Intermediate checkpoint upload step=90 (generator_train)
Browse files- .gitattributes +1 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/chat_template.jinja +54 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/config.json +57 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/generation_config.json +13 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/model.safetensors +3 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/optimizer.pt +3 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/rng_state.pth +3 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/scheduler.pt +3 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/tokenizer.json +3 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/tokenizer_config.json +32 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/trainer_state.json +1683 -0
- self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -74,3 +74,4 @@ self_play_hf_l40s_full/round_003/answerer_train/checkpoint-120/tokenizer.json fi
|
|
| 74 |
self_play_hf_l40s_full/round_003/answerer_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 75 |
self_play_hf_l40s_full/round_004/generator_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 76 |
self_play_hf_l40s_full/round_004/generator_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 74 |
self_play_hf_l40s_full/round_003/answerer_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 75 |
self_play_hf_l40s_full/round_004/generator_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 76 |
self_play_hf_l40s_full/round_004/generator_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 77 |
+
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen2ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"bos_token_id": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 151645,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 896,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 4864,
|
| 13 |
+
"layer_types": [
|
| 14 |
+
"full_attention",
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention"
|
| 38 |
+
],
|
| 39 |
+
"max_position_embeddings": 32768,
|
| 40 |
+
"max_window_layers": 21,
|
| 41 |
+
"model_type": "qwen2",
|
| 42 |
+
"num_attention_heads": 14,
|
| 43 |
+
"num_hidden_layers": 24,
|
| 44 |
+
"num_key_value_heads": 2,
|
| 45 |
+
"pad_token_id": 151643,
|
| 46 |
+
"rms_norm_eps": 1e-06,
|
| 47 |
+
"rope_parameters": {
|
| 48 |
+
"rope_theta": 1000000.0,
|
| 49 |
+
"rope_type": "default"
|
| 50 |
+
},
|
| 51 |
+
"sliding_window": null,
|
| 52 |
+
"tie_word_embeddings": true,
|
| 53 |
+
"transformers_version": "5.6.2",
|
| 54 |
+
"use_cache": false,
|
| 55 |
+
"use_sliding_window": false,
|
| 56 |
+
"vocab_size": 151936
|
| 57 |
+
}
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/generation_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
151645,
|
| 5 |
+
151643
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 151643,
|
| 8 |
+
"repetition_penalty": 1.1,
|
| 9 |
+
"temperature": 0.7,
|
| 10 |
+
"top_k": 20,
|
| 11 |
+
"top_p": 0.8,
|
| 12 |
+
"transformers_version": "5.6.2"
|
| 13 |
+
}
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:975134b72d4c339d56b813cc9b91b93bcc14b8b494b4c6acf6d21fc521a0066f
|
| 3 |
+
size 1976163472
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:968c38c269e5efe94e7a2e304c381a48c9b1ed338e0044bc1e000ca8819d4a5c
|
| 3 |
+
size 3952509771
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b5001a1413e890182ec459c1edbf745017684c89f471fbb0cafcaadbc1070a9f
|
| 3 |
+
size 14645
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9466ecaa07ea9f7aa3b70a452615802059d7a87e6220a892a7f18c7703f928c0
|
| 3 |
+
size 1465
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
|
| 3 |
+
size 11421892
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/tokenizer_config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": true,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 131072,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"padding_side": "left",
|
| 28 |
+
"split_special_tokens": false,
|
| 29 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 30 |
+
"truncation_side": "left",
|
| 31 |
+
"unk_token": null
|
| 32 |
+
}
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/trainer_state.json
ADDED
|
@@ -0,0 +1,1683 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.75,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 90,
|
| 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": 1.0,
|
| 19 |
+
"completions/max_length": 384.0,
|
| 20 |
+
"completions/max_terminated_length": 0.0,
|
| 21 |
+
"completions/mean_length": 384.0,
|
| 22 |
+
"completions/mean_terminated_length": 0.0,
|
| 23 |
+
"completions/min_length": 384.0,
|
| 24 |
+
"completions/min_terminated_length": 0.0,
|
| 25 |
+
"entropy": 3.7540836334228516,
|
| 26 |
+
"epoch": 0.041666666666666664,
|
| 27 |
+
"frac_reward_zero_std": 0.5,
|
| 28 |
+
"grad_norm": 2.3061139583587646,
|
| 29 |
+
"kl": 0.0004934151656925678,
|
| 30 |
+
"learning_rate": 5e-06,
|
| 31 |
+
"loss": 0.062387701123952866,
|
| 32 |
+
"num_tokens": 25136.0,
|
| 33 |
+
"reward": -0.054375000298023224,
|
| 34 |
+
"reward_std": 0.01249999925494194,
|
| 35 |
+
"rewards/GeneratorRewardFunction/mean": -0.054375000298023224,
|
| 36 |
+
"rewards/GeneratorRewardFunction/std": 0.012500000186264515,
|
| 37 |
+
"step": 1,
|
| 38 |
+
"step_time": 12.28940931599982
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"clip_ratio/high_max": 0.03515625,
|
| 42 |
+
"clip_ratio/high_mean": 0.03515625,
|
| 43 |
+
"clip_ratio/low_mean": 0.0,
|
| 44 |
+
"clip_ratio/low_min": 0.0,
|
| 45 |
+
"clip_ratio/region_mean": 0.03515625,
|
| 46 |
+
"entropy": 3.8097915649414062,
|
| 47 |
+
"epoch": 0.08333333333333333,
|
| 48 |
+
"grad_norm": 1.1211391687393188,
|
| 49 |
+
"kl": 0.017926404252648354,
|
| 50 |
+
"learning_rate": 4.958333333333334e-06,
|
| 51 |
+
"loss": -0.05972687527537346,
|
| 52 |
+
"step": 2,
|
| 53 |
+
"step_time": 0.22016475100008392
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"clip_ratio/high_max": 0.0,
|
| 57 |
+
"clip_ratio/high_mean": 0.0,
|
| 58 |
+
"clip_ratio/low_mean": 0.0559895820915699,
|
| 59 |
+
"clip_ratio/low_min": 0.0559895820915699,
|
| 60 |
+
"clip_ratio/region_mean": 0.0559895820915699,
|
| 61 |
+
"entropy": 3.656827688217163,
|
| 62 |
+
"epoch": 0.125,
|
| 63 |
+
"grad_norm": 2.1082754135131836,
|
| 64 |
+
"kl": 0.025636015459895134,
|
| 65 |
+
"learning_rate": 4.9166666666666665e-06,
|
| 66 |
+
"loss": 0.09194447845220566,
|
| 67 |
+
"step": 3,
|
| 68 |
+
"step_time": 0.21996182600014436
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"clip_ratio/high_max": 0.0963541641831398,
|
| 72 |
+
"clip_ratio/high_mean": 0.0963541641831398,
|
| 73 |
+
"clip_ratio/low_mean": 0.0,
|
| 74 |
+
"clip_ratio/low_min": 0.0,
|
| 75 |
+
"clip_ratio/region_mean": 0.0963541641831398,
|
| 76 |
+
"entropy": 3.8841819763183594,
|
| 77 |
+
"epoch": 0.16666666666666666,
|
| 78 |
+
"grad_norm": 1.4164149761199951,
|
| 79 |
+
"kl": 0.030918851494789124,
|
| 80 |
+
"learning_rate": 4.875e-06,
|
| 81 |
+
"loss": -0.08998001366853714,
|
| 82 |
+
"step": 4,
|
| 83 |
+
"step_time": 0.22261608200005867
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"clip_ratio/high_max": 0.0,
|
| 87 |
+
"clip_ratio/high_mean": 0.0,
|
| 88 |
+
"clip_ratio/low_mean": 0.0,
|
| 89 |
+
"clip_ratio/low_min": 0.0,
|
| 90 |
+
"clip_ratio/region_mean": 0.0,
|
| 91 |
+
"completions/clipped_ratio": 1.0,
|
| 92 |
+
"completions/max_length": 384.0,
|
| 93 |
+
"completions/max_terminated_length": 0.0,
|
| 94 |
+
"completions/mean_length": 384.0,
|
| 95 |
+
"completions/mean_terminated_length": 0.0,
|
| 96 |
+
"completions/min_length": 384.0,
|
| 97 |
+
"completions/min_terminated_length": 0.0,
|
| 98 |
+
"entropy": 3.838679075241089,
|
| 99 |
+
"epoch": 0.20833333333333334,
|
| 100 |
+
"frac_reward_zero_std": 1.0,
|
| 101 |
+
"grad_norm": 0.019757986068725586,
|
| 102 |
+
"kl": 0.038326989859342575,
|
| 103 |
+
"learning_rate": 4.833333333333333e-06,
|
| 104 |
+
"loss": 9.581747144693509e-05,
|
| 105 |
+
"num_tokens": 50436.0,
|
| 106 |
+
"reward": -0.05000000074505806,
|
| 107 |
+
"reward_std": 0.0,
|
| 108 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 109 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 110 |
+
"step": 5,
|
| 111 |
+
"step_time": 12.358392488999925
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"clip_ratio/high_max": 0.0,
|
| 115 |
+
"clip_ratio/high_mean": 0.0,
|
| 116 |
+
"clip_ratio/low_mean": 0.0,
|
| 117 |
+
"clip_ratio/low_min": 0.0,
|
| 118 |
+
"clip_ratio/region_mean": 0.0,
|
| 119 |
+
"entropy": 3.878758192062378,
|
| 120 |
+
"epoch": 0.25,
|
| 121 |
+
"grad_norm": 0.02407441847026348,
|
| 122 |
+
"kl": 0.047756802290678024,
|
| 123 |
+
"learning_rate": 4.791666666666668e-06,
|
| 124 |
+
"loss": 0.00011939200339838862,
|
| 125 |
+
"step": 6,
|
| 126 |
+
"step_time": 0.22772415899999032
|
| 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 |
+
"entropy": 4.379356861114502,
|
| 135 |
+
"epoch": 0.2916666666666667,
|
| 136 |
+
"grad_norm": 0.021465642377734184,
|
| 137 |
+
"kl": 0.04809901490807533,
|
| 138 |
+
"learning_rate": 4.75e-06,
|
| 139 |
+
"loss": 0.00012024753959849477,
|
| 140 |
+
"step": 7,
|
| 141 |
+
"step_time": 0.22733985799959555
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"clip_ratio/high_max": 0.0,
|
| 145 |
+
"clip_ratio/high_mean": 0.0,
|
| 146 |
+
"clip_ratio/low_mean": 0.0,
|
| 147 |
+
"clip_ratio/low_min": 0.0,
|
| 148 |
+
"clip_ratio/region_mean": 0.0,
|
| 149 |
+
"entropy": 4.279653072357178,
|
| 150 |
+
"epoch": 0.3333333333333333,
|
| 151 |
+
"grad_norm": 0.0189784225076437,
|
| 152 |
+
"kl": 0.04601867124438286,
|
| 153 |
+
"learning_rate": 4.708333333333334e-06,
|
| 154 |
+
"loss": 0.00011504666326800361,
|
| 155 |
+
"step": 8,
|
| 156 |
+
"step_time": 0.22832407399982912
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"clip_ratio/high_max": 0.0,
|
| 160 |
+
"clip_ratio/high_mean": 0.0,
|
| 161 |
+
"clip_ratio/low_mean": 0.0,
|
| 162 |
+
"clip_ratio/low_min": 0.0,
|
| 163 |
+
"clip_ratio/region_mean": 0.0,
|
| 164 |
+
"completions/clipped_ratio": 1.0,
|
| 165 |
+
"completions/max_length": 384.0,
|
| 166 |
+
"completions/max_terminated_length": 0.0,
|
| 167 |
+
"completions/mean_length": 384.0,
|
| 168 |
+
"completions/mean_terminated_length": 0.0,
|
| 169 |
+
"completions/min_length": 384.0,
|
| 170 |
+
"completions/min_terminated_length": 0.0,
|
| 171 |
+
"entropy": 3.9055023193359375,
|
| 172 |
+
"epoch": 0.375,
|
| 173 |
+
"frac_reward_zero_std": 0.5,
|
| 174 |
+
"grad_norm": 2.340376377105713,
|
| 175 |
+
"kl": 0.06852344423532486,
|
| 176 |
+
"learning_rate": 4.666666666666667e-06,
|
| 177 |
+
"loss": 0.09388399124145508,
|
| 178 |
+
"num_tokens": 75808.0,
|
| 179 |
+
"reward": -0.11156250536441803,
|
| 180 |
+
"reward_std": 0.16717973351478577,
|
| 181 |
+
"rewards/GeneratorRewardFunction/mean": -0.11156250536441803,
|
| 182 |
+
"rewards/GeneratorRewardFunction/std": 0.16717973351478577,
|
| 183 |
+
"step": 9,
|
| 184 |
+
"step_time": 12.377079298000353
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"clip_ratio/high_max": 0.0013020833721384406,
|
| 188 |
+
"clip_ratio/high_mean": 0.0013020833721384406,
|
| 189 |
+
"clip_ratio/low_mean": 0.001953125,
|
| 190 |
+
"clip_ratio/low_min": 0.001953125,
|
| 191 |
+
"clip_ratio/region_mean": 0.0032552082557231188,
|
| 192 |
+
"entropy": 4.169831275939941,
|
| 193 |
+
"epoch": 0.4166666666666667,
|
| 194 |
+
"grad_norm": 2.4143083095550537,
|
| 195 |
+
"kl": 0.08672171831130981,
|
| 196 |
+
"learning_rate": 4.625000000000001e-06,
|
| 197 |
+
"loss": 0.06073353812098503,
|
| 198 |
+
"step": 10,
|
| 199 |
+
"step_time": 0.2275676320000457
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"clip_ratio/high_max": 0.017578125,
|
| 203 |
+
"clip_ratio/high_mean": 0.017578125,
|
| 204 |
+
"clip_ratio/low_mean": 0.0,
|
| 205 |
+
"clip_ratio/low_min": 0.0,
|
| 206 |
+
"clip_ratio/region_mean": 0.017578125,
|
| 207 |
+
"entropy": 4.554050445556641,
|
| 208 |
+
"epoch": 0.4583333333333333,
|
| 209 |
+
"grad_norm": 1.0698223114013672,
|
| 210 |
+
"kl": 0.07464811205863953,
|
| 211 |
+
"learning_rate": 4.583333333333333e-06,
|
| 212 |
+
"loss": -0.06039423868060112,
|
| 213 |
+
"step": 11,
|
| 214 |
+
"step_time": 0.22664018099976602
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"clip_ratio/high_max": 0.060546875,
|
| 218 |
+
"clip_ratio/high_mean": 0.060546875,
|
| 219 |
+
"clip_ratio/low_mean": 0.0,
|
| 220 |
+
"clip_ratio/low_min": 0.0,
|
| 221 |
+
"clip_ratio/region_mean": 0.060546875,
|
| 222 |
+
"entropy": 4.123767375946045,
|
| 223 |
+
"epoch": 0.5,
|
| 224 |
+
"grad_norm": 1.3299524784088135,
|
| 225 |
+
"kl": 0.11542084068059921,
|
| 226 |
+
"learning_rate": 4.541666666666667e-06,
|
| 227 |
+
"loss": -0.09344702959060669,
|
| 228 |
+
"step": 12,
|
| 229 |
+
"step_time": 0.22850085099980788
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"clip_ratio/high_max": 0.0,
|
| 233 |
+
"clip_ratio/high_mean": 0.0,
|
| 234 |
+
"clip_ratio/low_mean": 0.0,
|
| 235 |
+
"clip_ratio/low_min": 0.0,
|
| 236 |
+
"clip_ratio/region_mean": 0.0,
|
| 237 |
+
"completions/clipped_ratio": 1.0,
|
| 238 |
+
"completions/max_length": 384.0,
|
| 239 |
+
"completions/max_terminated_length": 0.0,
|
| 240 |
+
"completions/mean_length": 384.0,
|
| 241 |
+
"completions/mean_terminated_length": 0.0,
|
| 242 |
+
"completions/min_length": 384.0,
|
| 243 |
+
"completions/min_terminated_length": 0.0,
|
| 244 |
+
"entropy": 4.361082553863525,
|
| 245 |
+
"epoch": 0.5416666666666666,
|
| 246 |
+
"frac_reward_zero_std": 0.5,
|
| 247 |
+
"grad_norm": 2.413485288619995,
|
| 248 |
+
"kl": 0.09952286630868912,
|
| 249 |
+
"learning_rate": 4.5e-06,
|
| 250 |
+
"loss": 0.062085848301649094,
|
| 251 |
+
"num_tokens": 100880.0,
|
| 252 |
+
"reward": -0.08906249701976776,
|
| 253 |
+
"reward_std": 0.14971326291561127,
|
| 254 |
+
"rewards/GeneratorRewardFunction/mean": -0.08906249701976776,
|
| 255 |
+
"rewards/GeneratorRewardFunction/std": 0.14971326291561127,
|
| 256 |
+
"step": 13,
|
| 257 |
+
"step_time": 12.362482152999746
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"clip_ratio/high_max": 0.0045572915114462376,
|
| 261 |
+
"clip_ratio/high_mean": 0.0045572915114462376,
|
| 262 |
+
"clip_ratio/low_mean": 0.0013020833721384406,
|
| 263 |
+
"clip_ratio/low_min": 0.0013020833721384406,
|
| 264 |
+
"clip_ratio/region_mean": 0.005859375,
|
| 265 |
+
"entropy": 4.488595485687256,
|
| 266 |
+
"epoch": 0.5833333333333334,
|
| 267 |
+
"grad_norm": 2.6962649822235107,
|
| 268 |
+
"kl": 0.10717153549194336,
|
| 269 |
+
"learning_rate": 4.4583333333333336e-06,
|
| 270 |
+
"loss": 0.0006585324299521744,
|
| 271 |
+
"step": 14,
|
| 272 |
+
"step_time": 0.2181926339999336
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"clip_ratio/high_max": 0.01888020895421505,
|
| 276 |
+
"clip_ratio/high_mean": 0.01888020895421505,
|
| 277 |
+
"clip_ratio/low_mean": 0.0,
|
| 278 |
+
"clip_ratio/low_min": 0.0,
|
| 279 |
+
"clip_ratio/region_mean": 0.01888020895421505,
|
| 280 |
+
"entropy": 4.414994716644287,
|
| 281 |
+
"epoch": 0.625,
|
| 282 |
+
"grad_norm": 1.1328541040420532,
|
| 283 |
+
"kl": 0.09972727298736572,
|
| 284 |
+
"learning_rate": 4.416666666666667e-06,
|
| 285 |
+
"loss": -0.0626036524772644,
|
| 286 |
+
"step": 15,
|
| 287 |
+
"step_time": 0.21817723499998465
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"clip_ratio/high_max": 0.0,
|
| 291 |
+
"clip_ratio/high_mean": 0.0,
|
| 292 |
+
"clip_ratio/low_mean": 0.0,
|
| 293 |
+
"clip_ratio/low_min": 0.0,
|
| 294 |
+
"clip_ratio/region_mean": 0.0,
|
| 295 |
+
"entropy": 4.725884437561035,
|
| 296 |
+
"epoch": 0.6666666666666666,
|
| 297 |
+
"grad_norm": 0.02289302460849285,
|
| 298 |
+
"kl": 0.08768561482429504,
|
| 299 |
+
"learning_rate": 4.3750000000000005e-06,
|
| 300 |
+
"loss": 0.00021921405277680606,
|
| 301 |
+
"step": 16,
|
| 302 |
+
"step_time": 0.21883753800011618
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"clip_ratio/high_max": 0.0,
|
| 306 |
+
"clip_ratio/high_mean": 0.0,
|
| 307 |
+
"clip_ratio/low_mean": 0.0,
|
| 308 |
+
"clip_ratio/low_min": 0.0,
|
| 309 |
+
"clip_ratio/region_mean": 0.0,
|
| 310 |
+
"completions/clipped_ratio": 1.0,
|
| 311 |
+
"completions/max_length": 384.0,
|
| 312 |
+
"completions/max_terminated_length": 0.0,
|
| 313 |
+
"completions/mean_length": 384.0,
|
| 314 |
+
"completions/mean_terminated_length": 0.0,
|
| 315 |
+
"completions/min_length": 384.0,
|
| 316 |
+
"completions/min_terminated_length": 0.0,
|
| 317 |
+
"entropy": 4.040867328643799,
|
| 318 |
+
"epoch": 0.7083333333333334,
|
| 319 |
+
"frac_reward_zero_std": 0.5,
|
| 320 |
+
"grad_norm": 2.39467191696167,
|
| 321 |
+
"kl": 0.1182798221707344,
|
| 322 |
+
"learning_rate": 4.333333333333334e-06,
|
| 323 |
+
"loss": 0.09359735995531082,
|
| 324 |
+
"num_tokens": 125920.0,
|
| 325 |
+
"reward": -0.05625000223517418,
|
| 326 |
+
"reward_std": 0.019364915788173676,
|
| 327 |
+
"rewards/GeneratorRewardFunction/mean": -0.05625000223517418,
|
| 328 |
+
"rewards/GeneratorRewardFunction/std": 0.019364915788173676,
|
| 329 |
+
"step": 17,
|
| 330 |
+
"step_time": 12.371318035000058
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"clip_ratio/high_max": 0.0013020833721384406,
|
| 334 |
+
"clip_ratio/high_mean": 0.0013020833721384406,
|
| 335 |
+
"clip_ratio/low_mean": 0.0,
|
| 336 |
+
"clip_ratio/low_min": 0.0,
|
| 337 |
+
"clip_ratio/region_mean": 0.0013020833721384406,
|
| 338 |
+
"entropy": 4.358741760253906,
|
| 339 |
+
"epoch": 0.75,
|
| 340 |
+
"grad_norm": 1.4153165817260742,
|
| 341 |
+
"kl": 0.1291649043560028,
|
| 342 |
+
"learning_rate": 4.2916666666666665e-06,
|
| 343 |
+
"loss": -0.09299901872873306,
|
| 344 |
+
"step": 18,
|
| 345 |
+
"step_time": 0.220845523999742
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"clip_ratio/high_max": 0.01171875,
|
| 349 |
+
"clip_ratio/high_mean": 0.01171875,
|
| 350 |
+
"clip_ratio/low_mean": 0.0,
|
| 351 |
+
"clip_ratio/low_min": 0.0,
|
| 352 |
+
"clip_ratio/region_mean": 0.01171875,
|
| 353 |
+
"entropy": 4.570446014404297,
|
| 354 |
+
"epoch": 0.7916666666666666,
|
| 355 |
+
"grad_norm": 1.1514215469360352,
|
| 356 |
+
"kl": 0.11547613888978958,
|
| 357 |
+
"learning_rate": 4.25e-06,
|
| 358 |
+
"loss": -0.061982929706573486,
|
| 359 |
+
"step": 19,
|
| 360 |
+
"step_time": 0.2208292430000256
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"clip_ratio/high_max": 0.01627604104578495,
|
| 364 |
+
"clip_ratio/high_mean": 0.01627604104578495,
|
| 365 |
+
"clip_ratio/low_mean": 0.011067708022892475,
|
| 366 |
+
"clip_ratio/low_min": 0.011067708022892475,
|
| 367 |
+
"clip_ratio/region_mean": 0.02734375,
|
| 368 |
+
"entropy": 3.9513370990753174,
|
| 369 |
+
"epoch": 0.8333333333333334,
|
| 370 |
+
"grad_norm": 2.3550095558166504,
|
| 371 |
+
"kl": 0.1719018667936325,
|
| 372 |
+
"learning_rate": 4.208333333333333e-06,
|
| 373 |
+
"loss": 0.06247749552130699,
|
| 374 |
+
"step": 20,
|
| 375 |
+
"step_time": 0.22178658199982237
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"clip_ratio/high_max": 0.0,
|
| 379 |
+
"clip_ratio/high_mean": 0.0,
|
| 380 |
+
"clip_ratio/low_mean": 0.0,
|
| 381 |
+
"clip_ratio/low_min": 0.0,
|
| 382 |
+
"clip_ratio/region_mean": 0.0,
|
| 383 |
+
"completions/clipped_ratio": 1.0,
|
| 384 |
+
"completions/max_length": 384.0,
|
| 385 |
+
"completions/max_terminated_length": 0.0,
|
| 386 |
+
"completions/mean_length": 384.0,
|
| 387 |
+
"completions/mean_terminated_length": 0.0,
|
| 388 |
+
"completions/min_length": 384.0,
|
| 389 |
+
"completions/min_terminated_length": 0.0,
|
| 390 |
+
"entropy": 3.850313901901245,
|
| 391 |
+
"epoch": 0.875,
|
| 392 |
+
"frac_reward_zero_std": 0.0,
|
| 393 |
+
"grad_norm": 2.7302346229553223,
|
| 394 |
+
"kl": 0.1612752377986908,
|
| 395 |
+
"learning_rate": 4.166666666666667e-06,
|
| 396 |
+
"loss": -0.10781875252723694,
|
| 397 |
+
"num_tokens": 150972.0,
|
| 398 |
+
"reward": -0.27531248331069946,
|
| 399 |
+
"reward_std": 0.29975247383117676,
|
| 400 |
+
"rewards/GeneratorRewardFunction/mean": -0.27531248331069946,
|
| 401 |
+
"rewards/GeneratorRewardFunction/std": 0.29975250363349915,
|
| 402 |
+
"step": 21,
|
| 403 |
+
"step_time": 12.298712889999933
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"clip_ratio/high_max": 0.0052083334885537624,
|
| 407 |
+
"clip_ratio/high_mean": 0.0052083334885537624,
|
| 408 |
+
"clip_ratio/low_mean": 0.0,
|
| 409 |
+
"clip_ratio/low_min": 0.0,
|
| 410 |
+
"clip_ratio/region_mean": 0.0052083334885537624,
|
| 411 |
+
"entropy": 3.952542543411255,
|
| 412 |
+
"epoch": 0.9166666666666666,
|
| 413 |
+
"grad_norm": 2.175267457962036,
|
| 414 |
+
"kl": 0.15720854699611664,
|
| 415 |
+
"learning_rate": 4.125e-06,
|
| 416 |
+
"loss": -0.061483949422836304,
|
| 417 |
+
"step": 22,
|
| 418 |
+
"step_time": 0.2193120709998766
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"clip_ratio/high_max": 0.0032552082557231188,
|
| 422 |
+
"clip_ratio/high_mean": 0.0032552082557231188,
|
| 423 |
+
"clip_ratio/low_mean": 0.009765625,
|
| 424 |
+
"clip_ratio/low_min": 0.009765625,
|
| 425 |
+
"clip_ratio/region_mean": 0.013020833022892475,
|
| 426 |
+
"entropy": 4.322933673858643,
|
| 427 |
+
"epoch": 0.9583333333333334,
|
| 428 |
+
"grad_norm": 3.3126373291015625,
|
| 429 |
+
"kl": 0.1569279283285141,
|
| 430 |
+
"learning_rate": 4.083333333333334e-06,
|
| 431 |
+
"loss": 0.17323081195354462,
|
| 432 |
+
"step": 23,
|
| 433 |
+
"step_time": 0.21882151099998737
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"clip_ratio/high_max": 0.0572916679084301,
|
| 437 |
+
"clip_ratio/high_mean": 0.0572916679084301,
|
| 438 |
+
"clip_ratio/low_mean": 0.0078125,
|
| 439 |
+
"clip_ratio/low_min": 0.0078125,
|
| 440 |
+
"clip_ratio/region_mean": 0.0651041641831398,
|
| 441 |
+
"entropy": 4.215339660644531,
|
| 442 |
+
"epoch": 1.0,
|
| 443 |
+
"grad_norm": 3.0476796627044678,
|
| 444 |
+
"kl": 0.197212815284729,
|
| 445 |
+
"learning_rate": 4.041666666666667e-06,
|
| 446 |
+
"loss": 0.0013918876647949219,
|
| 447 |
+
"step": 24,
|
| 448 |
+
"step_time": 0.22078252000028442
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"clip_ratio/high_max": 0.0,
|
| 452 |
+
"clip_ratio/high_mean": 0.0,
|
| 453 |
+
"clip_ratio/low_mean": 0.0,
|
| 454 |
+
"clip_ratio/low_min": 0.0,
|
| 455 |
+
"clip_ratio/region_mean": 0.0,
|
| 456 |
+
"completions/clipped_ratio": 1.0,
|
| 457 |
+
"completions/max_length": 384.0,
|
| 458 |
+
"completions/max_terminated_length": 0.0,
|
| 459 |
+
"completions/mean_length": 384.0,
|
| 460 |
+
"completions/mean_terminated_length": 0.0,
|
| 461 |
+
"completions/min_length": 384.0,
|
| 462 |
+
"completions/min_terminated_length": 0.0,
|
| 463 |
+
"entropy": 4.335132598876953,
|
| 464 |
+
"epoch": 1.0416666666666667,
|
| 465 |
+
"frac_reward_zero_std": 0.25,
|
| 466 |
+
"grad_norm": 2.326383113861084,
|
| 467 |
+
"kl": 0.16707448661327362,
|
| 468 |
+
"learning_rate": 4.000000000000001e-06,
|
| 469 |
+
"loss": 0.05758856609463692,
|
| 470 |
+
"num_tokens": 176436.0,
|
| 471 |
+
"reward": -0.12687499821186066,
|
| 472 |
+
"reward_std": 0.20430266857147217,
|
| 473 |
+
"rewards/GeneratorRewardFunction/mean": -0.12687499821186066,
|
| 474 |
+
"rewards/GeneratorRewardFunction/std": 0.20430268347263336,
|
| 475 |
+
"step": 25,
|
| 476 |
+
"step_time": 12.420606664000388
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"clip_ratio/high_max": 0.0032552082557231188,
|
| 480 |
+
"clip_ratio/high_mean": 0.0032552082557231188,
|
| 481 |
+
"clip_ratio/low_mean": 0.0,
|
| 482 |
+
"clip_ratio/low_min": 0.0,
|
| 483 |
+
"clip_ratio/region_mean": 0.0032552082557231188,
|
| 484 |
+
"entropy": 4.311023712158203,
|
| 485 |
+
"epoch": 1.0833333333333333,
|
| 486 |
+
"grad_norm": 2.787942409515381,
|
| 487 |
+
"kl": 0.1736934781074524,
|
| 488 |
+
"learning_rate": 3.958333333333333e-06,
|
| 489 |
+
"loss": 0.0001516403426649049,
|
| 490 |
+
"step": 26,
|
| 491 |
+
"step_time": 0.2286363840003105
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"clip_ratio/high_max": 0.009114583022892475,
|
| 495 |
+
"clip_ratio/high_mean": 0.009114583022892475,
|
| 496 |
+
"clip_ratio/low_mean": 0.0032552082557231188,
|
| 497 |
+
"clip_ratio/low_min": 0.0032552082557231188,
|
| 498 |
+
"clip_ratio/region_mean": 0.012369791977107525,
|
| 499 |
+
"entropy": 4.165957927703857,
|
| 500 |
+
"epoch": 1.125,
|
| 501 |
+
"grad_norm": 2.553452968597412,
|
| 502 |
+
"kl": 0.2077862024307251,
|
| 503 |
+
"learning_rate": 3.916666666666667e-06,
|
| 504 |
+
"loss": 0.035733964294195175,
|
| 505 |
+
"step": 27,
|
| 506 |
+
"step_time": 0.22833154999989347
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"clip_ratio/high_max": 0.0358072929084301,
|
| 510 |
+
"clip_ratio/high_mean": 0.0358072929084301,
|
| 511 |
+
"clip_ratio/low_mean": 0.0,
|
| 512 |
+
"clip_ratio/low_min": 0.0,
|
| 513 |
+
"clip_ratio/region_mean": 0.0358072929084301,
|
| 514 |
+
"entropy": 4.27438497543335,
|
| 515 |
+
"epoch": 1.1666666666666667,
|
| 516 |
+
"grad_norm": 1.3039305210113525,
|
| 517 |
+
"kl": 0.1941693276166916,
|
| 518 |
+
"learning_rate": 3.875e-06,
|
| 519 |
+
"loss": -0.09091150015592575,
|
| 520 |
+
"step": 28,
|
| 521 |
+
"step_time": 0.22796160400048393
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"clip_ratio/high_max": 0.0,
|
| 525 |
+
"clip_ratio/high_mean": 0.0,
|
| 526 |
+
"clip_ratio/low_mean": 0.0,
|
| 527 |
+
"clip_ratio/low_min": 0.0,
|
| 528 |
+
"clip_ratio/region_mean": 0.0,
|
| 529 |
+
"completions/clipped_ratio": 1.0,
|
| 530 |
+
"completions/max_length": 384.0,
|
| 531 |
+
"completions/max_terminated_length": 0.0,
|
| 532 |
+
"completions/mean_length": 384.0,
|
| 533 |
+
"completions/mean_terminated_length": 0.0,
|
| 534 |
+
"completions/min_length": 384.0,
|
| 535 |
+
"completions/min_terminated_length": 0.0,
|
| 536 |
+
"entropy": 4.181336879730225,
|
| 537 |
+
"epoch": 1.2083333333333333,
|
| 538 |
+
"frac_reward_zero_std": 0.75,
|
| 539 |
+
"grad_norm": 1.4635995626449585,
|
| 540 |
+
"kl": 0.17187662422657013,
|
| 541 |
+
"learning_rate": 3.833333333333334e-06,
|
| 542 |
+
"loss": 0.05418387055397034,
|
| 543 |
+
"num_tokens": 201524.0,
|
| 544 |
+
"reward": -0.05312500149011612,
|
| 545 |
+
"reward_std": 0.008539125323295593,
|
| 546 |
+
"rewards/GeneratorRewardFunction/mean": -0.05312500149011612,
|
| 547 |
+
"rewards/GeneratorRewardFunction/std": 0.008539126254618168,
|
| 548 |
+
"step": 29,
|
| 549 |
+
"step_time": 12.34016204099953
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"clip_ratio/high_max": 0.0,
|
| 553 |
+
"clip_ratio/high_mean": 0.0,
|
| 554 |
+
"clip_ratio/low_mean": 0.0,
|
| 555 |
+
"clip_ratio/low_min": 0.0,
|
| 556 |
+
"clip_ratio/region_mean": 0.0,
|
| 557 |
+
"entropy": 4.342355251312256,
|
| 558 |
+
"epoch": 1.25,
|
| 559 |
+
"grad_norm": 1.3520798683166504,
|
| 560 |
+
"kl": 0.19560372829437256,
|
| 561 |
+
"learning_rate": 3.7916666666666666e-06,
|
| 562 |
+
"loss": -0.05310399457812309,
|
| 563 |
+
"step": 30,
|
| 564 |
+
"step_time": 0.22077936900041095
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"clip_ratio/high_max": 0.0,
|
| 568 |
+
"clip_ratio/high_mean": 0.0,
|
| 569 |
+
"clip_ratio/low_mean": 0.0,
|
| 570 |
+
"clip_ratio/low_min": 0.0,
|
| 571 |
+
"clip_ratio/region_mean": 0.0,
|
| 572 |
+
"entropy": 4.002376079559326,
|
| 573 |
+
"epoch": 1.2916666666666667,
|
| 574 |
+
"grad_norm": 0.03668312355875969,
|
| 575 |
+
"kl": 0.18400262296199799,
|
| 576 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 577 |
+
"loss": 0.0004600065294653177,
|
| 578 |
+
"step": 31,
|
| 579 |
+
"step_time": 0.23486327800037543
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"clip_ratio/high_max": 0.0078125,
|
| 583 |
+
"clip_ratio/high_mean": 0.0078125,
|
| 584 |
+
"clip_ratio/low_mean": 0.00390625,
|
| 585 |
+
"clip_ratio/low_min": 0.00390625,
|
| 586 |
+
"clip_ratio/region_mean": 0.01171875,
|
| 587 |
+
"entropy": 4.291225433349609,
|
| 588 |
+
"epoch": 1.3333333333333333,
|
| 589 |
+
"grad_norm": 1.7566320896148682,
|
| 590 |
+
"kl": 0.17402441799640656,
|
| 591 |
+
"learning_rate": 3.708333333333334e-06,
|
| 592 |
+
"loss": 0.0007328187930397689,
|
| 593 |
+
"step": 32,
|
| 594 |
+
"step_time": 0.21991783900011797
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"clip_ratio/high_max": 0.0,
|
| 598 |
+
"clip_ratio/high_mean": 0.0,
|
| 599 |
+
"clip_ratio/low_mean": 0.0,
|
| 600 |
+
"clip_ratio/low_min": 0.0,
|
| 601 |
+
"clip_ratio/region_mean": 0.0,
|
| 602 |
+
"completions/clipped_ratio": 1.0,
|
| 603 |
+
"completions/max_length": 384.0,
|
| 604 |
+
"completions/max_terminated_length": 0.0,
|
| 605 |
+
"completions/mean_length": 384.0,
|
| 606 |
+
"completions/mean_terminated_length": 0.0,
|
| 607 |
+
"completions/min_length": 384.0,
|
| 608 |
+
"completions/min_terminated_length": 0.0,
|
| 609 |
+
"entropy": 4.679595470428467,
|
| 610 |
+
"epoch": 1.375,
|
| 611 |
+
"frac_reward_zero_std": 1.0,
|
| 612 |
+
"grad_norm": 0.03962837904691696,
|
| 613 |
+
"kl": 0.18714486062526703,
|
| 614 |
+
"learning_rate": 3.6666666666666666e-06,
|
| 615 |
+
"loss": 0.00046786211896687746,
|
| 616 |
+
"num_tokens": 226532.0,
|
| 617 |
+
"reward": -0.05000000074505806,
|
| 618 |
+
"reward_std": 0.0,
|
| 619 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 620 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 621 |
+
"step": 33,
|
| 622 |
+
"step_time": 12.39098299699981
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"clip_ratio/high_max": 0.0,
|
| 626 |
+
"clip_ratio/high_mean": 0.0,
|
| 627 |
+
"clip_ratio/low_mean": 0.0,
|
| 628 |
+
"clip_ratio/low_min": 0.0,
|
| 629 |
+
"clip_ratio/region_mean": 0.0,
|
| 630 |
+
"entropy": 4.106312274932861,
|
| 631 |
+
"epoch": 1.4166666666666667,
|
| 632 |
+
"grad_norm": 0.036486461758613586,
|
| 633 |
+
"kl": 0.1916690319776535,
|
| 634 |
+
"learning_rate": 3.625e-06,
|
| 635 |
+
"loss": 0.0004791725368704647,
|
| 636 |
+
"step": 34,
|
| 637 |
+
"step_time": 0.22225405700010015
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"clip_ratio/high_max": 0.0,
|
| 641 |
+
"clip_ratio/high_mean": 0.0,
|
| 642 |
+
"clip_ratio/low_mean": 0.0,
|
| 643 |
+
"clip_ratio/low_min": 0.0,
|
| 644 |
+
"clip_ratio/region_mean": 0.0,
|
| 645 |
+
"entropy": 4.338494300842285,
|
| 646 |
+
"epoch": 1.4583333333333333,
|
| 647 |
+
"grad_norm": 0.04015596583485603,
|
| 648 |
+
"kl": 0.17945243418216705,
|
| 649 |
+
"learning_rate": 3.5833333333333335e-06,
|
| 650 |
+
"loss": 0.00044863109360449016,
|
| 651 |
+
"step": 35,
|
| 652 |
+
"step_time": 0.22212749400023313
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"clip_ratio/high_max": 0.0,
|
| 656 |
+
"clip_ratio/high_mean": 0.0,
|
| 657 |
+
"clip_ratio/low_mean": 0.0,
|
| 658 |
+
"clip_ratio/low_min": 0.0,
|
| 659 |
+
"clip_ratio/region_mean": 0.0,
|
| 660 |
+
"entropy": 4.584677219390869,
|
| 661 |
+
"epoch": 1.5,
|
| 662 |
+
"grad_norm": 0.05036609247326851,
|
| 663 |
+
"kl": 0.21357925236225128,
|
| 664 |
+
"learning_rate": 3.5416666666666673e-06,
|
| 665 |
+
"loss": 0.0005339480703696609,
|
| 666 |
+
"step": 36,
|
| 667 |
+
"step_time": 0.22190836000027048
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"clip_ratio/high_max": 0.0,
|
| 671 |
+
"clip_ratio/high_mean": 0.0,
|
| 672 |
+
"clip_ratio/low_mean": 0.0,
|
| 673 |
+
"clip_ratio/low_min": 0.0,
|
| 674 |
+
"clip_ratio/region_mean": 0.0,
|
| 675 |
+
"completions/clipped_ratio": 1.0,
|
| 676 |
+
"completions/max_length": 384.0,
|
| 677 |
+
"completions/max_terminated_length": 0.0,
|
| 678 |
+
"completions/mean_length": 384.0,
|
| 679 |
+
"completions/mean_terminated_length": 0.0,
|
| 680 |
+
"completions/min_length": 384.0,
|
| 681 |
+
"completions/min_terminated_length": 0.0,
|
| 682 |
+
"entropy": 4.206821918487549,
|
| 683 |
+
"epoch": 1.5416666666666665,
|
| 684 |
+
"frac_reward_zero_std": 0.75,
|
| 685 |
+
"grad_norm": 2.2474279403686523,
|
| 686 |
+
"kl": 0.17469431459903717,
|
| 687 |
+
"learning_rate": 3.5e-06,
|
| 688 |
+
"loss": 0.09434103965759277,
|
| 689 |
+
"num_tokens": 251600.0,
|
| 690 |
+
"reward": -0.08749999850988388,
|
| 691 |
+
"reward_std": 0.14999999105930328,
|
| 692 |
+
"rewards/GeneratorRewardFunction/mean": -0.08749999850988388,
|
| 693 |
+
"rewards/GeneratorRewardFunction/std": 0.15000000596046448,
|
| 694 |
+
"step": 37,
|
| 695 |
+
"step_time": 12.279682073000004
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"clip_ratio/high_max": 0.0,
|
| 699 |
+
"clip_ratio/high_mean": 0.0,
|
| 700 |
+
"clip_ratio/low_mean": 0.0,
|
| 701 |
+
"clip_ratio/low_min": 0.0,
|
| 702 |
+
"clip_ratio/region_mean": 0.0,
|
| 703 |
+
"entropy": 4.365553379058838,
|
| 704 |
+
"epoch": 1.5833333333333335,
|
| 705 |
+
"grad_norm": 0.7859151363372803,
|
| 706 |
+
"kl": 0.20030474662780762,
|
| 707 |
+
"learning_rate": 3.4583333333333334e-06,
|
| 708 |
+
"loss": -0.03081577830016613,
|
| 709 |
+
"step": 38,
|
| 710 |
+
"step_time": 0.22089524599959987
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"clip_ratio/high_max": 0.0013020833721384406,
|
| 714 |
+
"clip_ratio/high_mean": 0.0013020833721384406,
|
| 715 |
+
"clip_ratio/low_mean": 0.0,
|
| 716 |
+
"clip_ratio/low_min": 0.0,
|
| 717 |
+
"clip_ratio/region_mean": 0.0013020833721384406,
|
| 718 |
+
"entropy": 4.566871166229248,
|
| 719 |
+
"epoch": 1.625,
|
| 720 |
+
"grad_norm": 0.7892794013023376,
|
| 721 |
+
"kl": 0.23452065885066986,
|
| 722 |
+
"learning_rate": 3.416666666666667e-06,
|
| 723 |
+
"loss": -0.03072456084191799,
|
| 724 |
+
"step": 39,
|
| 725 |
+
"step_time": 0.22001067800010787
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"clip_ratio/high_max": 0.0052083334885537624,
|
| 729 |
+
"clip_ratio/high_mean": 0.0052083334885537624,
|
| 730 |
+
"clip_ratio/low_mean": 0.0,
|
| 731 |
+
"clip_ratio/low_min": 0.0,
|
| 732 |
+
"clip_ratio/region_mean": 0.0052083334885537624,
|
| 733 |
+
"entropy": 4.517628192901611,
|
| 734 |
+
"epoch": 1.6666666666666665,
|
| 735 |
+
"grad_norm": 0.7610399127006531,
|
| 736 |
+
"kl": 0.22405748069286346,
|
| 737 |
+
"learning_rate": 3.3750000000000003e-06,
|
| 738 |
+
"loss": -0.030617548152804375,
|
| 739 |
+
"step": 40,
|
| 740 |
+
"step_time": 0.22107800999947358
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"clip_ratio/high_max": 0.0,
|
| 744 |
+
"clip_ratio/high_mean": 0.0,
|
| 745 |
+
"clip_ratio/low_mean": 0.0,
|
| 746 |
+
"clip_ratio/low_min": 0.0,
|
| 747 |
+
"clip_ratio/region_mean": 0.0,
|
| 748 |
+
"completions/clipped_ratio": 1.0,
|
| 749 |
+
"completions/max_length": 384.0,
|
| 750 |
+
"completions/max_terminated_length": 0.0,
|
| 751 |
+
"completions/mean_length": 384.0,
|
| 752 |
+
"completions/mean_terminated_length": 0.0,
|
| 753 |
+
"completions/min_length": 384.0,
|
| 754 |
+
"completions/min_terminated_length": 0.0,
|
| 755 |
+
"entropy": 4.207302093505859,
|
| 756 |
+
"epoch": 1.7083333333333335,
|
| 757 |
+
"frac_reward_zero_std": 1.0,
|
| 758 |
+
"grad_norm": 0.03834126889705658,
|
| 759 |
+
"kl": 0.20183412730693817,
|
| 760 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 761 |
+
"loss": 0.000504585332237184,
|
| 762 |
+
"num_tokens": 276896.0,
|
| 763 |
+
"reward": -0.05000000074505806,
|
| 764 |
+
"reward_std": 0.0,
|
| 765 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 766 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 767 |
+
"step": 41,
|
| 768 |
+
"step_time": 12.462519400999554
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"clip_ratio/high_max": 0.0,
|
| 772 |
+
"clip_ratio/high_mean": 0.0,
|
| 773 |
+
"clip_ratio/low_mean": 0.0,
|
| 774 |
+
"clip_ratio/low_min": 0.0,
|
| 775 |
+
"clip_ratio/region_mean": 0.0,
|
| 776 |
+
"entropy": 4.1995134353637695,
|
| 777 |
+
"epoch": 1.75,
|
| 778 |
+
"grad_norm": 0.03475196659564972,
|
| 779 |
+
"kl": 0.20933139324188232,
|
| 780 |
+
"learning_rate": 3.2916666666666668e-06,
|
| 781 |
+
"loss": 0.0005233284318819642,
|
| 782 |
+
"step": 42,
|
| 783 |
+
"step_time": 0.2306131320001441
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"clip_ratio/high_max": 0.0,
|
| 787 |
+
"clip_ratio/high_mean": 0.0,
|
| 788 |
+
"clip_ratio/low_mean": 0.0,
|
| 789 |
+
"clip_ratio/low_min": 0.0,
|
| 790 |
+
"clip_ratio/region_mean": 0.0,
|
| 791 |
+
"entropy": 4.859399318695068,
|
| 792 |
+
"epoch": 1.7916666666666665,
|
| 793 |
+
"grad_norm": 0.043436359614133835,
|
| 794 |
+
"kl": 0.21264620125293732,
|
| 795 |
+
"learning_rate": 3.2500000000000002e-06,
|
| 796 |
+
"loss": 0.0005316155147738755,
|
| 797 |
+
"step": 43,
|
| 798 |
+
"step_time": 0.22955826899942622
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"clip_ratio/high_max": 0.0,
|
| 802 |
+
"clip_ratio/high_mean": 0.0,
|
| 803 |
+
"clip_ratio/low_mean": 0.0,
|
| 804 |
+
"clip_ratio/low_min": 0.0,
|
| 805 |
+
"clip_ratio/region_mean": 0.0,
|
| 806 |
+
"entropy": 4.589152812957764,
|
| 807 |
+
"epoch": 1.8333333333333335,
|
| 808 |
+
"grad_norm": 0.043015655130147934,
|
| 809 |
+
"kl": 0.22540146112442017,
|
| 810 |
+
"learning_rate": 3.2083333333333337e-06,
|
| 811 |
+
"loss": 0.000563503650482744,
|
| 812 |
+
"step": 44,
|
| 813 |
+
"step_time": 0.22972263199972076
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"clip_ratio/high_max": 0.0,
|
| 817 |
+
"clip_ratio/high_mean": 0.0,
|
| 818 |
+
"clip_ratio/low_mean": 0.0,
|
| 819 |
+
"clip_ratio/low_min": 0.0,
|
| 820 |
+
"clip_ratio/region_mean": 0.0,
|
| 821 |
+
"completions/clipped_ratio": 1.0,
|
| 822 |
+
"completions/max_length": 384.0,
|
| 823 |
+
"completions/max_terminated_length": 0.0,
|
| 824 |
+
"completions/mean_length": 384.0,
|
| 825 |
+
"completions/mean_terminated_length": 0.0,
|
| 826 |
+
"completions/min_length": 384.0,
|
| 827 |
+
"completions/min_terminated_length": 0.0,
|
| 828 |
+
"entropy": 4.461331844329834,
|
| 829 |
+
"epoch": 1.875,
|
| 830 |
+
"frac_reward_zero_std": 0.5,
|
| 831 |
+
"grad_norm": 0.03550349920988083,
|
| 832 |
+
"kl": 0.1928916722536087,
|
| 833 |
+
"learning_rate": 3.1666666666666667e-06,
|
| 834 |
+
"loss": 0.0004822291957680136,
|
| 835 |
+
"num_tokens": 301944.0,
|
| 836 |
+
"reward": -0.125,
|
| 837 |
+
"reward_std": 0.20493900775909424,
|
| 838 |
+
"rewards/GeneratorRewardFunction/mean": -0.125,
|
| 839 |
+
"rewards/GeneratorRewardFunction/std": 0.20493900775909424,
|
| 840 |
+
"step": 45,
|
| 841 |
+
"step_time": 12.369378186999711
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"clip_ratio/high_max": 0.0006510416860692203,
|
| 845 |
+
"clip_ratio/high_mean": 0.0006510416860692203,
|
| 846 |
+
"clip_ratio/low_mean": 0.0,
|
| 847 |
+
"clip_ratio/low_min": 0.0,
|
| 848 |
+
"clip_ratio/region_mean": 0.0006510416860692203,
|
| 849 |
+
"entropy": 4.596534252166748,
|
| 850 |
+
"epoch": 1.9166666666666665,
|
| 851 |
+
"grad_norm": 1.093070149421692,
|
| 852 |
+
"kl": 0.1923203319311142,
|
| 853 |
+
"learning_rate": 3.125e-06,
|
| 854 |
+
"loss": -0.062203943729400635,
|
| 855 |
+
"step": 46,
|
| 856 |
+
"step_time": 0.22156830799940508
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"clip_ratio/high_max": 0.0026041667442768812,
|
| 860 |
+
"clip_ratio/high_mean": 0.0026041667442768812,
|
| 861 |
+
"clip_ratio/low_mean": 0.0,
|
| 862 |
+
"clip_ratio/low_min": 0.0,
|
| 863 |
+
"clip_ratio/region_mean": 0.0026041667442768812,
|
| 864 |
+
"entropy": 4.4759135246276855,
|
| 865 |
+
"epoch": 1.9583333333333335,
|
| 866 |
+
"grad_norm": 2.8484206199645996,
|
| 867 |
+
"kl": 0.22040079534053802,
|
| 868 |
+
"learning_rate": 3.0833333333333336e-06,
|
| 869 |
+
"loss": 0.0003803384315688163,
|
| 870 |
+
"step": 47,
|
| 871 |
+
"step_time": 0.22012663999976212
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"clip_ratio/high_max": 0.001953125,
|
| 875 |
+
"clip_ratio/high_mean": 0.001953125,
|
| 876 |
+
"clip_ratio/low_mean": 0.0013020833721384406,
|
| 877 |
+
"clip_ratio/low_min": 0.0013020833721384406,
|
| 878 |
+
"clip_ratio/region_mean": 0.0032552082557231188,
|
| 879 |
+
"entropy": 4.502986907958984,
|
| 880 |
+
"epoch": 2.0,
|
| 881 |
+
"grad_norm": 2.476116418838501,
|
| 882 |
+
"kl": 0.17874129116535187,
|
| 883 |
+
"learning_rate": 3.0416666666666666e-06,
|
| 884 |
+
"loss": 0.06300241500139236,
|
| 885 |
+
"step": 48,
|
| 886 |
+
"step_time": 0.22058342899981653
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"clip_ratio/high_max": 0.0,
|
| 890 |
+
"clip_ratio/high_mean": 0.0,
|
| 891 |
+
"clip_ratio/low_mean": 0.0,
|
| 892 |
+
"clip_ratio/low_min": 0.0,
|
| 893 |
+
"clip_ratio/region_mean": 0.0,
|
| 894 |
+
"completions/clipped_ratio": 1.0,
|
| 895 |
+
"completions/max_length": 384.0,
|
| 896 |
+
"completions/max_terminated_length": 0.0,
|
| 897 |
+
"completions/mean_length": 384.0,
|
| 898 |
+
"completions/mean_terminated_length": 0.0,
|
| 899 |
+
"completions/min_length": 384.0,
|
| 900 |
+
"completions/min_terminated_length": 0.0,
|
| 901 |
+
"entropy": 4.52464485168457,
|
| 902 |
+
"epoch": 2.0416666666666665,
|
| 903 |
+
"frac_reward_zero_std": 0.75,
|
| 904 |
+
"grad_norm": 1.1072802543640137,
|
| 905 |
+
"kl": 0.17327140271663666,
|
| 906 |
+
"learning_rate": 3e-06,
|
| 907 |
+
"loss": -0.0618751160800457,
|
| 908 |
+
"num_tokens": 327088.0,
|
| 909 |
+
"reward": -0.05406250059604645,
|
| 910 |
+
"reward_std": 0.016249997541308403,
|
| 911 |
+
"rewards/GeneratorRewardFunction/mean": -0.05406250059604645,
|
| 912 |
+
"rewards/GeneratorRewardFunction/std": 0.0162500012665987,
|
| 913 |
+
"step": 49,
|
| 914 |
+
"step_time": 12.398161551999692
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"clip_ratio/high_max": 0.0,
|
| 918 |
+
"clip_ratio/high_mean": 0.0,
|
| 919 |
+
"clip_ratio/low_mean": 0.0,
|
| 920 |
+
"clip_ratio/low_min": 0.0,
|
| 921 |
+
"clip_ratio/region_mean": 0.0,
|
| 922 |
+
"entropy": 4.568341255187988,
|
| 923 |
+
"epoch": 2.0833333333333335,
|
| 924 |
+
"grad_norm": 0.03384983912110329,
|
| 925 |
+
"kl": 0.19154612720012665,
|
| 926 |
+
"learning_rate": 2.9583333333333335e-06,
|
| 927 |
+
"loss": 0.0004788653168361634,
|
| 928 |
+
"step": 50,
|
| 929 |
+
"step_time": 0.2228688120003426
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"clip_ratio/high_max": 0.0026041667442768812,
|
| 933 |
+
"clip_ratio/high_mean": 0.0026041667442768812,
|
| 934 |
+
"clip_ratio/low_mean": 0.0,
|
| 935 |
+
"clip_ratio/low_min": 0.0,
|
| 936 |
+
"clip_ratio/region_mean": 0.0026041667442768812,
|
| 937 |
+
"entropy": 4.437350749969482,
|
| 938 |
+
"epoch": 2.125,
|
| 939 |
+
"grad_norm": 0.7250102758407593,
|
| 940 |
+
"kl": 0.1893668919801712,
|
| 941 |
+
"learning_rate": 2.916666666666667e-06,
|
| 942 |
+
"loss": -0.030638111755251884,
|
| 943 |
+
"step": 51,
|
| 944 |
+
"step_time": 0.2229313830002866
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"clip_ratio/high_max": 0.0,
|
| 948 |
+
"clip_ratio/high_mean": 0.0,
|
| 949 |
+
"clip_ratio/low_mean": 0.0,
|
| 950 |
+
"clip_ratio/low_min": 0.0,
|
| 951 |
+
"clip_ratio/region_mean": 0.0,
|
| 952 |
+
"entropy": 4.27592134475708,
|
| 953 |
+
"epoch": 2.1666666666666665,
|
| 954 |
+
"grad_norm": 2.2188467979431152,
|
| 955 |
+
"kl": 0.17189562320709229,
|
| 956 |
+
"learning_rate": 2.875e-06,
|
| 957 |
+
"loss": 0.09424998611211777,
|
| 958 |
+
"step": 52,
|
| 959 |
+
"step_time": 0.2230674660004297
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"clip_ratio/high_max": 0.0,
|
| 963 |
+
"clip_ratio/high_mean": 0.0,
|
| 964 |
+
"clip_ratio/low_mean": 0.0,
|
| 965 |
+
"clip_ratio/low_min": 0.0,
|
| 966 |
+
"clip_ratio/region_mean": 0.0,
|
| 967 |
+
"completions/clipped_ratio": 1.0,
|
| 968 |
+
"completions/max_length": 384.0,
|
| 969 |
+
"completions/max_terminated_length": 0.0,
|
| 970 |
+
"completions/mean_length": 384.0,
|
| 971 |
+
"completions/mean_terminated_length": 0.0,
|
| 972 |
+
"completions/min_length": 384.0,
|
| 973 |
+
"completions/min_terminated_length": 0.0,
|
| 974 |
+
"entropy": 4.4265666007995605,
|
| 975 |
+
"epoch": 2.2083333333333335,
|
| 976 |
+
"frac_reward_zero_std": 1.0,
|
| 977 |
+
"grad_norm": 0.038845159113407135,
|
| 978 |
+
"kl": 0.21073134243488312,
|
| 979 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 980 |
+
"loss": 0.0005268284003250301,
|
| 981 |
+
"num_tokens": 352328.0,
|
| 982 |
+
"reward": -0.05000000074505806,
|
| 983 |
+
"reward_std": 0.0,
|
| 984 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 985 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 986 |
+
"step": 53,
|
| 987 |
+
"step_time": 12.3143936510005
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"clip_ratio/high_max": 0.0,
|
| 991 |
+
"clip_ratio/high_mean": 0.0,
|
| 992 |
+
"clip_ratio/low_mean": 0.0,
|
| 993 |
+
"clip_ratio/low_min": 0.0,
|
| 994 |
+
"clip_ratio/region_mean": 0.0,
|
| 995 |
+
"entropy": 4.656918525695801,
|
| 996 |
+
"epoch": 2.25,
|
| 997 |
+
"grad_norm": 0.0484224371612072,
|
| 998 |
+
"kl": 0.2176627665758133,
|
| 999 |
+
"learning_rate": 2.791666666666667e-06,
|
| 1000 |
+
"loss": 0.0005441569373942912,
|
| 1001 |
+
"step": 54,
|
| 1002 |
+
"step_time": 0.22369184699982725
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"clip_ratio/high_max": 0.0,
|
| 1006 |
+
"clip_ratio/high_mean": 0.0,
|
| 1007 |
+
"clip_ratio/low_mean": 0.0,
|
| 1008 |
+
"clip_ratio/low_min": 0.0,
|
| 1009 |
+
"clip_ratio/region_mean": 0.0,
|
| 1010 |
+
"entropy": 4.2594475746154785,
|
| 1011 |
+
"epoch": 2.2916666666666665,
|
| 1012 |
+
"grad_norm": 0.03165994957089424,
|
| 1013 |
+
"kl": 0.19008274376392365,
|
| 1014 |
+
"learning_rate": 2.7500000000000004e-06,
|
| 1015 |
+
"loss": 0.0004752068780362606,
|
| 1016 |
+
"step": 55,
|
| 1017 |
+
"step_time": 0.22343963099956454
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"clip_ratio/high_max": 0.0,
|
| 1021 |
+
"clip_ratio/high_mean": 0.0,
|
| 1022 |
+
"clip_ratio/low_mean": 0.0,
|
| 1023 |
+
"clip_ratio/low_min": 0.0,
|
| 1024 |
+
"clip_ratio/region_mean": 0.0,
|
| 1025 |
+
"entropy": 4.303634166717529,
|
| 1026 |
+
"epoch": 2.3333333333333335,
|
| 1027 |
+
"grad_norm": 0.03170448914170265,
|
| 1028 |
+
"kl": 0.18705779314041138,
|
| 1029 |
+
"learning_rate": 2.7083333333333334e-06,
|
| 1030 |
+
"loss": 0.00046764445141889155,
|
| 1031 |
+
"step": 56,
|
| 1032 |
+
"step_time": 0.22311556299973745
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"clip_ratio/high_max": 0.0,
|
| 1036 |
+
"clip_ratio/high_mean": 0.0,
|
| 1037 |
+
"clip_ratio/low_mean": 0.0,
|
| 1038 |
+
"clip_ratio/low_min": 0.0,
|
| 1039 |
+
"clip_ratio/region_mean": 0.0,
|
| 1040 |
+
"completions/clipped_ratio": 1.0,
|
| 1041 |
+
"completions/max_length": 384.0,
|
| 1042 |
+
"completions/max_terminated_length": 0.0,
|
| 1043 |
+
"completions/mean_length": 384.0,
|
| 1044 |
+
"completions/mean_terminated_length": 0.0,
|
| 1045 |
+
"completions/min_length": 384.0,
|
| 1046 |
+
"completions/min_terminated_length": 0.0,
|
| 1047 |
+
"entropy": 4.496265888214111,
|
| 1048 |
+
"epoch": 2.375,
|
| 1049 |
+
"frac_reward_zero_std": 0.5,
|
| 1050 |
+
"grad_norm": 2.193422794342041,
|
| 1051 |
+
"kl": 0.22647613286972046,
|
| 1052 |
+
"learning_rate": 2.666666666666667e-06,
|
| 1053 |
+
"loss": 0.09425372630357742,
|
| 1054 |
+
"num_tokens": 377544.0,
|
| 1055 |
+
"reward": -0.0690625011920929,
|
| 1056 |
+
"reward_std": 0.07492704689502716,
|
| 1057 |
+
"rewards/GeneratorRewardFunction/mean": -0.0690625011920929,
|
| 1058 |
+
"rewards/GeneratorRewardFunction/std": 0.07492704689502716,
|
| 1059 |
+
"step": 57,
|
| 1060 |
+
"step_time": 12.497021862999645
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"clip_ratio/high_max": 0.0,
|
| 1064 |
+
"clip_ratio/high_mean": 0.0,
|
| 1065 |
+
"clip_ratio/low_mean": 0.0,
|
| 1066 |
+
"clip_ratio/low_min": 0.0,
|
| 1067 |
+
"clip_ratio/region_mean": 0.0,
|
| 1068 |
+
"entropy": 4.573244094848633,
|
| 1069 |
+
"epoch": 2.4166666666666665,
|
| 1070 |
+
"grad_norm": 1.3519999980926514,
|
| 1071 |
+
"kl": 0.21581268310546875,
|
| 1072 |
+
"learning_rate": 2.6250000000000003e-06,
|
| 1073 |
+
"loss": -0.09204903244972229,
|
| 1074 |
+
"step": 58,
|
| 1075 |
+
"step_time": 0.23012131300038163
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"clip_ratio/high_max": 0.001953125,
|
| 1079 |
+
"clip_ratio/high_mean": 0.001953125,
|
| 1080 |
+
"clip_ratio/low_mean": 0.0,
|
| 1081 |
+
"clip_ratio/low_min": 0.0,
|
| 1082 |
+
"clip_ratio/region_mean": 0.001953125,
|
| 1083 |
+
"entropy": 4.067762851715088,
|
| 1084 |
+
"epoch": 2.4583333333333335,
|
| 1085 |
+
"grad_norm": 2.2663047313690186,
|
| 1086 |
+
"kl": 0.19602923095226288,
|
| 1087 |
+
"learning_rate": 2.5833333333333337e-06,
|
| 1088 |
+
"loss": 0.05952334776520729,
|
| 1089 |
+
"step": 59,
|
| 1090 |
+
"step_time": 0.23155366399987543
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"clip_ratio/high_max": 0.008463541977107525,
|
| 1094 |
+
"clip_ratio/high_mean": 0.008463541977107525,
|
| 1095 |
+
"clip_ratio/low_mean": 0.0,
|
| 1096 |
+
"clip_ratio/low_min": 0.0,
|
| 1097 |
+
"clip_ratio/region_mean": 0.008463541977107525,
|
| 1098 |
+
"entropy": 4.1496710777282715,
|
| 1099 |
+
"epoch": 2.5,
|
| 1100 |
+
"grad_norm": 1.0887759923934937,
|
| 1101 |
+
"kl": 0.2114473581314087,
|
| 1102 |
+
"learning_rate": 2.5416666666666668e-06,
|
| 1103 |
+
"loss": -0.05951349437236786,
|
| 1104 |
+
"step": 60,
|
| 1105 |
+
"step_time": 0.22960056100055226
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"clip_ratio/high_max": 0.0,
|
| 1109 |
+
"clip_ratio/high_mean": 0.0,
|
| 1110 |
+
"clip_ratio/low_mean": 0.0,
|
| 1111 |
+
"clip_ratio/low_min": 0.0,
|
| 1112 |
+
"clip_ratio/region_mean": 0.0,
|
| 1113 |
+
"completions/clipped_ratio": 1.0,
|
| 1114 |
+
"completions/max_length": 384.0,
|
| 1115 |
+
"completions/max_terminated_length": 0.0,
|
| 1116 |
+
"completions/mean_length": 384.0,
|
| 1117 |
+
"completions/mean_terminated_length": 0.0,
|
| 1118 |
+
"completions/min_length": 384.0,
|
| 1119 |
+
"completions/min_terminated_length": 0.0,
|
| 1120 |
+
"entropy": 4.408988952636719,
|
| 1121 |
+
"epoch": 2.5416666666666665,
|
| 1122 |
+
"frac_reward_zero_std": 0.5,
|
| 1123 |
+
"grad_norm": 1.3472992181777954,
|
| 1124 |
+
"kl": 0.18820710480213165,
|
| 1125 |
+
"learning_rate": 2.5e-06,
|
| 1126 |
+
"loss": -0.09331253170967102,
|
| 1127 |
+
"num_tokens": 402808.0,
|
| 1128 |
+
"reward": -0.1315625011920929,
|
| 1129 |
+
"reward_std": 0.2236941158771515,
|
| 1130 |
+
"rewards/GeneratorRewardFunction/mean": -0.1315625011920929,
|
| 1131 |
+
"rewards/GeneratorRewardFunction/std": 0.22369414567947388,
|
| 1132 |
+
"step": 61,
|
| 1133 |
+
"step_time": 12.39430396500029
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"clip_ratio/high_max": 0.0,
|
| 1137 |
+
"clip_ratio/high_mean": 0.0,
|
| 1138 |
+
"clip_ratio/low_mean": 0.0,
|
| 1139 |
+
"clip_ratio/low_min": 0.0,
|
| 1140 |
+
"clip_ratio/region_mean": 0.0,
|
| 1141 |
+
"entropy": 3.8388965129852295,
|
| 1142 |
+
"epoch": 2.5833333333333335,
|
| 1143 |
+
"grad_norm": 3.653455972671509,
|
| 1144 |
+
"kl": 0.21382872760295868,
|
| 1145 |
+
"learning_rate": 2.4583333333333332e-06,
|
| 1146 |
+
"loss": 0.18890579044818878,
|
| 1147 |
+
"step": 62,
|
| 1148 |
+
"step_time": 0.22953293099999428
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"clip_ratio/high_max": 0.0006510416860692203,
|
| 1152 |
+
"clip_ratio/high_mean": 0.0006510416860692203,
|
| 1153 |
+
"clip_ratio/low_mean": 0.0,
|
| 1154 |
+
"clip_ratio/low_min": 0.0,
|
| 1155 |
+
"clip_ratio/region_mean": 0.0006510416860692203,
|
| 1156 |
+
"entropy": 3.9201724529266357,
|
| 1157 |
+
"epoch": 2.625,
|
| 1158 |
+
"grad_norm": 1.142210602760315,
|
| 1159 |
+
"kl": 0.2663920223712921,
|
| 1160 |
+
"learning_rate": 2.4166666666666667e-06,
|
| 1161 |
+
"loss": -0.061935603618621826,
|
| 1162 |
+
"step": 63,
|
| 1163 |
+
"step_time": 0.22865817100046115
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"clip_ratio/high_max": 0.0045572915114462376,
|
| 1167 |
+
"clip_ratio/high_mean": 0.0045572915114462376,
|
| 1168 |
+
"clip_ratio/low_mean": 0.0,
|
| 1169 |
+
"clip_ratio/low_min": 0.0,
|
| 1170 |
+
"clip_ratio/region_mean": 0.0045572915114462376,
|
| 1171 |
+
"entropy": 4.356930255889893,
|
| 1172 |
+
"epoch": 2.6666666666666665,
|
| 1173 |
+
"grad_norm": 0.7153739333152771,
|
| 1174 |
+
"kl": 0.20749028027057648,
|
| 1175 |
+
"learning_rate": 2.375e-06,
|
| 1176 |
+
"loss": -0.03097183257341385,
|
| 1177 |
+
"step": 64,
|
| 1178 |
+
"step_time": 0.2284142079997764
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"clip_ratio/high_max": 0.0,
|
| 1182 |
+
"clip_ratio/high_mean": 0.0,
|
| 1183 |
+
"clip_ratio/low_mean": 0.0,
|
| 1184 |
+
"clip_ratio/low_min": 0.0,
|
| 1185 |
+
"clip_ratio/region_mean": 0.0,
|
| 1186 |
+
"completions/clipped_ratio": 1.0,
|
| 1187 |
+
"completions/max_length": 384.0,
|
| 1188 |
+
"completions/max_terminated_length": 0.0,
|
| 1189 |
+
"completions/mean_length": 384.0,
|
| 1190 |
+
"completions/mean_terminated_length": 0.0,
|
| 1191 |
+
"completions/min_length": 384.0,
|
| 1192 |
+
"completions/min_terminated_length": 0.0,
|
| 1193 |
+
"entropy": 4.17288064956665,
|
| 1194 |
+
"epoch": 2.7083333333333335,
|
| 1195 |
+
"frac_reward_zero_std": 1.0,
|
| 1196 |
+
"grad_norm": 0.042102206498384476,
|
| 1197 |
+
"kl": 0.21204183995723724,
|
| 1198 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 1199 |
+
"loss": 0.0005301045603118837,
|
| 1200 |
+
"num_tokens": 427876.0,
|
| 1201 |
+
"reward": -0.05000000074505806,
|
| 1202 |
+
"reward_std": 0.0,
|
| 1203 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 1204 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 1205 |
+
"step": 65,
|
| 1206 |
+
"step_time": 12.319688841999778
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"clip_ratio/high_max": 0.0,
|
| 1210 |
+
"clip_ratio/high_mean": 0.0,
|
| 1211 |
+
"clip_ratio/low_mean": 0.0,
|
| 1212 |
+
"clip_ratio/low_min": 0.0,
|
| 1213 |
+
"clip_ratio/region_mean": 0.0,
|
| 1214 |
+
"entropy": 4.3474955558776855,
|
| 1215 |
+
"epoch": 2.75,
|
| 1216 |
+
"grad_norm": 0.03690265491604805,
|
| 1217 |
+
"kl": 0.21480290591716766,
|
| 1218 |
+
"learning_rate": 2.2916666666666666e-06,
|
| 1219 |
+
"loss": 0.000537007232196629,
|
| 1220 |
+
"step": 66,
|
| 1221 |
+
"step_time": 0.2216859949994614
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"clip_ratio/high_max": 0.0,
|
| 1225 |
+
"clip_ratio/high_mean": 0.0,
|
| 1226 |
+
"clip_ratio/low_mean": 0.0,
|
| 1227 |
+
"clip_ratio/low_min": 0.0,
|
| 1228 |
+
"clip_ratio/region_mean": 0.0,
|
| 1229 |
+
"entropy": 4.275708198547363,
|
| 1230 |
+
"epoch": 2.7916666666666665,
|
| 1231 |
+
"grad_norm": 0.035951025784015656,
|
| 1232 |
+
"kl": 0.19490306079387665,
|
| 1233 |
+
"learning_rate": 2.25e-06,
|
| 1234 |
+
"loss": 0.0004872575809713453,
|
| 1235 |
+
"step": 67,
|
| 1236 |
+
"step_time": 0.22137762800048222
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"clip_ratio/high_max": 0.0,
|
| 1240 |
+
"clip_ratio/high_mean": 0.0,
|
| 1241 |
+
"clip_ratio/low_mean": 0.0,
|
| 1242 |
+
"clip_ratio/low_min": 0.0,
|
| 1243 |
+
"clip_ratio/region_mean": 0.0,
|
| 1244 |
+
"entropy": 4.227705001831055,
|
| 1245 |
+
"epoch": 2.8333333333333335,
|
| 1246 |
+
"grad_norm": 0.035950690507888794,
|
| 1247 |
+
"kl": 0.22681327164173126,
|
| 1248 |
+
"learning_rate": 2.2083333333333335e-06,
|
| 1249 |
+
"loss": 0.0005670331884175539,
|
| 1250 |
+
"step": 68,
|
| 1251 |
+
"step_time": 0.22117000499929418
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"clip_ratio/high_max": 0.0,
|
| 1255 |
+
"clip_ratio/high_mean": 0.0,
|
| 1256 |
+
"clip_ratio/low_mean": 0.0,
|
| 1257 |
+
"clip_ratio/low_min": 0.0,
|
| 1258 |
+
"clip_ratio/region_mean": 0.0,
|
| 1259 |
+
"completions/clipped_ratio": 1.0,
|
| 1260 |
+
"completions/max_length": 384.0,
|
| 1261 |
+
"completions/max_terminated_length": 0.0,
|
| 1262 |
+
"completions/mean_length": 384.0,
|
| 1263 |
+
"completions/mean_terminated_length": 0.0,
|
| 1264 |
+
"completions/min_length": 384.0,
|
| 1265 |
+
"completions/min_terminated_length": 0.0,
|
| 1266 |
+
"entropy": 4.621903896331787,
|
| 1267 |
+
"epoch": 2.875,
|
| 1268 |
+
"frac_reward_zero_std": 0.75,
|
| 1269 |
+
"grad_norm": 2.626718521118164,
|
| 1270 |
+
"kl": 0.21730534732341766,
|
| 1271 |
+
"learning_rate": 2.166666666666667e-06,
|
| 1272 |
+
"loss": 0.0005432550678960979,
|
| 1273 |
+
"num_tokens": 452916.0,
|
| 1274 |
+
"reward": -0.08749999850988388,
|
| 1275 |
+
"reward_std": 0.14999999105930328,
|
| 1276 |
+
"rewards/GeneratorRewardFunction/mean": -0.08749999850988388,
|
| 1277 |
+
"rewards/GeneratorRewardFunction/std": 0.15000000596046448,
|
| 1278 |
+
"step": 69,
|
| 1279 |
+
"step_time": 12.363676371999645
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"clip_ratio/high_max": 0.0,
|
| 1283 |
+
"clip_ratio/high_mean": 0.0,
|
| 1284 |
+
"clip_ratio/low_mean": 0.0,
|
| 1285 |
+
"clip_ratio/low_min": 0.0,
|
| 1286 |
+
"clip_ratio/region_mean": 0.0,
|
| 1287 |
+
"entropy": 4.34497594833374,
|
| 1288 |
+
"epoch": 2.9166666666666665,
|
| 1289 |
+
"grad_norm": 0.04092087969183922,
|
| 1290 |
+
"kl": 0.23031370341777802,
|
| 1291 |
+
"learning_rate": 2.125e-06,
|
| 1292 |
+
"loss": 0.0005757841863669455,
|
| 1293 |
+
"step": 70,
|
| 1294 |
+
"step_time": 0.22308474500005104
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"clip_ratio/high_max": 0.0,
|
| 1298 |
+
"clip_ratio/high_mean": 0.0,
|
| 1299 |
+
"clip_ratio/low_mean": 0.0,
|
| 1300 |
+
"clip_ratio/low_min": 0.0,
|
| 1301 |
+
"clip_ratio/region_mean": 0.0,
|
| 1302 |
+
"entropy": 4.401247024536133,
|
| 1303 |
+
"epoch": 2.9583333333333335,
|
| 1304 |
+
"grad_norm": 0.04431963711977005,
|
| 1305 |
+
"kl": 0.20616620779037476,
|
| 1306 |
+
"learning_rate": 2.0833333333333334e-06,
|
| 1307 |
+
"loss": 0.0005154155660420656,
|
| 1308 |
+
"step": 71,
|
| 1309 |
+
"step_time": 0.22127309600000444
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"clip_ratio/high_max": 0.0,
|
| 1313 |
+
"clip_ratio/high_mean": 0.0,
|
| 1314 |
+
"clip_ratio/low_mean": 0.0,
|
| 1315 |
+
"clip_ratio/low_min": 0.0,
|
| 1316 |
+
"clip_ratio/region_mean": 0.0,
|
| 1317 |
+
"entropy": 4.58444356918335,
|
| 1318 |
+
"epoch": 3.0,
|
| 1319 |
+
"grad_norm": 0.04426565393805504,
|
| 1320 |
+
"kl": 0.24117839336395264,
|
| 1321 |
+
"learning_rate": 2.041666666666667e-06,
|
| 1322 |
+
"loss": 0.0006029459764249623,
|
| 1323 |
+
"step": 72,
|
| 1324 |
+
"step_time": 0.221652592999817
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"clip_ratio/high_max": 0.0,
|
| 1328 |
+
"clip_ratio/high_mean": 0.0,
|
| 1329 |
+
"clip_ratio/low_mean": 0.0,
|
| 1330 |
+
"clip_ratio/low_min": 0.0,
|
| 1331 |
+
"clip_ratio/region_mean": 0.0,
|
| 1332 |
+
"completions/clipped_ratio": 1.0,
|
| 1333 |
+
"completions/max_length": 384.0,
|
| 1334 |
+
"completions/max_terminated_length": 0.0,
|
| 1335 |
+
"completions/mean_length": 384.0,
|
| 1336 |
+
"completions/mean_terminated_length": 0.0,
|
| 1337 |
+
"completions/min_length": 384.0,
|
| 1338 |
+
"completions/min_terminated_length": 0.0,
|
| 1339 |
+
"entropy": 4.410362720489502,
|
| 1340 |
+
"epoch": 3.0416666666666665,
|
| 1341 |
+
"frac_reward_zero_std": 1.0,
|
| 1342 |
+
"grad_norm": 0.03857680782675743,
|
| 1343 |
+
"kl": 0.21758143603801727,
|
| 1344 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 1345 |
+
"loss": 0.0005439535598270595,
|
| 1346 |
+
"num_tokens": 478212.0,
|
| 1347 |
+
"reward": -0.05000000074505806,
|
| 1348 |
+
"reward_std": 0.0,
|
| 1349 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 1350 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 1351 |
+
"step": 73,
|
| 1352 |
+
"step_time": 12.467288267000185
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"clip_ratio/high_max": 0.0,
|
| 1356 |
+
"clip_ratio/high_mean": 0.0,
|
| 1357 |
+
"clip_ratio/low_mean": 0.0,
|
| 1358 |
+
"clip_ratio/low_min": 0.0,
|
| 1359 |
+
"clip_ratio/region_mean": 0.0,
|
| 1360 |
+
"entropy": 4.536159038543701,
|
| 1361 |
+
"epoch": 3.0833333333333335,
|
| 1362 |
+
"grad_norm": 0.05269327014684677,
|
| 1363 |
+
"kl": 0.24625271558761597,
|
| 1364 |
+
"learning_rate": 1.9583333333333334e-06,
|
| 1365 |
+
"loss": 0.0006156317540444434,
|
| 1366 |
+
"step": 74,
|
| 1367 |
+
"step_time": 0.23044603900052607
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"clip_ratio/high_max": 0.0,
|
| 1371 |
+
"clip_ratio/high_mean": 0.0,
|
| 1372 |
+
"clip_ratio/low_mean": 0.0,
|
| 1373 |
+
"clip_ratio/low_min": 0.0,
|
| 1374 |
+
"clip_ratio/region_mean": 0.0,
|
| 1375 |
+
"entropy": 4.654154300689697,
|
| 1376 |
+
"epoch": 3.125,
|
| 1377 |
+
"grad_norm": 0.045070722699165344,
|
| 1378 |
+
"kl": 0.2163999080657959,
|
| 1379 |
+
"learning_rate": 1.916666666666667e-06,
|
| 1380 |
+
"loss": 0.0005409997538663447,
|
| 1381 |
+
"step": 75,
|
| 1382 |
+
"step_time": 0.22983702600049583
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"clip_ratio/high_max": 0.0,
|
| 1386 |
+
"clip_ratio/high_mean": 0.0,
|
| 1387 |
+
"clip_ratio/low_mean": 0.0,
|
| 1388 |
+
"clip_ratio/low_min": 0.0,
|
| 1389 |
+
"clip_ratio/region_mean": 0.0,
|
| 1390 |
+
"entropy": 4.584025859832764,
|
| 1391 |
+
"epoch": 3.1666666666666665,
|
| 1392 |
+
"grad_norm": 0.042388372123241425,
|
| 1393 |
+
"kl": 0.23418636620044708,
|
| 1394 |
+
"learning_rate": 1.8750000000000003e-06,
|
| 1395 |
+
"loss": 0.0005854659248143435,
|
| 1396 |
+
"step": 76,
|
| 1397 |
+
"step_time": 0.22974896400046418
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"clip_ratio/high_max": 0.0,
|
| 1401 |
+
"clip_ratio/high_mean": 0.0,
|
| 1402 |
+
"clip_ratio/low_mean": 0.0,
|
| 1403 |
+
"clip_ratio/low_min": 0.0,
|
| 1404 |
+
"clip_ratio/region_mean": 0.0,
|
| 1405 |
+
"completions/clipped_ratio": 1.0,
|
| 1406 |
+
"completions/max_length": 384.0,
|
| 1407 |
+
"completions/max_terminated_length": 0.0,
|
| 1408 |
+
"completions/mean_length": 384.0,
|
| 1409 |
+
"completions/mean_terminated_length": 0.0,
|
| 1410 |
+
"completions/min_length": 384.0,
|
| 1411 |
+
"completions/min_terminated_length": 0.0,
|
| 1412 |
+
"entropy": 3.430957794189453,
|
| 1413 |
+
"epoch": 3.2083333333333335,
|
| 1414 |
+
"frac_reward_zero_std": 0.75,
|
| 1415 |
+
"grad_norm": 0.24369363486766815,
|
| 1416 |
+
"kl": 0.1837681084871292,
|
| 1417 |
+
"learning_rate": 1.8333333333333333e-06,
|
| 1418 |
+
"loss": -0.029588719829916954,
|
| 1419 |
+
"num_tokens": 503340.0,
|
| 1420 |
+
"reward": -0.050312504172325134,
|
| 1421 |
+
"reward_std": 0.0012500007869675756,
|
| 1422 |
+
"rewards/GeneratorRewardFunction/mean": -0.050312504172325134,
|
| 1423 |
+
"rewards/GeneratorRewardFunction/std": 0.001249999739229679,
|
| 1424 |
+
"step": 77,
|
| 1425 |
+
"step_time": 12.413958975999776
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"clip_ratio/high_max": 0.0,
|
| 1429 |
+
"clip_ratio/high_mean": 0.0,
|
| 1430 |
+
"clip_ratio/low_mean": 0.0,
|
| 1431 |
+
"clip_ratio/low_min": 0.0,
|
| 1432 |
+
"clip_ratio/region_mean": 0.0,
|
| 1433 |
+
"entropy": 4.394783020019531,
|
| 1434 |
+
"epoch": 3.25,
|
| 1435 |
+
"grad_norm": 0.7777454853057861,
|
| 1436 |
+
"kl": 0.2489025592803955,
|
| 1437 |
+
"learning_rate": 1.7916666666666667e-06,
|
| 1438 |
+
"loss": -0.029388101771473885,
|
| 1439 |
+
"step": 78,
|
| 1440 |
+
"step_time": 0.22276251499988575
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"clip_ratio/high_max": 0.0,
|
| 1444 |
+
"clip_ratio/high_mean": 0.0,
|
| 1445 |
+
"clip_ratio/low_mean": 0.0,
|
| 1446 |
+
"clip_ratio/low_min": 0.0,
|
| 1447 |
+
"clip_ratio/region_mean": 0.0,
|
| 1448 |
+
"entropy": 4.531010150909424,
|
| 1449 |
+
"epoch": 3.2916666666666665,
|
| 1450 |
+
"grad_norm": 0.7285988926887512,
|
| 1451 |
+
"kl": 0.2649364471435547,
|
| 1452 |
+
"learning_rate": 1.75e-06,
|
| 1453 |
+
"loss": -0.029455339536070824,
|
| 1454 |
+
"step": 79,
|
| 1455 |
+
"step_time": 0.22254095900007087
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"clip_ratio/high_max": 0.0,
|
| 1459 |
+
"clip_ratio/high_mean": 0.0,
|
| 1460 |
+
"clip_ratio/low_mean": 0.0,
|
| 1461 |
+
"clip_ratio/low_min": 0.0,
|
| 1462 |
+
"clip_ratio/region_mean": 0.0,
|
| 1463 |
+
"entropy": 4.429391860961914,
|
| 1464 |
+
"epoch": 3.3333333333333335,
|
| 1465 |
+
"grad_norm": 2.166607618331909,
|
| 1466 |
+
"kl": 0.2291480153799057,
|
| 1467 |
+
"learning_rate": 1.7083333333333334e-06,
|
| 1468 |
+
"loss": 0.09083986282348633,
|
| 1469 |
+
"step": 80,
|
| 1470 |
+
"step_time": 0.2218179120000059
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"clip_ratio/high_max": 0.0,
|
| 1474 |
+
"clip_ratio/high_mean": 0.0,
|
| 1475 |
+
"clip_ratio/low_mean": 0.0,
|
| 1476 |
+
"clip_ratio/low_min": 0.0,
|
| 1477 |
+
"clip_ratio/region_mean": 0.0,
|
| 1478 |
+
"completions/clipped_ratio": 1.0,
|
| 1479 |
+
"completions/max_length": 384.0,
|
| 1480 |
+
"completions/max_terminated_length": 0.0,
|
| 1481 |
+
"completions/mean_length": 384.0,
|
| 1482 |
+
"completions/mean_terminated_length": 0.0,
|
| 1483 |
+
"completions/min_length": 384.0,
|
| 1484 |
+
"completions/min_terminated_length": 0.0,
|
| 1485 |
+
"entropy": 4.478769779205322,
|
| 1486 |
+
"epoch": 3.375,
|
| 1487 |
+
"frac_reward_zero_std": 1.0,
|
| 1488 |
+
"grad_norm": 0.03470289707183838,
|
| 1489 |
+
"kl": 0.2259281426668167,
|
| 1490 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 1491 |
+
"loss": 0.0005648203077726066,
|
| 1492 |
+
"num_tokens": 528368.0,
|
| 1493 |
+
"reward": -0.05000000074505806,
|
| 1494 |
+
"reward_std": 0.0,
|
| 1495 |
+
"rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
|
| 1496 |
+
"rewards/GeneratorRewardFunction/std": 0.0,
|
| 1497 |
+
"step": 81,
|
| 1498 |
+
"step_time": 12.270316088000072
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"clip_ratio/high_max": 0.0,
|
| 1502 |
+
"clip_ratio/high_mean": 0.0,
|
| 1503 |
+
"clip_ratio/low_mean": 0.0,
|
| 1504 |
+
"clip_ratio/low_min": 0.0,
|
| 1505 |
+
"clip_ratio/region_mean": 0.0,
|
| 1506 |
+
"entropy": 4.2660441398620605,
|
| 1507 |
+
"epoch": 3.4166666666666665,
|
| 1508 |
+
"grad_norm": 0.04197094962000847,
|
| 1509 |
+
"kl": 0.2527250647544861,
|
| 1510 |
+
"learning_rate": 1.6250000000000001e-06,
|
| 1511 |
+
"loss": 0.0006318126688711345,
|
| 1512 |
+
"step": 82,
|
| 1513 |
+
"step_time": 0.22032965800008242
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"clip_ratio/high_max": 0.0,
|
| 1517 |
+
"clip_ratio/high_mean": 0.0,
|
| 1518 |
+
"clip_ratio/low_mean": 0.0,
|
| 1519 |
+
"clip_ratio/low_min": 0.0,
|
| 1520 |
+
"clip_ratio/region_mean": 0.0,
|
| 1521 |
+
"entropy": 4.4261088371276855,
|
| 1522 |
+
"epoch": 3.4583333333333335,
|
| 1523 |
+
"grad_norm": 0.04092609882354736,
|
| 1524 |
+
"kl": 0.22817325592041016,
|
| 1525 |
+
"learning_rate": 1.5833333333333333e-06,
|
| 1526 |
+
"loss": 0.0005704330396838486,
|
| 1527 |
+
"step": 83,
|
| 1528 |
+
"step_time": 0.22033832899978734
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"clip_ratio/high_max": 0.0,
|
| 1532 |
+
"clip_ratio/high_mean": 0.0,
|
| 1533 |
+
"clip_ratio/low_mean": 0.0,
|
| 1534 |
+
"clip_ratio/low_min": 0.0,
|
| 1535 |
+
"clip_ratio/region_mean": 0.0,
|
| 1536 |
+
"entropy": 4.307797908782959,
|
| 1537 |
+
"epoch": 3.5,
|
| 1538 |
+
"grad_norm": 0.03771496191620827,
|
| 1539 |
+
"kl": 0.23303823173046112,
|
| 1540 |
+
"learning_rate": 1.5416666666666668e-06,
|
| 1541 |
+
"loss": 0.0005825955304317176,
|
| 1542 |
+
"step": 84,
|
| 1543 |
+
"step_time": 0.2203848580002159
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"clip_ratio/high_max": 0.0,
|
| 1547 |
+
"clip_ratio/high_mean": 0.0,
|
| 1548 |
+
"clip_ratio/low_mean": 0.0,
|
| 1549 |
+
"clip_ratio/low_min": 0.0,
|
| 1550 |
+
"clip_ratio/region_mean": 0.0,
|
| 1551 |
+
"completions/clipped_ratio": 1.0,
|
| 1552 |
+
"completions/max_length": 384.0,
|
| 1553 |
+
"completions/max_terminated_length": 0.0,
|
| 1554 |
+
"completions/mean_length": 384.0,
|
| 1555 |
+
"completions/mean_terminated_length": 0.0,
|
| 1556 |
+
"completions/min_length": 384.0,
|
| 1557 |
+
"completions/min_terminated_length": 0.0,
|
| 1558 |
+
"entropy": 4.24134635925293,
|
| 1559 |
+
"epoch": 3.5416666666666665,
|
| 1560 |
+
"frac_reward_zero_std": 0.5,
|
| 1561 |
+
"grad_norm": 2.4447741508483887,
|
| 1562 |
+
"kl": 0.23069417476654053,
|
| 1563 |
+
"learning_rate": 1.5e-06,
|
| 1564 |
+
"loss": 0.06225435063242912,
|
| 1565 |
+
"num_tokens": 553444.0,
|
| 1566 |
+
"reward": -0.06968750059604645,
|
| 1567 |
+
"reward_std": 0.07484358549118042,
|
| 1568 |
+
"rewards/GeneratorRewardFunction/mean": -0.06968750059604645,
|
| 1569 |
+
"rewards/GeneratorRewardFunction/std": 0.07484357804059982,
|
| 1570 |
+
"step": 85,
|
| 1571 |
+
"step_time": 12.391525436999473
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"clip_ratio/high_max": 0.0,
|
| 1575 |
+
"clip_ratio/high_mean": 0.0,
|
| 1576 |
+
"clip_ratio/low_mean": 0.0,
|
| 1577 |
+
"clip_ratio/low_min": 0.0,
|
| 1578 |
+
"clip_ratio/region_mean": 0.0,
|
| 1579 |
+
"entropy": 4.249276638031006,
|
| 1580 |
+
"epoch": 3.5833333333333335,
|
| 1581 |
+
"grad_norm": 2.462290048599243,
|
| 1582 |
+
"kl": 0.240711972117424,
|
| 1583 |
+
"learning_rate": 1.4583333333333335e-06,
|
| 1584 |
+
"loss": 0.06312751024961472,
|
| 1585 |
+
"step": 86,
|
| 1586 |
+
"step_time": 0.22052581699972507
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"clip_ratio/high_max": 0.0006510416860692203,
|
| 1590 |
+
"clip_ratio/high_mean": 0.0006510416860692203,
|
| 1591 |
+
"clip_ratio/low_mean": 0.0,
|
| 1592 |
+
"clip_ratio/low_min": 0.0,
|
| 1593 |
+
"clip_ratio/region_mean": 0.0006510416860692203,
|
| 1594 |
+
"entropy": 4.230101108551025,
|
| 1595 |
+
"epoch": 3.625,
|
| 1596 |
+
"grad_norm": 1.0876790285110474,
|
| 1597 |
+
"kl": 0.23581016063690186,
|
| 1598 |
+
"learning_rate": 1.4166666666666667e-06,
|
| 1599 |
+
"loss": -0.061899349093437195,
|
| 1600 |
+
"step": 87,
|
| 1601 |
+
"step_time": 0.22024468099971273
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"clip_ratio/high_max": 0.0006510416860692203,
|
| 1605 |
+
"clip_ratio/high_mean": 0.0006510416860692203,
|
| 1606 |
+
"clip_ratio/low_mean": 0.0,
|
| 1607 |
+
"clip_ratio/low_min": 0.0,
|
| 1608 |
+
"clip_ratio/region_mean": 0.0006510416860692203,
|
| 1609 |
+
"entropy": 4.287677764892578,
|
| 1610 |
+
"epoch": 3.6666666666666665,
|
| 1611 |
+
"grad_norm": 1.0642911195755005,
|
| 1612 |
+
"kl": 0.23578937351703644,
|
| 1613 |
+
"learning_rate": 1.3750000000000002e-06,
|
| 1614 |
+
"loss": -0.06103720888495445,
|
| 1615 |
+
"step": 88,
|
| 1616 |
+
"step_time": 0.21951269500004855
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"clip_ratio/high_max": 0.0,
|
| 1620 |
+
"clip_ratio/high_mean": 0.0,
|
| 1621 |
+
"clip_ratio/low_mean": 0.0,
|
| 1622 |
+
"clip_ratio/low_min": 0.0,
|
| 1623 |
+
"clip_ratio/region_mean": 0.0,
|
| 1624 |
+
"completions/clipped_ratio": 1.0,
|
| 1625 |
+
"completions/max_length": 384.0,
|
| 1626 |
+
"completions/max_terminated_length": 0.0,
|
| 1627 |
+
"completions/mean_length": 384.0,
|
| 1628 |
+
"completions/mean_terminated_length": 0.0,
|
| 1629 |
+
"completions/min_length": 384.0,
|
| 1630 |
+
"completions/min_terminated_length": 0.0,
|
| 1631 |
+
"entropy": 4.740299701690674,
|
| 1632 |
+
"epoch": 3.7083333333333335,
|
| 1633 |
+
"frac_reward_zero_std": 0.75,
|
| 1634 |
+
"grad_norm": 0.7542988657951355,
|
| 1635 |
+
"kl": 0.25239482522010803,
|
| 1636 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 1637 |
+
"loss": -0.02941715158522129,
|
| 1638 |
+
"num_tokens": 578952.0,
|
| 1639 |
+
"reward": -0.050312504172325134,
|
| 1640 |
+
"reward_std": 0.0012500007869675756,
|
| 1641 |
+
"rewards/GeneratorRewardFunction/mean": -0.050312504172325134,
|
| 1642 |
+
"rewards/GeneratorRewardFunction/std": 0.001249999739229679,
|
| 1643 |
+
"step": 89,
|
| 1644 |
+
"step_time": 12.352254893999998
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"clip_ratio/high_max": 0.0,
|
| 1648 |
+
"clip_ratio/high_mean": 0.0,
|
| 1649 |
+
"clip_ratio/low_mean": 0.0,
|
| 1650 |
+
"clip_ratio/low_min": 0.0,
|
| 1651 |
+
"clip_ratio/region_mean": 0.0,
|
| 1652 |
+
"entropy": 4.547545909881592,
|
| 1653 |
+
"epoch": 3.75,
|
| 1654 |
+
"grad_norm": 0.04400306195020676,
|
| 1655 |
+
"kl": 0.22126352787017822,
|
| 1656 |
+
"learning_rate": 1.2916666666666669e-06,
|
| 1657 |
+
"loss": 0.0005531588103622198,
|
| 1658 |
+
"step": 90,
|
| 1659 |
+
"step_time": 0.2290401660002317
|
| 1660 |
+
}
|
| 1661 |
+
],
|
| 1662 |
+
"logging_steps": 1,
|
| 1663 |
+
"max_steps": 120,
|
| 1664 |
+
"num_input_tokens_seen": 578952,
|
| 1665 |
+
"num_train_epochs": 5,
|
| 1666 |
+
"save_steps": 30,
|
| 1667 |
+
"stateful_callbacks": {
|
| 1668 |
+
"TrainerControl": {
|
| 1669 |
+
"args": {
|
| 1670 |
+
"should_epoch_stop": false,
|
| 1671 |
+
"should_evaluate": false,
|
| 1672 |
+
"should_log": false,
|
| 1673 |
+
"should_save": true,
|
| 1674 |
+
"should_training_stop": false
|
| 1675 |
+
},
|
| 1676 |
+
"attributes": {}
|
| 1677 |
+
}
|
| 1678 |
+
},
|
| 1679 |
+
"total_flos": 0.0,
|
| 1680 |
+
"train_batch_size": 4,
|
| 1681 |
+
"trial_name": null,
|
| 1682 |
+
"trial_params": null
|
| 1683 |
+
}
|
self_play_hf_l40s_full/round_004/generator_train/checkpoint-90/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b5dc67a9771849b225a6392756d4a1f5c89c7689c4ad63ef1c3f5f8879020ff
|
| 3 |
+
size 7249
|