": 10,
|
| 37 |
+
"step_time": 35.20241980789724,
|
| 38 |
+
"tools/call_frequency": 5.49375,
|
| 39 |
+
"tools/failure_frequency": 0.019680690905079245
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"clip_ratio/high_max": 0.0,
|
|
|
|
| 45 |
"clip_ratio/low_min": 0.0,
|
| 46 |
"clip_ratio/region_mean": 0.0,
|
| 47 |
"completions/clipped_ratio": 0.0,
|
| 48 |
+
"completions/max_length": 294.3,
|
| 49 |
+
"completions/max_terminated_length": 294.3,
|
| 50 |
+
"completions/mean_length": 172.56875,
|
| 51 |
+
"completions/mean_terminated_length": 172.56875,
|
| 52 |
+
"completions/min_length": 50.3,
|
| 53 |
+
"completions/min_terminated_length": 50.3,
|
| 54 |
+
"entropy": 0.23517516404390335,
|
| 55 |
+
"epoch": 0.625,
|
| 56 |
+
"frac_reward_zero_std": 0.2625,
|
| 57 |
+
"grad_norm": 3.8192903995513916,
|
| 58 |
+
"learning_rate": 7.031249999999999e-07,
|
| 59 |
+
"loss": 0.01356593519449234,
|
| 60 |
+
"num_tokens": 726552.0,
|
| 61 |
+
"reward": 0.09853125289082527,
|
| 62 |
+
"reward_std": 0.583341383934021,
|
| 63 |
+
"rewards/pulse_reward/mean": 0.09853125289082527,
|
| 64 |
+
"rewards/pulse_reward/std": 0.583341383934021,
|
| 65 |
+
"step": 20,
|
| 66 |
+
"step_time": 39.57589099079487,
|
| 67 |
+
"tools/call_frequency": 6.35,
|
| 68 |
+
"tools/failure_frequency": 0.005996426660567522
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"clip_ratio/high_max": 0.0,
|
| 72 |
+
"clip_ratio/high_mean": 0.0,
|
| 73 |
+
"clip_ratio/low_mean": 0.0,
|
| 74 |
+
"clip_ratio/low_min": 0.0,
|
| 75 |
+
"clip_ratio/region_mean": 0.0,
|
| 76 |
+
"completions/clipped_ratio": 0.0,
|
| 77 |
+
"completions/max_length": 273.0,
|
| 78 |
+
"completions/max_terminated_length": 273.0,
|
| 79 |
+
"completions/mean_length": 149.115625,
|
| 80 |
+
"completions/mean_terminated_length": 149.115625,
|
| 81 |
+
"completions/min_length": 50.9,
|
| 82 |
+
"completions/min_terminated_length": 50.9,
|
| 83 |
+
"entropy": 0.1797773003578186,
|
| 84 |
+
"epoch": 0.9375,
|
| 85 |
+
"frac_reward_zero_std": 0.0875,
|
| 86 |
+
"grad_norm": 5.692322254180908,
|
| 87 |
+
"learning_rate": 5.46875e-07,
|
| 88 |
+
"loss": 0.06958286762237549,
|
| 89 |
+
"num_tokens": 1082109.0,
|
| 90 |
+
"reward": 0.3879156395792961,
|
| 91 |
+
"reward_std": 0.6661841243505477,
|
| 92 |
+
"rewards/pulse_reward/mean": 0.3879156395792961,
|
| 93 |
+
"rewards/pulse_reward/std": 0.6661841526627541,
|
| 94 |
+
"step": 30,
|
| 95 |
+
"step_time": 30.443527194001945,
|
| 96 |
+
"tools/call_frequency": 5.58125,
|
| 97 |
+
"tools/failure_frequency": 0.005842319130897522
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"clip_ratio/high_max": 0.0,
|
| 101 |
+
"clip_ratio/high_mean": 0.0,
|
| 102 |
+
"clip_ratio/low_mean": 0.0,
|
| 103 |
+
"clip_ratio/low_min": 0.0,
|
| 104 |
+
"clip_ratio/region_mean": 0.0,
|
| 105 |
+
"completions/clipped_ratio": 0.0,
|
| 106 |
+
"completions/max_length": 115.3,
|
| 107 |
+
"completions/max_terminated_length": 115.3,
|
| 108 |
+
"completions/mean_length": 70.528125,
|
| 109 |
+
"completions/mean_terminated_length": 70.528125,
|
| 110 |
+
"completions/min_length": 38.1,
|
| 111 |
+
"completions/min_terminated_length": 38.1,
|
| 112 |
+
"entropy": 0.1547975329682231,
|
| 113 |
+
"epoch": 1.25,
|
| 114 |
+
"frac_reward_zero_std": 0.1625,
|
| 115 |
+
"grad_norm": 3.677471399307251,
|
| 116 |
+
"learning_rate": 3.9062499999999997e-07,
|
| 117 |
+
"loss": 0.02682075500488281,
|
| 118 |
+
"num_tokens": 1412518.0,
|
| 119 |
+
"reward": 1.0730250239372254,
|
| 120 |
+
"reward_std": 0.7086882412433624,
|
| 121 |
+
"rewards/pulse_reward/mean": 1.0730250239372254,
|
| 122 |
+
"rewards/pulse_reward/std": 0.7086882352828979,
|
| 123 |
+
"step": 40,
|
| 124 |
+
"step_time": 11.361506971600466,
|
| 125 |
+
"tools/call_frequency": 2.215625,
|
| 126 |
+
"tools/failure_frequency": 0.0015384615398943424
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"clip_ratio/high_max": 0.0,
|
| 130 |
+
"clip_ratio/high_mean": 0.0,
|
| 131 |
+
"clip_ratio/low_mean": 0.0,
|
| 132 |
+
"clip_ratio/low_min": 0.0,
|
| 133 |
+
"clip_ratio/region_mean": 0.0,
|
| 134 |
+
"completions/clipped_ratio": 0.0,
|
| 135 |
+
"completions/max_length": 83.1,
|
| 136 |
+
"completions/max_terminated_length": 83.1,
|
| 137 |
+
"completions/mean_length": 65.421875,
|
| 138 |
+
"completions/mean_terminated_length": 65.421875,
|
| 139 |
+
"completions/min_length": 48.0,
|
| 140 |
+
"completions/min_terminated_length": 48.0,
|
| 141 |
+
"entropy": 0.12452723067253828,
|
| 142 |
+
"epoch": 1.5625,
|
| 143 |
+
"frac_reward_zero_std": 0.6625,
|
| 144 |
+
"grad_norm": 3.609057664871216,
|
| 145 |
+
"learning_rate": 2.3437499999999998e-07,
|
| 146 |
+
"loss": -3.346521407365799e-05,
|
| 147 |
+
"num_tokens": 1741293.0,
|
| 148 |
+
"reward": 1.4852312803268433,
|
| 149 |
+
"reward_std": 0.3715872406959534,
|
| 150 |
+
"rewards/pulse_reward/mean": 1.4852312803268433,
|
| 151 |
+
"rewards/pulse_reward/std": 0.3715872406959534,
|
| 152 |
+
"step": 50,
|
| 153 |
+
"step_time": 8.49522676919878,
|
| 154 |
+
"tools/call_frequency": 2.0,
|
| 155 |
+
"tools/failure_frequency": 0.0
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"clip_ratio/high_max": 0.0,
|
| 159 |
+
"clip_ratio/high_mean": 0.0,
|
| 160 |
+
"clip_ratio/low_mean": 0.0,
|
| 161 |
+
"clip_ratio/low_min": 0.0,
|
| 162 |
+
"clip_ratio/region_mean": 0.0,
|
| 163 |
+
"completions/clipped_ratio": 0.0,
|
| 164 |
+
"completions/max_length": 86.6,
|
| 165 |
+
"completions/max_terminated_length": 86.6,
|
| 166 |
+
"completions/mean_length": 65.471875,
|
| 167 |
+
"completions/mean_terminated_length": 65.471875,
|
| 168 |
+
"completions/min_length": 53.6,
|
| 169 |
+
"completions/min_terminated_length": 53.6,
|
| 170 |
+
"entropy": 0.11676975060254335,
|
| 171 |
+
"epoch": 1.875,
|
| 172 |
+
"frac_reward_zero_std": 0.8125,
|
| 173 |
+
"grad_norm": 2.142204761505127,
|
| 174 |
+
"learning_rate": 7.812499999999999e-08,
|
| 175 |
+
"loss": -0.00026753861457109453,
|
| 176 |
+
"num_tokens": 2070084.0,
|
| 177 |
+
"reward": 1.538143789768219,
|
| 178 |
+
"reward_std": 0.2553505107760429,
|
| 179 |
+
"rewards/pulse_reward/mean": 1.538143789768219,
|
| 180 |
+
"rewards/pulse_reward/std": 0.25535051375627515,
|
| 181 |
+
"step": 60,
|
| 182 |
+
"step_time": 8.590878555196104,
|
| 183 |
+
"tools/call_frequency": 1.996875,
|
| 184 |
+
"tools/failure_frequency": 0.0
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"clip_ratio/high_max": 0.0,
|
| 188 |
+
"clip_ratio/high_mean": 0.0,
|
| 189 |
+
"clip_ratio/low_mean": 0.0,
|
| 190 |
+
"clip_ratio/low_min": 0.0,
|
| 191 |
+
"clip_ratio/region_mean": 0.0,
|
| 192 |
+
"completions/clipped_ratio": 0.0,
|
| 193 |
+
"completions/max_length": 74.75,
|
| 194 |
+
"completions/max_terminated_length": 74.75,
|
| 195 |
+
"completions/mean_length": 65.3203125,
|
| 196 |
+
"completions/mean_terminated_length": 65.3203125,
|
| 197 |
+
"completions/min_length": 61.25,
|
| 198 |
+
"completions/min_terminated_length": 61.25,
|
| 199 |
+
"entropy": 0.11779927462339401,
|
| 200 |
+
"epoch": 2.0,
|
| 201 |
+
"frac_reward_zero_std": 0.9375,
|
| 202 |
+
"num_tokens": 2201581.0,
|
| 203 |
+
"reward": 1.57924222946167,
|
| 204 |
+
"reward_std": 0.11742392182350159,
|
| 205 |
+
"rewards/pulse_reward/mean": 1.57924222946167,
|
| 206 |
+
"rewards/pulse_reward/std": 0.11742392182350159,
|
| 207 |
+
"step": 64,
|
| 208 |
+
"step_time": 7.905290340990177,
|
| 209 |
+
"tools/call_frequency": 2.0078125,
|
| 210 |
"tools/failure_frequency": 0.0,
|
| 211 |
"total_flos": 0.0,
|
| 212 |
+
"train_loss": 0.011853500996949151,
|
| 213 |
+
"train_runtime": 1404.4837,
|
| 214 |
+
"train_samples_per_second": 0.365,
|
| 215 |
+
"train_steps_per_second": 0.046
|
| 216 |
}
|
| 217 |
],
|
| 218 |
"logging_steps": 10,
|
| 219 |
+
"max_steps": 64,
|
| 220 |
+
"num_input_tokens_seen": 2201581,
|
| 221 |
+
"num_train_epochs": 2,
|
| 222 |
"save_steps": 500,
|
| 223 |
"stateful_callbacks": {
|
| 224 |
"TrainerControl": {
|
|
|
|
| 233 |
}
|
| 234 |
},
|
| 235 |
"total_flos": 0.0,
|
| 236 |
+
"train_batch_size": 8,
|
| 237 |
"trial_name": null,
|
| 238 |
"trial_params": null
|
| 239 |
}
|