Intermediate checkpoint upload step=90 (answerer_train)
Browse files- .gitattributes +1 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/chat_template.jinja +54 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/config.json +57 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/generation_config.json +13 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/model.safetensors +3 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/optimizer.pt +3 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/rng_state.pth +3 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/scheduler.pt +3 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/tokenizer.json +3 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/tokenizer_config.json +32 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/trainer_state.json +1683 -0
- self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -89,3 +89,4 @@ self_play_hf_l40s_full/round_005/generator_train/checkpoint-120/tokenizer.json f
|
|
| 89 |
self_play_hf_l40s_full/round_005/generator_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 90 |
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 91 |
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 89 |
self_play_hf_l40s_full/round_005/generator_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 90 |
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 91 |
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 92 |
+
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
self_play_hf_l40s_full/round_005/answerer_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_005/answerer_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_005/answerer_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_005/answerer_train/checkpoint-90/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a046fbdedf0fb58a8725f05539a2ef5725f8aff34ef30f6f229d144d6110b345
|
| 3 |
+
size 1976163472
|
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f3d58bbe8dd0ab8dba4896215355fa702c65d7baabfa9bf0d0a2253b287e23f
|
| 3 |
+
size 3952509771
|
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26f6240389ad2111a606d4aef5c6832d0136df7debe9f496847f25e20ecf886b
|
| 3 |
+
size 14645
|
self_play_hf_l40s_full/round_005/answerer_train/checkpoint-90/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:716b80a1d03fa94e9f2454097632e924f71019b15d068a8837fcbbbce1910893
|
| 3 |
+
size 1465
|
self_play_hf_l40s_full/round_005/answerer_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_005/answerer_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_005/answerer_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": 2.25,
|
| 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": 256.0,
|
| 20 |
+
"completions/max_terminated_length": 0.0,
|
| 21 |
+
"completions/mean_length": 256.0,
|
| 22 |
+
"completions/mean_terminated_length": 0.0,
|
| 23 |
+
"completions/min_length": 256.0,
|
| 24 |
+
"completions/min_terminated_length": 0.0,
|
| 25 |
+
"entropy": 1.7747459411621094,
|
| 26 |
+
"epoch": 0.025,
|
| 27 |
+
"frac_reward_zero_std": 0.5,
|
| 28 |
+
"grad_norm": 2.0766546726226807,
|
| 29 |
+
"kl": 0.0,
|
| 30 |
+
"learning_rate": 3e-06,
|
| 31 |
+
"loss": -0.10802466422319412,
|
| 32 |
+
"num_tokens": 9960.0,
|
| 33 |
+
"reward": 0.861758291721344,
|
| 34 |
+
"reward_std": 0.1182720735669136,
|
| 35 |
+
"rewards/AnswererRewardFunction/mean": 0.861758291721344,
|
| 36 |
+
"rewards/AnswererRewardFunction/std": 0.1182720735669136,
|
| 37 |
+
"step": 1,
|
| 38 |
+
"step_time": 7.9753973850001785
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"clip_ratio/high_max": 0.048828125,
|
| 42 |
+
"clip_ratio/high_mean": 0.048828125,
|
| 43 |
+
"clip_ratio/low_mean": 0.0,
|
| 44 |
+
"clip_ratio/low_min": 0.0,
|
| 45 |
+
"clip_ratio/region_mean": 0.048828125,
|
| 46 |
+
"entropy": 1.9674677848815918,
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"grad_norm": 1.2367463111877441,
|
| 49 |
+
"kl": 0.12442716211080551,
|
| 50 |
+
"learning_rate": 2.9750000000000003e-06,
|
| 51 |
+
"loss": -0.061602603644132614,
|
| 52 |
+
"step": 2,
|
| 53 |
+
"step_time": 0.1180487430001449
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"clip_ratio/high_max": 0.0,
|
| 57 |
+
"clip_ratio/high_mean": 0.0,
|
| 58 |
+
"clip_ratio/low_mean": 0.046875,
|
| 59 |
+
"clip_ratio/low_min": 0.046875,
|
| 60 |
+
"clip_ratio/region_mean": 0.046875,
|
| 61 |
+
"entropy": 2.1487021446228027,
|
| 62 |
+
"epoch": 0.075,
|
| 63 |
+
"grad_norm": 3.5523228645324707,
|
| 64 |
+
"kl": 0.12029533833265305,
|
| 65 |
+
"learning_rate": 2.9499999999999997e-06,
|
| 66 |
+
"loss": 0.15017201006412506,
|
| 67 |
+
"step": 3,
|
| 68 |
+
"step_time": 0.11217075500007923
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"clip_ratio/high_max": 0.0263671875,
|
| 72 |
+
"clip_ratio/high_mean": 0.0263671875,
|
| 73 |
+
"clip_ratio/low_mean": 0.03125,
|
| 74 |
+
"clip_ratio/low_min": 0.03125,
|
| 75 |
+
"clip_ratio/region_mean": 0.0576171875,
|
| 76 |
+
"entropy": 2.007148265838623,
|
| 77 |
+
"epoch": 0.1,
|
| 78 |
+
"grad_norm": 1.966639518737793,
|
| 79 |
+
"kl": 0.10988782346248627,
|
| 80 |
+
"learning_rate": 2.925e-06,
|
| 81 |
+
"loss": 0.02496827021241188,
|
| 82 |
+
"step": 4,
|
| 83 |
+
"step_time": 0.11107316000015999
|
| 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": 256.0,
|
| 93 |
+
"completions/max_terminated_length": 0.0,
|
| 94 |
+
"completions/mean_length": 256.0,
|
| 95 |
+
"completions/mean_terminated_length": 0.0,
|
| 96 |
+
"completions/min_length": 256.0,
|
| 97 |
+
"completions/min_terminated_length": 0.0,
|
| 98 |
+
"entropy": 2.053436756134033,
|
| 99 |
+
"epoch": 0.125,
|
| 100 |
+
"frac_reward_zero_std": 1.0,
|
| 101 |
+
"grad_norm": 0.023533064872026443,
|
| 102 |
+
"kl": 0.03712207078933716,
|
| 103 |
+
"learning_rate": 2.9e-06,
|
| 104 |
+
"loss": 9.280518133891746e-05,
|
| 105 |
+
"num_tokens": 19772.0,
|
| 106 |
+
"reward": 0.775258481502533,
|
| 107 |
+
"reward_std": 0.09679953753948212,
|
| 108 |
+
"rewards/AnswererRewardFunction/mean": 0.775258481502533,
|
| 109 |
+
"rewards/AnswererRewardFunction/std": 0.09679954499006271,
|
| 110 |
+
"step": 5,
|
| 111 |
+
"step_time": 7.949325107000732
|
| 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": 1.9749951362609863,
|
| 120 |
+
"epoch": 0.15,
|
| 121 |
+
"grad_norm": 0.019975125789642334,
|
| 122 |
+
"kl": 0.04075080528855324,
|
| 123 |
+
"learning_rate": 2.875e-06,
|
| 124 |
+
"loss": 0.00010187701263930649,
|
| 125 |
+
"step": 6,
|
| 126 |
+
"step_time": 0.11253355799999554
|
| 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": 1.901890754699707,
|
| 135 |
+
"epoch": 0.175,
|
| 136 |
+
"grad_norm": 0.040124714374542236,
|
| 137 |
+
"kl": 0.06329751014709473,
|
| 138 |
+
"learning_rate": 2.85e-06,
|
| 139 |
+
"loss": 0.00015824378351680934,
|
| 140 |
+
"step": 7,
|
| 141 |
+
"step_time": 0.1121210709998195
|
| 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": 1.9866925477981567,
|
| 150 |
+
"epoch": 0.2,
|
| 151 |
+
"grad_norm": 0.28295770287513733,
|
| 152 |
+
"kl": 0.07546718418598175,
|
| 153 |
+
"learning_rate": 2.825e-06,
|
| 154 |
+
"loss": 0.00018866793834604323,
|
| 155 |
+
"step": 8,
|
| 156 |
+
"step_time": 0.11143768600049953
|
| 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": 256.0,
|
| 166 |
+
"completions/max_terminated_length": 0.0,
|
| 167 |
+
"completions/mean_length": 256.0,
|
| 168 |
+
"completions/mean_terminated_length": 0.0,
|
| 169 |
+
"completions/min_length": 256.0,
|
| 170 |
+
"completions/min_terminated_length": 0.0,
|
| 171 |
+
"entropy": 1.887934923171997,
|
| 172 |
+
"epoch": 0.225,
|
| 173 |
+
"frac_reward_zero_std": 1.0,
|
| 174 |
+
"grad_norm": 0.02717243880033493,
|
| 175 |
+
"kl": 0.05967765301465988,
|
| 176 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 177 |
+
"loss": 0.00014919412205927074,
|
| 178 |
+
"num_tokens": 29608.0,
|
| 179 |
+
"reward": 0.8826282620429993,
|
| 180 |
+
"reward_std": 0.10600326210260391,
|
| 181 |
+
"rewards/AnswererRewardFunction/mean": 0.8826282620429993,
|
| 182 |
+
"rewards/AnswererRewardFunction/std": 0.10600324720144272,
|
| 183 |
+
"step": 9,
|
| 184 |
+
"step_time": 7.95751762799955
|
| 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 |
+
"entropy": 1.672785758972168,
|
| 193 |
+
"epoch": 0.25,
|
| 194 |
+
"grad_norm": 0.03259706869721413,
|
| 195 |
+
"kl": 0.07357163727283478,
|
| 196 |
+
"learning_rate": 2.775e-06,
|
| 197 |
+
"loss": 0.00018392910715192556,
|
| 198 |
+
"step": 10,
|
| 199 |
+
"step_time": 0.1110475139994378
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"clip_ratio/high_max": 0.0,
|
| 203 |
+
"clip_ratio/high_mean": 0.0,
|
| 204 |
+
"clip_ratio/low_mean": 0.0,
|
| 205 |
+
"clip_ratio/low_min": 0.0,
|
| 206 |
+
"clip_ratio/region_mean": 0.0,
|
| 207 |
+
"entropy": 2.054617404937744,
|
| 208 |
+
"epoch": 0.275,
|
| 209 |
+
"grad_norm": 0.034495558589696884,
|
| 210 |
+
"kl": 0.07685650885105133,
|
| 211 |
+
"learning_rate": 2.75e-06,
|
| 212 |
+
"loss": 0.00019214126223232597,
|
| 213 |
+
"step": 11,
|
| 214 |
+
"step_time": 0.1110302529996261
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"clip_ratio/high_max": 0.0,
|
| 218 |
+
"clip_ratio/high_mean": 0.0,
|
| 219 |
+
"clip_ratio/low_mean": 0.0,
|
| 220 |
+
"clip_ratio/low_min": 0.0,
|
| 221 |
+
"clip_ratio/region_mean": 0.0,
|
| 222 |
+
"entropy": 2.0550127029418945,
|
| 223 |
+
"epoch": 0.3,
|
| 224 |
+
"grad_norm": 0.029428452253341675,
|
| 225 |
+
"kl": 0.07294180989265442,
|
| 226 |
+
"learning_rate": 2.725e-06,
|
| 227 |
+
"loss": 0.0001823545026127249,
|
| 228 |
+
"step": 12,
|
| 229 |
+
"step_time": 0.11090638099994976
|
| 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": 256.0,
|
| 239 |
+
"completions/max_terminated_length": 0.0,
|
| 240 |
+
"completions/mean_length": 256.0,
|
| 241 |
+
"completions/mean_terminated_length": 0.0,
|
| 242 |
+
"completions/min_length": 256.0,
|
| 243 |
+
"completions/min_terminated_length": 0.0,
|
| 244 |
+
"entropy": 1.8360331058502197,
|
| 245 |
+
"epoch": 0.325,
|
| 246 |
+
"frac_reward_zero_std": 0.75,
|
| 247 |
+
"grad_norm": 2.9922423362731934,
|
| 248 |
+
"kl": 0.07551632076501846,
|
| 249 |
+
"learning_rate": 2.7e-06,
|
| 250 |
+
"loss": -0.06226314604282379,
|
| 251 |
+
"num_tokens": 39580.0,
|
| 252 |
+
"reward": 0.9046050310134888,
|
| 253 |
+
"reward_std": 0.08236623555421829,
|
| 254 |
+
"rewards/AnswererRewardFunction/mean": 0.9046050310134888,
|
| 255 |
+
"rewards/AnswererRewardFunction/std": 0.0823662281036377,
|
| 256 |
+
"step": 13,
|
| 257 |
+
"step_time": 7.9857917509998515
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"clip_ratio/high_max": 0.0,
|
| 261 |
+
"clip_ratio/high_mean": 0.0,
|
| 262 |
+
"clip_ratio/low_mean": 0.0,
|
| 263 |
+
"clip_ratio/low_min": 0.0,
|
| 264 |
+
"clip_ratio/region_mean": 0.0,
|
| 265 |
+
"entropy": 2.2757010459899902,
|
| 266 |
+
"epoch": 0.35,
|
| 267 |
+
"grad_norm": 0.032990459352731705,
|
| 268 |
+
"kl": 0.07481437176465988,
|
| 269 |
+
"learning_rate": 2.6750000000000002e-06,
|
| 270 |
+
"loss": 0.0001870359410531819,
|
| 271 |
+
"step": 14,
|
| 272 |
+
"step_time": 0.11140432199954375
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"clip_ratio/high_max": 0.0,
|
| 276 |
+
"clip_ratio/high_mean": 0.0,
|
| 277 |
+
"clip_ratio/low_mean": 0.0009765625,
|
| 278 |
+
"clip_ratio/low_min": 0.0009765625,
|
| 279 |
+
"clip_ratio/region_mean": 0.0009765625,
|
| 280 |
+
"entropy": 1.8182785511016846,
|
| 281 |
+
"epoch": 0.375,
|
| 282 |
+
"grad_norm": 0.8941988348960876,
|
| 283 |
+
"kl": 0.10425875335931778,
|
| 284 |
+
"learning_rate": 2.65e-06,
|
| 285 |
+
"loss": 0.0317252017557621,
|
| 286 |
+
"step": 15,
|
| 287 |
+
"step_time": 0.1321398499994757
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"clip_ratio/high_max": 0.0,
|
| 291 |
+
"clip_ratio/high_mean": 0.0,
|
| 292 |
+
"clip_ratio/low_mean": 0.0068359375,
|
| 293 |
+
"clip_ratio/low_min": 0.0068359375,
|
| 294 |
+
"clip_ratio/region_mean": 0.0068359375,
|
| 295 |
+
"entropy": 1.4916911125183105,
|
| 296 |
+
"epoch": 0.4,
|
| 297 |
+
"grad_norm": 0.8881436586380005,
|
| 298 |
+
"kl": 0.08638335019350052,
|
| 299 |
+
"learning_rate": 2.6250000000000003e-06,
|
| 300 |
+
"loss": 0.031212178990244865,
|
| 301 |
+
"step": 16,
|
| 302 |
+
"step_time": 0.11174384800051484
|
| 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": 256.0,
|
| 312 |
+
"completions/max_terminated_length": 0.0,
|
| 313 |
+
"completions/mean_length": 256.0,
|
| 314 |
+
"completions/mean_terminated_length": 0.0,
|
| 315 |
+
"completions/min_length": 256.0,
|
| 316 |
+
"completions/min_terminated_length": 0.0,
|
| 317 |
+
"entropy": 2.031963348388672,
|
| 318 |
+
"epoch": 0.425,
|
| 319 |
+
"frac_reward_zero_std": 1.0,
|
| 320 |
+
"grad_norm": 0.045982908457517624,
|
| 321 |
+
"kl": 0.10716703534126282,
|
| 322 |
+
"learning_rate": 2.6e-06,
|
| 323 |
+
"loss": 0.00026791758136823773,
|
| 324 |
+
"num_tokens": 49960.0,
|
| 325 |
+
"reward": 0.9323133230209351,
|
| 326 |
+
"reward_std": 0.0,
|
| 327 |
+
"rewards/AnswererRewardFunction/mean": 0.9323133230209351,
|
| 328 |
+
"rewards/AnswererRewardFunction/std": 0.0,
|
| 329 |
+
"step": 17,
|
| 330 |
+
"step_time": 7.907421807999526
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"clip_ratio/high_max": 0.0,
|
| 334 |
+
"clip_ratio/high_mean": 0.0,
|
| 335 |
+
"clip_ratio/low_mean": 0.0,
|
| 336 |
+
"clip_ratio/low_min": 0.0,
|
| 337 |
+
"clip_ratio/region_mean": 0.0,
|
| 338 |
+
"entropy": 1.8369736671447754,
|
| 339 |
+
"epoch": 0.45,
|
| 340 |
+
"grad_norm": 0.045823127031326294,
|
| 341 |
+
"kl": 0.10512565076351166,
|
| 342 |
+
"learning_rate": 2.575e-06,
|
| 343 |
+
"loss": 0.00026281410828232765,
|
| 344 |
+
"step": 18,
|
| 345 |
+
"step_time": 0.11191840300034528
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"clip_ratio/high_max": 0.0,
|
| 349 |
+
"clip_ratio/high_mean": 0.0,
|
| 350 |
+
"clip_ratio/low_mean": 0.0,
|
| 351 |
+
"clip_ratio/low_min": 0.0,
|
| 352 |
+
"clip_ratio/region_mean": 0.0,
|
| 353 |
+
"entropy": 1.9750962257385254,
|
| 354 |
+
"epoch": 0.475,
|
| 355 |
+
"grad_norm": 0.03864246979355812,
|
| 356 |
+
"kl": 0.0918373316526413,
|
| 357 |
+
"learning_rate": 2.55e-06,
|
| 358 |
+
"loss": 0.00022959333728067577,
|
| 359 |
+
"step": 19,
|
| 360 |
+
"step_time": 0.11069956699975592
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"clip_ratio/high_max": 0.0,
|
| 364 |
+
"clip_ratio/high_mean": 0.0,
|
| 365 |
+
"clip_ratio/low_mean": 0.0,
|
| 366 |
+
"clip_ratio/low_min": 0.0,
|
| 367 |
+
"clip_ratio/region_mean": 0.0,
|
| 368 |
+
"entropy": 2.031843662261963,
|
| 369 |
+
"epoch": 0.5,
|
| 370 |
+
"grad_norm": 0.05079154297709465,
|
| 371 |
+
"kl": 0.12332479655742645,
|
| 372 |
+
"learning_rate": 2.525e-06,
|
| 373 |
+
"loss": 0.0003083120100200176,
|
| 374 |
+
"step": 20,
|
| 375 |
+
"step_time": 0.11133921100008592
|
| 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": 256.0,
|
| 385 |
+
"completions/max_terminated_length": 0.0,
|
| 386 |
+
"completions/mean_length": 256.0,
|
| 387 |
+
"completions/mean_terminated_length": 0.0,
|
| 388 |
+
"completions/min_length": 256.0,
|
| 389 |
+
"completions/min_terminated_length": 0.0,
|
| 390 |
+
"entropy": 1.8194365501403809,
|
| 391 |
+
"epoch": 0.525,
|
| 392 |
+
"frac_reward_zero_std": 1.0,
|
| 393 |
+
"grad_norm": 0.23991960287094116,
|
| 394 |
+
"kl": 0.13239404559135437,
|
| 395 |
+
"learning_rate": 2.5e-06,
|
| 396 |
+
"loss": 0.00033098511630669236,
|
| 397 |
+
"num_tokens": 60148.0,
|
| 398 |
+
"reward": 0.819347620010376,
|
| 399 |
+
"reward_std": 0.11957819759845734,
|
| 400 |
+
"rewards/AnswererRewardFunction/mean": 0.819347620010376,
|
| 401 |
+
"rewards/AnswererRewardFunction/std": 0.11957821249961853,
|
| 402 |
+
"step": 21,
|
| 403 |
+
"step_time": 7.9142402379993655
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"clip_ratio/high_max": 0.0,
|
| 407 |
+
"clip_ratio/high_mean": 0.0,
|
| 408 |
+
"clip_ratio/low_mean": 0.0,
|
| 409 |
+
"clip_ratio/low_min": 0.0,
|
| 410 |
+
"clip_ratio/region_mean": 0.0,
|
| 411 |
+
"entropy": 1.333341360092163,
|
| 412 |
+
"epoch": 0.55,
|
| 413 |
+
"grad_norm": 0.04438238963484764,
|
| 414 |
+
"kl": 0.07663387060165405,
|
| 415 |
+
"learning_rate": 2.475e-06,
|
| 416 |
+
"loss": 0.00019158469513058662,
|
| 417 |
+
"step": 22,
|
| 418 |
+
"step_time": 0.11160182699950383
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"clip_ratio/high_max": 0.0,
|
| 422 |
+
"clip_ratio/high_mean": 0.0,
|
| 423 |
+
"clip_ratio/low_mean": 0.0,
|
| 424 |
+
"clip_ratio/low_min": 0.0,
|
| 425 |
+
"clip_ratio/region_mean": 0.0,
|
| 426 |
+
"entropy": 1.8490242958068848,
|
| 427 |
+
"epoch": 0.575,
|
| 428 |
+
"grad_norm": 0.04169222339987755,
|
| 429 |
+
"kl": 0.10077071189880371,
|
| 430 |
+
"learning_rate": 2.45e-06,
|
| 431 |
+
"loss": 0.00025192677276208997,
|
| 432 |
+
"step": 23,
|
| 433 |
+
"step_time": 0.11161970700050006
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"clip_ratio/high_max": 0.0,
|
| 437 |
+
"clip_ratio/high_mean": 0.0,
|
| 438 |
+
"clip_ratio/low_mean": 0.0,
|
| 439 |
+
"clip_ratio/low_min": 0.0,
|
| 440 |
+
"clip_ratio/region_mean": 0.0,
|
| 441 |
+
"entropy": 1.6197292804718018,
|
| 442 |
+
"epoch": 0.6,
|
| 443 |
+
"grad_norm": 0.05068081617355347,
|
| 444 |
+
"kl": 0.12023425102233887,
|
| 445 |
+
"learning_rate": 2.425e-06,
|
| 446 |
+
"loss": 0.0003005856415256858,
|
| 447 |
+
"step": 24,
|
| 448 |
+
"step_time": 0.11153809700044803
|
| 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": 256.0,
|
| 458 |
+
"completions/max_terminated_length": 0.0,
|
| 459 |
+
"completions/mean_length": 256.0,
|
| 460 |
+
"completions/mean_terminated_length": 0.0,
|
| 461 |
+
"completions/min_length": 256.0,
|
| 462 |
+
"completions/min_terminated_length": 0.0,
|
| 463 |
+
"entropy": 1.7150523662567139,
|
| 464 |
+
"epoch": 0.625,
|
| 465 |
+
"frac_reward_zero_std": 1.0,
|
| 466 |
+
"grad_norm": 0.03134142979979515,
|
| 467 |
+
"kl": 0.09788648784160614,
|
| 468 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 469 |
+
"loss": 0.0002447162405587733,
|
| 470 |
+
"num_tokens": 70092.0,
|
| 471 |
+
"reward": 0.8280093669891357,
|
| 472 |
+
"reward_std": 0.10857455432415009,
|
| 473 |
+
"rewards/AnswererRewardFunction/mean": 0.8280093669891357,
|
| 474 |
+
"rewards/AnswererRewardFunction/std": 0.10857456177473068,
|
| 475 |
+
"step": 25,
|
| 476 |
+
"step_time": 7.957390386000043
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"clip_ratio/high_max": 0.0,
|
| 480 |
+
"clip_ratio/high_mean": 0.0,
|
| 481 |
+
"clip_ratio/low_mean": 0.0,
|
| 482 |
+
"clip_ratio/low_min": 0.0,
|
| 483 |
+
"clip_ratio/region_mean": 0.0,
|
| 484 |
+
"entropy": 1.5895636081695557,
|
| 485 |
+
"epoch": 0.65,
|
| 486 |
+
"grad_norm": 0.04325919598340988,
|
| 487 |
+
"kl": 0.1107931137084961,
|
| 488 |
+
"learning_rate": 2.375e-06,
|
| 489 |
+
"loss": 0.00027698278427124023,
|
| 490 |
+
"step": 26,
|
| 491 |
+
"step_time": 0.1117739569999685
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"clip_ratio/high_max": 0.0,
|
| 495 |
+
"clip_ratio/high_mean": 0.0,
|
| 496 |
+
"clip_ratio/low_mean": 0.0,
|
| 497 |
+
"clip_ratio/low_min": 0.0,
|
| 498 |
+
"clip_ratio/region_mean": 0.0,
|
| 499 |
+
"entropy": 1.73920476436615,
|
| 500 |
+
"epoch": 0.675,
|
| 501 |
+
"grad_norm": 0.03872699663043022,
|
| 502 |
+
"kl": 0.10983294248580933,
|
| 503 |
+
"learning_rate": 2.35e-06,
|
| 504 |
+
"loss": 0.00027458235854282975,
|
| 505 |
+
"step": 27,
|
| 506 |
+
"step_time": 0.11149289300010423
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"clip_ratio/high_max": 0.0,
|
| 510 |
+
"clip_ratio/high_mean": 0.0,
|
| 511 |
+
"clip_ratio/low_mean": 0.0,
|
| 512 |
+
"clip_ratio/low_min": 0.0,
|
| 513 |
+
"clip_ratio/region_mean": 0.0,
|
| 514 |
+
"entropy": 1.898261308670044,
|
| 515 |
+
"epoch": 0.7,
|
| 516 |
+
"grad_norm": 0.04079338535666466,
|
| 517 |
+
"kl": 0.11259151250123978,
|
| 518 |
+
"learning_rate": 2.325e-06,
|
| 519 |
+
"loss": 0.00028147880220785737,
|
| 520 |
+
"step": 28,
|
| 521 |
+
"step_time": 0.11117797499991866
|
| 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": 256.0,
|
| 531 |
+
"completions/max_terminated_length": 0.0,
|
| 532 |
+
"completions/mean_length": 256.0,
|
| 533 |
+
"completions/mean_terminated_length": 0.0,
|
| 534 |
+
"completions/min_length": 256.0,
|
| 535 |
+
"completions/min_terminated_length": 0.0,
|
| 536 |
+
"entropy": 1.9487004280090332,
|
| 537 |
+
"epoch": 0.725,
|
| 538 |
+
"frac_reward_zero_std": 1.0,
|
| 539 |
+
"grad_norm": 0.047494642436504364,
|
| 540 |
+
"kl": 0.11588452756404877,
|
| 541 |
+
"learning_rate": 2.3000000000000004e-06,
|
| 542 |
+
"loss": 0.0002897112863138318,
|
| 543 |
+
"num_tokens": 80364.0,
|
| 544 |
+
"reward": 0.8668582439422607,
|
| 545 |
+
"reward_std": 0.1170896589756012,
|
| 546 |
+
"rewards/AnswererRewardFunction/mean": 0.8668582439422607,
|
| 547 |
+
"rewards/AnswererRewardFunction/std": 0.1170896664261818,
|
| 548 |
+
"step": 29,
|
| 549 |
+
"step_time": 7.972899151000092
|
| 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": 1.7704535722732544,
|
| 558 |
+
"epoch": 0.75,
|
| 559 |
+
"grad_norm": 0.02996332012116909,
|
| 560 |
+
"kl": 0.09660449624061584,
|
| 561 |
+
"learning_rate": 2.275e-06,
|
| 562 |
+
"loss": 0.00024151125398930162,
|
| 563 |
+
"step": 30,
|
| 564 |
+
"step_time": 0.11398699500023213
|
| 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": 1.766472339630127,
|
| 573 |
+
"epoch": 0.775,
|
| 574 |
+
"grad_norm": 0.03504747897386551,
|
| 575 |
+
"kl": 0.08477599918842316,
|
| 576 |
+
"learning_rate": 2.25e-06,
|
| 577 |
+
"loss": 0.00021194000146351755,
|
| 578 |
+
"step": 31,
|
| 579 |
+
"step_time": 0.11211207500036835
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"clip_ratio/high_max": 0.0,
|
| 583 |
+
"clip_ratio/high_mean": 0.0,
|
| 584 |
+
"clip_ratio/low_mean": 0.0,
|
| 585 |
+
"clip_ratio/low_min": 0.0,
|
| 586 |
+
"clip_ratio/region_mean": 0.0,
|
| 587 |
+
"entropy": 1.4578940868377686,
|
| 588 |
+
"epoch": 0.8,
|
| 589 |
+
"grad_norm": 0.02072232775390148,
|
| 590 |
+
"kl": 0.06009065359830856,
|
| 591 |
+
"learning_rate": 2.2250000000000003e-06,
|
| 592 |
+
"loss": 0.00015022660954855382,
|
| 593 |
+
"step": 32,
|
| 594 |
+
"step_time": 0.1105231419996926
|
| 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": 256.0,
|
| 604 |
+
"completions/max_terminated_length": 0.0,
|
| 605 |
+
"completions/mean_length": 256.0,
|
| 606 |
+
"completions/mean_terminated_length": 0.0,
|
| 607 |
+
"completions/min_length": 256.0,
|
| 608 |
+
"completions/min_terminated_length": 0.0,
|
| 609 |
+
"entropy": 1.7355934381484985,
|
| 610 |
+
"epoch": 0.825,
|
| 611 |
+
"frac_reward_zero_std": 0.75,
|
| 612 |
+
"grad_norm": 1.4318349361419678,
|
| 613 |
+
"kl": 0.10250318050384521,
|
| 614 |
+
"learning_rate": 2.1999999999999997e-06,
|
| 615 |
+
"loss": -0.06233750656247139,
|
| 616 |
+
"num_tokens": 90796.0,
|
| 617 |
+
"reward": 0.9159495830535889,
|
| 618 |
+
"reward_std": 0.06545510143041611,
|
| 619 |
+
"rewards/AnswererRewardFunction/mean": 0.9159495830535889,
|
| 620 |
+
"rewards/AnswererRewardFunction/std": 0.0654551088809967,
|
| 621 |
+
"step": 33,
|
| 622 |
+
"step_time": 7.989859212000738
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"clip_ratio/high_max": 0.0,
|
| 626 |
+
"clip_ratio/high_mean": 0.0,
|
| 627 |
+
"clip_ratio/low_mean": 0.0009765625,
|
| 628 |
+
"clip_ratio/low_min": 0.0009765625,
|
| 629 |
+
"clip_ratio/region_mean": 0.0009765625,
|
| 630 |
+
"entropy": 1.732600450515747,
|
| 631 |
+
"epoch": 0.85,
|
| 632 |
+
"grad_norm": 2.944581985473633,
|
| 633 |
+
"kl": 0.08287997543811798,
|
| 634 |
+
"learning_rate": 2.175e-06,
|
| 635 |
+
"loss": 0.06262950599193573,
|
| 636 |
+
"step": 34,
|
| 637 |
+
"step_time": 0.11203606500021124
|
| 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": 1.9220362901687622,
|
| 646 |
+
"epoch": 0.875,
|
| 647 |
+
"grad_norm": 0.03256967291235924,
|
| 648 |
+
"kl": 0.08927878737449646,
|
| 649 |
+
"learning_rate": 2.15e-06,
|
| 650 |
+
"loss": 0.00022319695563055575,
|
| 651 |
+
"step": 35,
|
| 652 |
+
"step_time": 0.1124586719997751
|
| 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": 1.7540044784545898,
|
| 661 |
+
"epoch": 0.9,
|
| 662 |
+
"grad_norm": 0.03057776764035225,
|
| 663 |
+
"kl": 0.08942355215549469,
|
| 664 |
+
"learning_rate": 2.125e-06,
|
| 665 |
+
"loss": 0.0002235588908661157,
|
| 666 |
+
"step": 36,
|
| 667 |
+
"step_time": 0.11161630500009778
|
| 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": 256.0,
|
| 677 |
+
"completions/max_terminated_length": 0.0,
|
| 678 |
+
"completions/mean_length": 256.0,
|
| 679 |
+
"completions/mean_terminated_length": 0.0,
|
| 680 |
+
"completions/min_length": 256.0,
|
| 681 |
+
"completions/min_terminated_length": 0.0,
|
| 682 |
+
"entropy": 1.8761208057403564,
|
| 683 |
+
"epoch": 0.925,
|
| 684 |
+
"frac_reward_zero_std": 1.0,
|
| 685 |
+
"grad_norm": 0.053888410329818726,
|
| 686 |
+
"kl": 0.10505490750074387,
|
| 687 |
+
"learning_rate": 2.1e-06,
|
| 688 |
+
"loss": 0.00026263727340847254,
|
| 689 |
+
"num_tokens": 100696.0,
|
| 690 |
+
"reward": 0.8227691054344177,
|
| 691 |
+
"reward_std": 0.11510813981294632,
|
| 692 |
+
"rewards/AnswererRewardFunction/mean": 0.8227691054344177,
|
| 693 |
+
"rewards/AnswererRewardFunction/std": 0.11510813981294632,
|
| 694 |
+
"step": 37,
|
| 695 |
+
"step_time": 7.976926989000276
|
| 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": 1.6150931119918823,
|
| 704 |
+
"epoch": 0.95,
|
| 705 |
+
"grad_norm": 0.028759539127349854,
|
| 706 |
+
"kl": 0.0988161563873291,
|
| 707 |
+
"learning_rate": 2.075e-06,
|
| 708 |
+
"loss": 0.0002470404142513871,
|
| 709 |
+
"step": 38,
|
| 710 |
+
"step_time": 0.1113082760002726
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"clip_ratio/high_max": 0.0,
|
| 714 |
+
"clip_ratio/high_mean": 0.0,
|
| 715 |
+
"clip_ratio/low_mean": 0.0,
|
| 716 |
+
"clip_ratio/low_min": 0.0,
|
| 717 |
+
"clip_ratio/region_mean": 0.0,
|
| 718 |
+
"entropy": 2.035503387451172,
|
| 719 |
+
"epoch": 0.975,
|
| 720 |
+
"grad_norm": 0.04202085733413696,
|
| 721 |
+
"kl": 0.11065234243869781,
|
| 722 |
+
"learning_rate": 2.0500000000000003e-06,
|
| 723 |
+
"loss": 0.00027663083164952695,
|
| 724 |
+
"step": 39,
|
| 725 |
+
"step_time": 0.11134837800000241
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"clip_ratio/high_max": 0.0,
|
| 729 |
+
"clip_ratio/high_mean": 0.0,
|
| 730 |
+
"clip_ratio/low_mean": 0.0,
|
| 731 |
+
"clip_ratio/low_min": 0.0,
|
| 732 |
+
"clip_ratio/region_mean": 0.0,
|
| 733 |
+
"entropy": 1.9588853120803833,
|
| 734 |
+
"epoch": 1.0,
|
| 735 |
+
"grad_norm": 0.029340602457523346,
|
| 736 |
+
"kl": 0.09918107837438583,
|
| 737 |
+
"learning_rate": 2.025e-06,
|
| 738 |
+
"loss": 0.0002479527029208839,
|
| 739 |
+
"step": 40,
|
| 740 |
+
"step_time": 0.11250619200018264
|
| 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": 256.0,
|
| 750 |
+
"completions/max_terminated_length": 0.0,
|
| 751 |
+
"completions/mean_length": 256.0,
|
| 752 |
+
"completions/mean_terminated_length": 0.0,
|
| 753 |
+
"completions/min_length": 256.0,
|
| 754 |
+
"completions/min_terminated_length": 0.0,
|
| 755 |
+
"entropy": 1.903963327407837,
|
| 756 |
+
"epoch": 1.025,
|
| 757 |
+
"frac_reward_zero_std": 0.25,
|
| 758 |
+
"grad_norm": 2.2654471397399902,
|
| 759 |
+
"kl": 0.10239098221063614,
|
| 760 |
+
"learning_rate": 2e-06,
|
| 761 |
+
"loss": 0.035676077008247375,
|
| 762 |
+
"num_tokens": 110284.0,
|
| 763 |
+
"reward": 0.7121027708053589,
|
| 764 |
+
"reward_std": 0.12839390337467194,
|
| 765 |
+
"rewards/AnswererRewardFunction/mean": 0.7121027708053589,
|
| 766 |
+
"rewards/AnswererRewardFunction/std": 0.12839393317699432,
|
| 767 |
+
"step": 41,
|
| 768 |
+
"step_time": 7.889583379000214
|
| 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": 2.0613226890563965,
|
| 777 |
+
"epoch": 1.05,
|
| 778 |
+
"grad_norm": 3.5304839611053467,
|
| 779 |
+
"kl": 0.10499992966651917,
|
| 780 |
+
"learning_rate": 1.975e-06,
|
| 781 |
+
"loss": 0.13377703726291656,
|
| 782 |
+
"step": 42,
|
| 783 |
+
"step_time": 0.11177234499973565
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"clip_ratio/high_max": 0.0146484375,
|
| 787 |
+
"clip_ratio/high_mean": 0.0146484375,
|
| 788 |
+
"clip_ratio/low_mean": 0.0,
|
| 789 |
+
"clip_ratio/low_min": 0.0,
|
| 790 |
+
"clip_ratio/region_mean": 0.0146484375,
|
| 791 |
+
"entropy": 1.7200109958648682,
|
| 792 |
+
"epoch": 1.075,
|
| 793 |
+
"grad_norm": 2.346640110015869,
|
| 794 |
+
"kl": 0.11238373816013336,
|
| 795 |
+
"learning_rate": 1.95e-06,
|
| 796 |
+
"loss": -0.1459183692932129,
|
| 797 |
+
"step": 43,
|
| 798 |
+
"step_time": 0.11070189299971389
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"clip_ratio/high_max": 0.0146484375,
|
| 802 |
+
"clip_ratio/high_mean": 0.0146484375,
|
| 803 |
+
"clip_ratio/low_mean": 0.00390625,
|
| 804 |
+
"clip_ratio/low_min": 0.00390625,
|
| 805 |
+
"clip_ratio/region_mean": 0.0185546875,
|
| 806 |
+
"entropy": 1.9567177295684814,
|
| 807 |
+
"epoch": 1.1,
|
| 808 |
+
"grad_norm": 3.576796531677246,
|
| 809 |
+
"kl": 0.12578752636909485,
|
| 810 |
+
"learning_rate": 1.925e-06,
|
| 811 |
+
"loss": -0.021488450467586517,
|
| 812 |
+
"step": 44,
|
| 813 |
+
"step_time": 0.11009783999998035
|
| 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": 256.0,
|
| 823 |
+
"completions/max_terminated_length": 0.0,
|
| 824 |
+
"completions/mean_length": 256.0,
|
| 825 |
+
"completions/mean_terminated_length": 0.0,
|
| 826 |
+
"completions/min_length": 256.0,
|
| 827 |
+
"completions/min_terminated_length": 0.0,
|
| 828 |
+
"entropy": 2.0115418434143066,
|
| 829 |
+
"epoch": 1.125,
|
| 830 |
+
"frac_reward_zero_std": 0.75,
|
| 831 |
+
"grad_norm": 0.046743664890527725,
|
| 832 |
+
"kl": 0.11043193191289902,
|
| 833 |
+
"learning_rate": 1.9e-06,
|
| 834 |
+
"loss": 0.00027607983793132007,
|
| 835 |
+
"num_tokens": 120284.0,
|
| 836 |
+
"reward": 0.8719930648803711,
|
| 837 |
+
"reward_std": 0.10185166448354721,
|
| 838 |
+
"rewards/AnswererRewardFunction/mean": 0.8719930648803711,
|
| 839 |
+
"rewards/AnswererRewardFunction/std": 0.10185166448354721,
|
| 840 |
+
"step": 45,
|
| 841 |
+
"step_time": 8.05356791899976
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"clip_ratio/high_max": 0.001953125,
|
| 845 |
+
"clip_ratio/high_mean": 0.001953125,
|
| 846 |
+
"clip_ratio/low_mean": 0.0,
|
| 847 |
+
"clip_ratio/low_min": 0.0,
|
| 848 |
+
"clip_ratio/region_mean": 0.001953125,
|
| 849 |
+
"entropy": 1.8525512218475342,
|
| 850 |
+
"epoch": 1.15,
|
| 851 |
+
"grad_norm": 0.9486414194107056,
|
| 852 |
+
"kl": 0.12013739347457886,
|
| 853 |
+
"learning_rate": 1.875e-06,
|
| 854 |
+
"loss": -0.03101479448378086,
|
| 855 |
+
"step": 46,
|
| 856 |
+
"step_time": 0.11169548499947268
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"clip_ratio/high_max": 0.0048828125,
|
| 860 |
+
"clip_ratio/high_mean": 0.0048828125,
|
| 861 |
+
"clip_ratio/low_mean": 0.0,
|
| 862 |
+
"clip_ratio/low_min": 0.0,
|
| 863 |
+
"clip_ratio/region_mean": 0.0048828125,
|
| 864 |
+
"entropy": 2.05437970161438,
|
| 865 |
+
"epoch": 1.175,
|
| 866 |
+
"grad_norm": 0.9893139600753784,
|
| 867 |
+
"kl": 0.0978676825761795,
|
| 868 |
+
"learning_rate": 1.85e-06,
|
| 869 |
+
"loss": -0.03111003153026104,
|
| 870 |
+
"step": 47,
|
| 871 |
+
"step_time": 0.1112122829999862
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"clip_ratio/high_max": 0.0048828125,
|
| 875 |
+
"clip_ratio/high_mean": 0.0048828125,
|
| 876 |
+
"clip_ratio/low_mean": 0.0,
|
| 877 |
+
"clip_ratio/low_min": 0.0,
|
| 878 |
+
"clip_ratio/region_mean": 0.0048828125,
|
| 879 |
+
"entropy": 1.9310542345046997,
|
| 880 |
+
"epoch": 1.2,
|
| 881 |
+
"grad_norm": 2.8787803649902344,
|
| 882 |
+
"kl": 0.10207061469554901,
|
| 883 |
+
"learning_rate": 1.8249999999999999e-06,
|
| 884 |
+
"loss": 0.06266620755195618,
|
| 885 |
+
"step": 48,
|
| 886 |
+
"step_time": 0.12508205599988287
|
| 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": 256.0,
|
| 896 |
+
"completions/max_terminated_length": 0.0,
|
| 897 |
+
"completions/mean_length": 256.0,
|
| 898 |
+
"completions/mean_terminated_length": 0.0,
|
| 899 |
+
"completions/min_length": 256.0,
|
| 900 |
+
"completions/min_terminated_length": 0.0,
|
| 901 |
+
"entropy": 1.9703929424285889,
|
| 902 |
+
"epoch": 1.225,
|
| 903 |
+
"frac_reward_zero_std": 0.5,
|
| 904 |
+
"grad_norm": 3.0901620388031006,
|
| 905 |
+
"kl": 0.0971868485212326,
|
| 906 |
+
"learning_rate": 1.8e-06,
|
| 907 |
+
"loss": 0.06205549091100693,
|
| 908 |
+
"num_tokens": 130188.0,
|
| 909 |
+
"reward": 0.7701587677001953,
|
| 910 |
+
"reward_std": 0.17226041853427887,
|
| 911 |
+
"rewards/AnswererRewardFunction/mean": 0.7701587677001953,
|
| 912 |
+
"rewards/AnswererRewardFunction/std": 0.17226043343544006,
|
| 913 |
+
"step": 49,
|
| 914 |
+
"step_time": 7.936510957000792
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"clip_ratio/high_max": 0.0048828125,
|
| 918 |
+
"clip_ratio/high_mean": 0.0048828125,
|
| 919 |
+
"clip_ratio/low_mean": 0.0,
|
| 920 |
+
"clip_ratio/low_min": 0.0,
|
| 921 |
+
"clip_ratio/region_mean": 0.0048828125,
|
| 922 |
+
"entropy": 1.9568400382995605,
|
| 923 |
+
"epoch": 1.25,
|
| 924 |
+
"grad_norm": 1.540932297706604,
|
| 925 |
+
"kl": 0.11889415979385376,
|
| 926 |
+
"learning_rate": 1.7750000000000002e-06,
|
| 927 |
+
"loss": -0.09337170422077179,
|
| 928 |
+
"step": 50,
|
| 929 |
+
"step_time": 0.10965072100043471
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"clip_ratio/high_max": 0.0068359375,
|
| 933 |
+
"clip_ratio/high_mean": 0.0068359375,
|
| 934 |
+
"clip_ratio/low_mean": 0.0009765625,
|
| 935 |
+
"clip_ratio/low_min": 0.0009765625,
|
| 936 |
+
"clip_ratio/region_mean": 0.0078125,
|
| 937 |
+
"entropy": 1.9066113233566284,
|
| 938 |
+
"epoch": 1.275,
|
| 939 |
+
"grad_norm": 3.2399439811706543,
|
| 940 |
+
"kl": 0.10549439489841461,
|
| 941 |
+
"learning_rate": 1.7500000000000002e-06,
|
| 942 |
+
"loss": 0.03147980943322182,
|
| 943 |
+
"step": 51,
|
| 944 |
+
"step_time": 0.10916036200069357
|
| 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": 1.7676732540130615,
|
| 953 |
+
"epoch": 1.3,
|
| 954 |
+
"grad_norm": 0.038625847548246384,
|
| 955 |
+
"kl": 0.1173200011253357,
|
| 956 |
+
"learning_rate": 1.725e-06,
|
| 957 |
+
"loss": 0.00029329999233596027,
|
| 958 |
+
"step": 52,
|
| 959 |
+
"step_time": 0.10919338299936499
|
| 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": 256.0,
|
| 969 |
+
"completions/max_terminated_length": 0.0,
|
| 970 |
+
"completions/mean_length": 256.0,
|
| 971 |
+
"completions/mean_terminated_length": 0.0,
|
| 972 |
+
"completions/min_length": 256.0,
|
| 973 |
+
"completions/min_terminated_length": 0.0,
|
| 974 |
+
"entropy": 1.6084643602371216,
|
| 975 |
+
"epoch": 1.325,
|
| 976 |
+
"frac_reward_zero_std": 1.0,
|
| 977 |
+
"grad_norm": 0.04859758913516998,
|
| 978 |
+
"kl": 0.11696670204401016,
|
| 979 |
+
"learning_rate": 1.7e-06,
|
| 980 |
+
"loss": 0.00029241672018542886,
|
| 981 |
+
"num_tokens": 140400.0,
|
| 982 |
+
"reward": 0.8848026990890503,
|
| 983 |
+
"reward_std": 0.08498957008123398,
|
| 984 |
+
"rewards/AnswererRewardFunction/mean": 0.8848026990890503,
|
| 985 |
+
"rewards/AnswererRewardFunction/std": 0.08498956263065338,
|
| 986 |
+
"step": 53,
|
| 987 |
+
"step_time": 7.985873733000517
|
| 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": 1.9126839637756348,
|
| 996 |
+
"epoch": 1.35,
|
| 997 |
+
"grad_norm": 0.05618106201291084,
|
| 998 |
+
"kl": 0.12067677080631256,
|
| 999 |
+
"learning_rate": 1.675e-06,
|
| 1000 |
+
"loss": 0.0003016919654328376,
|
| 1001 |
+
"step": 54,
|
| 1002 |
+
"step_time": 0.11533255099948292
|
| 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": 1.9060513973236084,
|
| 1011 |
+
"epoch": 1.375,
|
| 1012 |
+
"grad_norm": 0.03235335648059845,
|
| 1013 |
+
"kl": 0.10564382374286652,
|
| 1014 |
+
"learning_rate": 1.65e-06,
|
| 1015 |
+
"loss": 0.0002641095488797873,
|
| 1016 |
+
"step": 55,
|
| 1017 |
+
"step_time": 0.11275608800042392
|
| 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": 1.9660732746124268,
|
| 1026 |
+
"epoch": 1.4,
|
| 1027 |
+
"grad_norm": 0.03671698272228241,
|
| 1028 |
+
"kl": 0.10332442820072174,
|
| 1029 |
+
"learning_rate": 1.625e-06,
|
| 1030 |
+
"loss": 0.00025831107632257044,
|
| 1031 |
+
"step": 56,
|
| 1032 |
+
"step_time": 0.11206929399941146
|
| 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": 256.0,
|
| 1042 |
+
"completions/max_terminated_length": 0.0,
|
| 1043 |
+
"completions/mean_length": 256.0,
|
| 1044 |
+
"completions/mean_terminated_length": 0.0,
|
| 1045 |
+
"completions/min_length": 256.0,
|
| 1046 |
+
"completions/min_terminated_length": 0.0,
|
| 1047 |
+
"entropy": 2.006600856781006,
|
| 1048 |
+
"epoch": 1.425,
|
| 1049 |
+
"frac_reward_zero_std": 1.0,
|
| 1050 |
+
"grad_norm": 0.04220745339989662,
|
| 1051 |
+
"kl": 0.14282318949699402,
|
| 1052 |
+
"learning_rate": 1.6e-06,
|
| 1053 |
+
"loss": 0.00035705792834050953,
|
| 1054 |
+
"num_tokens": 150720.0,
|
| 1055 |
+
"reward": 0.9323133230209351,
|
| 1056 |
+
"reward_std": 0.0,
|
| 1057 |
+
"rewards/AnswererRewardFunction/mean": 0.9323133230209351,
|
| 1058 |
+
"rewards/AnswererRewardFunction/std": 0.0,
|
| 1059 |
+
"step": 57,
|
| 1060 |
+
"step_time": 7.929392387000007
|
| 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": 1.845897912979126,
|
| 1069 |
+
"epoch": 1.45,
|
| 1070 |
+
"grad_norm": 0.03133726865053177,
|
| 1071 |
+
"kl": 0.09734172374010086,
|
| 1072 |
+
"learning_rate": 1.5750000000000002e-06,
|
| 1073 |
+
"loss": 0.00024335432681255043,
|
| 1074 |
+
"step": 58,
|
| 1075 |
+
"step_time": 0.11144289299954835
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"clip_ratio/high_max": 0.0,
|
| 1079 |
+
"clip_ratio/high_mean": 0.0,
|
| 1080 |
+
"clip_ratio/low_mean": 0.0,
|
| 1081 |
+
"clip_ratio/low_min": 0.0,
|
| 1082 |
+
"clip_ratio/region_mean": 0.0,
|
| 1083 |
+
"entropy": 1.9323326349258423,
|
| 1084 |
+
"epoch": 1.475,
|
| 1085 |
+
"grad_norm": 0.031240344047546387,
|
| 1086 |
+
"kl": 0.1022108644247055,
|
| 1087 |
+
"learning_rate": 1.5500000000000002e-06,
|
| 1088 |
+
"loss": 0.0002555271494202316,
|
| 1089 |
+
"step": 59,
|
| 1090 |
+
"step_time": 0.11132388100031676
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"clip_ratio/high_max": 0.0,
|
| 1094 |
+
"clip_ratio/high_mean": 0.0,
|
| 1095 |
+
"clip_ratio/low_mean": 0.0,
|
| 1096 |
+
"clip_ratio/low_min": 0.0,
|
| 1097 |
+
"clip_ratio/region_mean": 0.0,
|
| 1098 |
+
"entropy": 1.9500852823257446,
|
| 1099 |
+
"epoch": 1.5,
|
| 1100 |
+
"grad_norm": 0.04029398411512375,
|
| 1101 |
+
"kl": 0.12107711285352707,
|
| 1102 |
+
"learning_rate": 1.525e-06,
|
| 1103 |
+
"loss": 0.0003026928170584142,
|
| 1104 |
+
"step": 60,
|
| 1105 |
+
"step_time": 0.11092958100016403
|
| 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": 256.0,
|
| 1115 |
+
"completions/max_terminated_length": 0.0,
|
| 1116 |
+
"completions/mean_length": 256.0,
|
| 1117 |
+
"completions/mean_terminated_length": 0.0,
|
| 1118 |
+
"completions/min_length": 256.0,
|
| 1119 |
+
"completions/min_terminated_length": 0.0,
|
| 1120 |
+
"entropy": 1.9172897338867188,
|
| 1121 |
+
"epoch": 1.525,
|
| 1122 |
+
"frac_reward_zero_std": 0.75,
|
| 1123 |
+
"grad_norm": 3.0262608528137207,
|
| 1124 |
+
"kl": 0.1267942488193512,
|
| 1125 |
+
"learning_rate": 1.5e-06,
|
| 1126 |
+
"loss": 0.06276892125606537,
|
| 1127 |
+
"num_tokens": 160920.0,
|
| 1128 |
+
"reward": 0.8680921792984009,
|
| 1129 |
+
"reward_std": 0.13140177726745605,
|
| 1130 |
+
"rewards/AnswererRewardFunction/mean": 0.8680921792984009,
|
| 1131 |
+
"rewards/AnswererRewardFunction/std": 0.13140179216861725,
|
| 1132 |
+
"step": 61,
|
| 1133 |
+
"step_time": 7.9190442310000435
|
| 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": 2.047105550765991,
|
| 1142 |
+
"epoch": 1.55,
|
| 1143 |
+
"grad_norm": 0.8910013437271118,
|
| 1144 |
+
"kl": 0.11219840496778488,
|
| 1145 |
+
"learning_rate": 1.4749999999999999e-06,
|
| 1146 |
+
"loss": -0.031001705676317215,
|
| 1147 |
+
"step": 62,
|
| 1148 |
+
"step_time": 0.11164996400020755
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"clip_ratio/high_max": 0.0,
|
| 1152 |
+
"clip_ratio/high_mean": 0.0,
|
| 1153 |
+
"clip_ratio/low_mean": 0.0,
|
| 1154 |
+
"clip_ratio/low_min": 0.0,
|
| 1155 |
+
"clip_ratio/region_mean": 0.0,
|
| 1156 |
+
"entropy": 1.8467944860458374,
|
| 1157 |
+
"epoch": 1.575,
|
| 1158 |
+
"grad_norm": 0.06872954964637756,
|
| 1159 |
+
"kl": 0.12846754491329193,
|
| 1160 |
+
"learning_rate": 1.45e-06,
|
| 1161 |
+
"loss": 0.0003211688599549234,
|
| 1162 |
+
"step": 63,
|
| 1163 |
+
"step_time": 0.11100759100008872
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"clip_ratio/high_max": 0.0,
|
| 1167 |
+
"clip_ratio/high_mean": 0.0,
|
| 1168 |
+
"clip_ratio/low_mean": 0.0,
|
| 1169 |
+
"clip_ratio/low_min": 0.0,
|
| 1170 |
+
"clip_ratio/region_mean": 0.0,
|
| 1171 |
+
"entropy": 2.0865182876586914,
|
| 1172 |
+
"epoch": 1.6,
|
| 1173 |
+
"grad_norm": 0.9386709332466125,
|
| 1174 |
+
"kl": 0.12258686870336533,
|
| 1175 |
+
"learning_rate": 1.425e-06,
|
| 1176 |
+
"loss": -0.030893009155988693,
|
| 1177 |
+
"step": 64,
|
| 1178 |
+
"step_time": 0.11031587699926604
|
| 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": 256.0,
|
| 1188 |
+
"completions/max_terminated_length": 0.0,
|
| 1189 |
+
"completions/mean_length": 256.0,
|
| 1190 |
+
"completions/mean_terminated_length": 0.0,
|
| 1191 |
+
"completions/min_length": 256.0,
|
| 1192 |
+
"completions/min_terminated_length": 0.0,
|
| 1193 |
+
"entropy": 1.8367336988449097,
|
| 1194 |
+
"epoch": 1.625,
|
| 1195 |
+
"frac_reward_zero_std": 1.0,
|
| 1196 |
+
"grad_norm": 0.0509357824921608,
|
| 1197 |
+
"kl": 0.11890356987714767,
|
| 1198 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 1199 |
+
"loss": 0.00029725892818532884,
|
| 1200 |
+
"num_tokens": 171180.0,
|
| 1201 |
+
"reward": 0.9323133230209351,
|
| 1202 |
+
"reward_std": 0.0,
|
| 1203 |
+
"rewards/AnswererRewardFunction/mean": 0.9323133230209351,
|
| 1204 |
+
"rewards/AnswererRewardFunction/std": 0.0,
|
| 1205 |
+
"step": 65,
|
| 1206 |
+
"step_time": 7.954669548999846
|
| 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": 1.813781499862671,
|
| 1215 |
+
"epoch": 1.65,
|
| 1216 |
+
"grad_norm": 0.041692521423101425,
|
| 1217 |
+
"kl": 0.12231048941612244,
|
| 1218 |
+
"learning_rate": 1.375e-06,
|
| 1219 |
+
"loss": 0.0003057762223761529,
|
| 1220 |
+
"step": 66,
|
| 1221 |
+
"step_time": 0.11109020200001396
|
| 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": 2.020078659057617,
|
| 1230 |
+
"epoch": 1.675,
|
| 1231 |
+
"grad_norm": 0.04829629138112068,
|
| 1232 |
+
"kl": 0.11636092513799667,
|
| 1233 |
+
"learning_rate": 1.35e-06,
|
| 1234 |
+
"loss": 0.0002909023023676127,
|
| 1235 |
+
"step": 67,
|
| 1236 |
+
"step_time": 0.11080152700014878
|
| 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": 1.8754607439041138,
|
| 1245 |
+
"epoch": 1.7,
|
| 1246 |
+
"grad_norm": 0.10382068902254105,
|
| 1247 |
+
"kl": 0.15152397751808167,
|
| 1248 |
+
"learning_rate": 1.325e-06,
|
| 1249 |
+
"loss": 0.00037880995660088956,
|
| 1250 |
+
"step": 68,
|
| 1251 |
+
"step_time": 0.11052616100005253
|
| 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": 256.0,
|
| 1261 |
+
"completions/max_terminated_length": 0.0,
|
| 1262 |
+
"completions/mean_length": 256.0,
|
| 1263 |
+
"completions/mean_terminated_length": 0.0,
|
| 1264 |
+
"completions/min_length": 256.0,
|
| 1265 |
+
"completions/min_terminated_length": 0.0,
|
| 1266 |
+
"entropy": 2.18009090423584,
|
| 1267 |
+
"epoch": 1.725,
|
| 1268 |
+
"frac_reward_zero_std": 1.0,
|
| 1269 |
+
"grad_norm": 0.04690709710121155,
|
| 1270 |
+
"kl": 0.129824697971344,
|
| 1271 |
+
"learning_rate": 1.3e-06,
|
| 1272 |
+
"loss": 0.0003245617263019085,
|
| 1273 |
+
"num_tokens": 181280.0,
|
| 1274 |
+
"reward": 0.9394216537475586,
|
| 1275 |
+
"reward_std": 0.00734142167493701,
|
| 1276 |
+
"rewards/AnswererRewardFunction/mean": 0.9394216537475586,
|
| 1277 |
+
"rewards/AnswererRewardFunction/std": 0.00734142167493701,
|
| 1278 |
+
"step": 69,
|
| 1279 |
+
"step_time": 7.907399953999629
|
| 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": 2.0286574363708496,
|
| 1288 |
+
"epoch": 1.75,
|
| 1289 |
+
"grad_norm": 0.03925037384033203,
|
| 1290 |
+
"kl": 0.11402641236782074,
|
| 1291 |
+
"learning_rate": 1.275e-06,
|
| 1292 |
+
"loss": 0.00028506602393463254,
|
| 1293 |
+
"step": 70,
|
| 1294 |
+
"step_time": 0.1109790789996623
|
| 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": 2.0765724182128906,
|
| 1303 |
+
"epoch": 1.775,
|
| 1304 |
+
"grad_norm": 0.03309269994497299,
|
| 1305 |
+
"kl": 0.11155493557453156,
|
| 1306 |
+
"learning_rate": 1.25e-06,
|
| 1307 |
+
"loss": 0.0002788873389363289,
|
| 1308 |
+
"step": 71,
|
| 1309 |
+
"step_time": 0.1113708470002166
|
| 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": 2.0393435955047607,
|
| 1318 |
+
"epoch": 1.8,
|
| 1319 |
+
"grad_norm": 0.031221741810441017,
|
| 1320 |
+
"kl": 0.11150261759757996,
|
| 1321 |
+
"learning_rate": 1.225e-06,
|
| 1322 |
+
"loss": 0.00027875651721842587,
|
| 1323 |
+
"step": 72,
|
| 1324 |
+
"step_time": 0.10620486800053186
|
| 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": 256.0,
|
| 1334 |
+
"completions/max_terminated_length": 0.0,
|
| 1335 |
+
"completions/mean_length": 256.0,
|
| 1336 |
+
"completions/mean_terminated_length": 0.0,
|
| 1337 |
+
"completions/min_length": 256.0,
|
| 1338 |
+
"completions/min_terminated_length": 0.0,
|
| 1339 |
+
"entropy": 2.1867928504943848,
|
| 1340 |
+
"epoch": 1.825,
|
| 1341 |
+
"frac_reward_zero_std": 1.0,
|
| 1342 |
+
"grad_norm": 0.04418026655912399,
|
| 1343 |
+
"kl": 0.13718846440315247,
|
| 1344 |
+
"learning_rate": 1.2000000000000002e-06,
|
| 1345 |
+
"loss": 0.00034297117963433266,
|
| 1346 |
+
"num_tokens": 191260.0,
|
| 1347 |
+
"reward": 0.7480312585830688,
|
| 1348 |
+
"reward_std": 0.11062110960483551,
|
| 1349 |
+
"rewards/AnswererRewardFunction/mean": 0.7480312585830688,
|
| 1350 |
+
"rewards/AnswererRewardFunction/std": 0.11062111705541611,
|
| 1351 |
+
"step": 73,
|
| 1352 |
+
"step_time": 7.9303118289999475
|
| 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": 2.141998052597046,
|
| 1361 |
+
"epoch": 1.85,
|
| 1362 |
+
"grad_norm": 0.03836562857031822,
|
| 1363 |
+
"kl": 0.12017437815666199,
|
| 1364 |
+
"learning_rate": 1.175e-06,
|
| 1365 |
+
"loss": 0.00030043593142181635,
|
| 1366 |
+
"step": 74,
|
| 1367 |
+
"step_time": 0.15656315299929702
|
| 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": 2.0122885704040527,
|
| 1376 |
+
"epoch": 1.875,
|
| 1377 |
+
"grad_norm": 0.07951362431049347,
|
| 1378 |
+
"kl": 0.13586632907390594,
|
| 1379 |
+
"learning_rate": 1.1500000000000002e-06,
|
| 1380 |
+
"loss": 0.0003396657994017005,
|
| 1381 |
+
"step": 75,
|
| 1382 |
+
"step_time": 0.11328490699997928
|
| 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": 2.159562826156616,
|
| 1391 |
+
"epoch": 1.9,
|
| 1392 |
+
"grad_norm": 0.04065697267651558,
|
| 1393 |
+
"kl": 0.11498960852622986,
|
| 1394 |
+
"learning_rate": 1.125e-06,
|
| 1395 |
+
"loss": 0.0002874740166589618,
|
| 1396 |
+
"step": 76,
|
| 1397 |
+
"step_time": 0.11609526600022946
|
| 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": 256.0,
|
| 1407 |
+
"completions/max_terminated_length": 0.0,
|
| 1408 |
+
"completions/mean_length": 256.0,
|
| 1409 |
+
"completions/mean_terminated_length": 0.0,
|
| 1410 |
+
"completions/min_length": 256.0,
|
| 1411 |
+
"completions/min_terminated_length": 0.0,
|
| 1412 |
+
"entropy": 2.0413670539855957,
|
| 1413 |
+
"epoch": 1.925,
|
| 1414 |
+
"frac_reward_zero_std": 1.0,
|
| 1415 |
+
"grad_norm": 0.0457465685904026,
|
| 1416 |
+
"kl": 0.12245075404644012,
|
| 1417 |
+
"learning_rate": 1.0999999999999998e-06,
|
| 1418 |
+
"loss": 0.00030612689442932606,
|
| 1419 |
+
"num_tokens": 201392.0,
|
| 1420 |
+
"reward": 0.8702797293663025,
|
| 1421 |
+
"reward_std": 0.11096905916929245,
|
| 1422 |
+
"rewards/AnswererRewardFunction/mean": 0.8702797293663025,
|
| 1423 |
+
"rewards/AnswererRewardFunction/std": 0.11096906661987305,
|
| 1424 |
+
"step": 77,
|
| 1425 |
+
"step_time": 8.033617556999161
|
| 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": 2.123049736022949,
|
| 1434 |
+
"epoch": 1.95,
|
| 1435 |
+
"grad_norm": 0.033493202179670334,
|
| 1436 |
+
"kl": 0.11246775090694427,
|
| 1437 |
+
"learning_rate": 1.075e-06,
|
| 1438 |
+
"loss": 0.0002811693702824414,
|
| 1439 |
+
"step": 78,
|
| 1440 |
+
"step_time": 0.11165324300054635
|
| 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": 1.990412950515747,
|
| 1449 |
+
"epoch": 1.975,
|
| 1450 |
+
"grad_norm": 0.048628926277160645,
|
| 1451 |
+
"kl": 0.13643597066402435,
|
| 1452 |
+
"learning_rate": 1.05e-06,
|
| 1453 |
+
"loss": 0.0003410899080336094,
|
| 1454 |
+
"step": 79,
|
| 1455 |
+
"step_time": 0.11406734499996674
|
| 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": 2.0939106941223145,
|
| 1464 |
+
"epoch": 2.0,
|
| 1465 |
+
"grad_norm": 0.047254763543605804,
|
| 1466 |
+
"kl": 0.12829910218715668,
|
| 1467 |
+
"learning_rate": 1.0250000000000001e-06,
|
| 1468 |
+
"loss": 0.00032074772752821445,
|
| 1469 |
+
"step": 80,
|
| 1470 |
+
"step_time": 0.11228936599945882
|
| 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": 256.0,
|
| 1480 |
+
"completions/max_terminated_length": 0.0,
|
| 1481 |
+
"completions/mean_length": 256.0,
|
| 1482 |
+
"completions/mean_terminated_length": 0.0,
|
| 1483 |
+
"completions/min_length": 256.0,
|
| 1484 |
+
"completions/min_terminated_length": 0.0,
|
| 1485 |
+
"entropy": 2.1104767322540283,
|
| 1486 |
+
"epoch": 2.025,
|
| 1487 |
+
"frac_reward_zero_std": 1.0,
|
| 1488 |
+
"grad_norm": 0.030738642439246178,
|
| 1489 |
+
"kl": 0.10635365545749664,
|
| 1490 |
+
"learning_rate": 1e-06,
|
| 1491 |
+
"loss": 0.0002658841258380562,
|
| 1492 |
+
"num_tokens": 211528.0,
|
| 1493 |
+
"reward": 0.8755199909210205,
|
| 1494 |
+
"reward_std": 0.10159505903720856,
|
| 1495 |
+
"rewards/AnswererRewardFunction/mean": 0.8755199909210205,
|
| 1496 |
+
"rewards/AnswererRewardFunction/std": 0.10159505903720856,
|
| 1497 |
+
"step": 81,
|
| 1498 |
+
"step_time": 7.942807592999998
|
| 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": 2.336609125137329,
|
| 1507 |
+
"epoch": 2.05,
|
| 1508 |
+
"grad_norm": 0.037570513784885406,
|
| 1509 |
+
"kl": 0.1238531619310379,
|
| 1510 |
+
"learning_rate": 9.75e-07,
|
| 1511 |
+
"loss": 0.00030963291646912694,
|
| 1512 |
+
"step": 82,
|
| 1513 |
+
"step_time": 0.11186360600004264
|
| 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": 2.1161727905273438,
|
| 1522 |
+
"epoch": 2.075,
|
| 1523 |
+
"grad_norm": 0.036076463758945465,
|
| 1524 |
+
"kl": 0.12031081318855286,
|
| 1525 |
+
"learning_rate": 9.5e-07,
|
| 1526 |
+
"loss": 0.00030077702831476927,
|
| 1527 |
+
"step": 83,
|
| 1528 |
+
"step_time": 0.11085510600059933
|
| 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": 2.109645128250122,
|
| 1537 |
+
"epoch": 2.1,
|
| 1538 |
+
"grad_norm": 0.028782352805137634,
|
| 1539 |
+
"kl": 0.09679339081048965,
|
| 1540 |
+
"learning_rate": 9.25e-07,
|
| 1541 |
+
"loss": 0.00024198347819037735,
|
| 1542 |
+
"step": 84,
|
| 1543 |
+
"step_time": 0.11177983600009611
|
| 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": 256.0,
|
| 1553 |
+
"completions/max_terminated_length": 0.0,
|
| 1554 |
+
"completions/mean_length": 256.0,
|
| 1555 |
+
"completions/mean_terminated_length": 0.0,
|
| 1556 |
+
"completions/min_length": 256.0,
|
| 1557 |
+
"completions/min_terminated_length": 0.0,
|
| 1558 |
+
"entropy": 1.9927191734313965,
|
| 1559 |
+
"epoch": 2.125,
|
| 1560 |
+
"frac_reward_zero_std": 1.0,
|
| 1561 |
+
"grad_norm": 0.036913640797138214,
|
| 1562 |
+
"kl": 0.10849414765834808,
|
| 1563 |
+
"learning_rate": 9e-07,
|
| 1564 |
+
"loss": 0.00027123535983264446,
|
| 1565 |
+
"num_tokens": 221264.0,
|
| 1566 |
+
"reward": 0.8402513265609741,
|
| 1567 |
+
"reward_std": 0.13532696664333344,
|
| 1568 |
+
"rewards/AnswererRewardFunction/mean": 0.8402513265609741,
|
| 1569 |
+
"rewards/AnswererRewardFunction/std": 0.13532698154449463,
|
| 1570 |
+
"step": 85,
|
| 1571 |
+
"step_time": 7.973877727000399
|
| 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": 2.0149073600769043,
|
| 1580 |
+
"epoch": 2.15,
|
| 1581 |
+
"grad_norm": 0.03311130777001381,
|
| 1582 |
+
"kl": 0.1119215339422226,
|
| 1583 |
+
"learning_rate": 8.750000000000001e-07,
|
| 1584 |
+
"loss": 0.0002798038476612419,
|
| 1585 |
+
"step": 86,
|
| 1586 |
+
"step_time": 0.10729382600038662
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"clip_ratio/high_max": 0.0,
|
| 1590 |
+
"clip_ratio/high_mean": 0.0,
|
| 1591 |
+
"clip_ratio/low_mean": 0.0,
|
| 1592 |
+
"clip_ratio/low_min": 0.0,
|
| 1593 |
+
"clip_ratio/region_mean": 0.0,
|
| 1594 |
+
"entropy": 1.7913100719451904,
|
| 1595 |
+
"epoch": 2.175,
|
| 1596 |
+
"grad_norm": 0.024518176913261414,
|
| 1597 |
+
"kl": 0.08977322280406952,
|
| 1598 |
+
"learning_rate": 8.5e-07,
|
| 1599 |
+
"loss": 0.00022443305351771414,
|
| 1600 |
+
"step": 87,
|
| 1601 |
+
"step_time": 0.10724336499970377
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"clip_ratio/high_max": 0.0,
|
| 1605 |
+
"clip_ratio/high_mean": 0.0,
|
| 1606 |
+
"clip_ratio/low_mean": 0.0,
|
| 1607 |
+
"clip_ratio/low_min": 0.0,
|
| 1608 |
+
"clip_ratio/region_mean": 0.0,
|
| 1609 |
+
"entropy": 2.058328151702881,
|
| 1610 |
+
"epoch": 2.2,
|
| 1611 |
+
"grad_norm": 0.031460244208574295,
|
| 1612 |
+
"kl": 0.10560563206672668,
|
| 1613 |
+
"learning_rate": 8.25e-07,
|
| 1614 |
+
"loss": 0.0002640140592120588,
|
| 1615 |
+
"step": 88,
|
| 1616 |
+
"step_time": 0.10991910000029748
|
| 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": 256.0,
|
| 1626 |
+
"completions/max_terminated_length": 0.0,
|
| 1627 |
+
"completions/mean_length": 256.0,
|
| 1628 |
+
"completions/mean_terminated_length": 0.0,
|
| 1629 |
+
"completions/min_length": 256.0,
|
| 1630 |
+
"completions/min_terminated_length": 0.0,
|
| 1631 |
+
"entropy": 1.8188772201538086,
|
| 1632 |
+
"epoch": 2.225,
|
| 1633 |
+
"frac_reward_zero_std": 1.0,
|
| 1634 |
+
"grad_norm": 0.033341169357299805,
|
| 1635 |
+
"kl": 0.10262635350227356,
|
| 1636 |
+
"learning_rate": 8e-07,
|
| 1637 |
+
"loss": 0.0002565658651292324,
|
| 1638 |
+
"num_tokens": 231000.0,
|
| 1639 |
+
"reward": 0.780498743057251,
|
| 1640 |
+
"reward_std": 0.09186852723360062,
|
| 1641 |
+
"rewards/AnswererRewardFunction/mean": 0.780498743057251,
|
| 1642 |
+
"rewards/AnswererRewardFunction/std": 0.09186854958534241,
|
| 1643 |
+
"step": 89,
|
| 1644 |
+
"step_time": 7.958869117999711
|
| 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": 2.034656047821045,
|
| 1653 |
+
"epoch": 2.25,
|
| 1654 |
+
"grad_norm": 0.05250053480267525,
|
| 1655 |
+
"kl": 0.1321343332529068,
|
| 1656 |
+
"learning_rate": 7.750000000000001e-07,
|
| 1657 |
+
"loss": 0.00033033580984920263,
|
| 1658 |
+
"step": 90,
|
| 1659 |
+
"step_time": 0.11438874499981466
|
| 1660 |
+
}
|
| 1661 |
+
],
|
| 1662 |
+
"logging_steps": 1,
|
| 1663 |
+
"max_steps": 120,
|
| 1664 |
+
"num_input_tokens_seen": 231000,
|
| 1665 |
+
"num_train_epochs": 3,
|
| 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_005/answerer_train/checkpoint-90/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:176482d0784a3473c5f8dbadcd601422b0dadd5afadeeee68d055f1eaaf7d577
|
| 3 |
+
size 7249
|