first push
Browse files- .gitattributes +1 -0
- added_tokens.json +28 -0
- chat_template.jinja +85 -0
- config.json +30 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +31 -0
- tokenizer.json +3 -0
- tokenizer_config.json +239 -0
- trainer_state.json +2642 -0
- training_args.bin +3 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
added_tokens.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</think>": 151668,
|
| 3 |
+
"</tool_call>": 151658,
|
| 4 |
+
"</tool_response>": 151666,
|
| 5 |
+
"<think>": 151667,
|
| 6 |
+
"<tool_call>": 151657,
|
| 7 |
+
"<tool_response>": 151665,
|
| 8 |
+
"<|box_end|>": 151649,
|
| 9 |
+
"<|box_start|>": 151648,
|
| 10 |
+
"<|endoftext|>": 151643,
|
| 11 |
+
"<|file_sep|>": 151664,
|
| 12 |
+
"<|fim_middle|>": 151660,
|
| 13 |
+
"<|fim_pad|>": 151662,
|
| 14 |
+
"<|fim_prefix|>": 151659,
|
| 15 |
+
"<|fim_suffix|>": 151661,
|
| 16 |
+
"<|im_end|>": 151645,
|
| 17 |
+
"<|im_start|>": 151644,
|
| 18 |
+
"<|image_pad|>": 151655,
|
| 19 |
+
"<|object_ref_end|>": 151647,
|
| 20 |
+
"<|object_ref_start|>": 151646,
|
| 21 |
+
"<|quad_end|>": 151651,
|
| 22 |
+
"<|quad_start|>": 151650,
|
| 23 |
+
"<|repo_name|>": 151663,
|
| 24 |
+
"<|video_pad|>": 151656,
|
| 25 |
+
"<|vision_end|>": 151653,
|
| 26 |
+
"<|vision_pad|>": 151654,
|
| 27 |
+
"<|vision_start|>": 151652
|
| 28 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,85 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 18 |
+
{%- for message in messages[::-1] %}
|
| 19 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 20 |
+
{%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
|
| 21 |
+
{%- set ns.multi_step_tool = false %}
|
| 22 |
+
{%- set ns.last_query_index = index %}
|
| 23 |
+
{%- endif %}
|
| 24 |
+
{%- endfor %}
|
| 25 |
+
{%- for message in messages %}
|
| 26 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 27 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 28 |
+
{%- elif message.role == "assistant" %}
|
| 29 |
+
{%- set content = message.content %}
|
| 30 |
+
{%- set reasoning_content = '' %}
|
| 31 |
+
{%- if message.reasoning_content is defined and message.reasoning_content is not none %}
|
| 32 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 33 |
+
{%- else %}
|
| 34 |
+
{%- if '</think>' in message.content %}
|
| 35 |
+
{%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
|
| 36 |
+
{%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 37 |
+
{%- endif %}
|
| 38 |
+
{%- endif %}
|
| 39 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 40 |
+
{%- if loop.last or (not loop.last and reasoning_content) %}
|
| 41 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
|
| 42 |
+
{%- else %}
|
| 43 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- else %}
|
| 46 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 47 |
+
{%- endif %}
|
| 48 |
+
{%- if message.tool_calls %}
|
| 49 |
+
{%- for tool_call in message.tool_calls %}
|
| 50 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 51 |
+
{{- '\n' }}
|
| 52 |
+
{%- endif %}
|
| 53 |
+
{%- if tool_call.function %}
|
| 54 |
+
{%- set tool_call = tool_call.function %}
|
| 55 |
+
{%- endif %}
|
| 56 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 57 |
+
{{- tool_call.name }}
|
| 58 |
+
{{- '", "arguments": ' }}
|
| 59 |
+
{%- if tool_call.arguments is string %}
|
| 60 |
+
{{- tool_call.arguments }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{{- tool_call.arguments | tojson }}
|
| 63 |
+
{%- endif %}
|
| 64 |
+
{{- '}\n</tool_call>' }}
|
| 65 |
+
{%- endfor %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{{- '<|im_end|>\n' }}
|
| 68 |
+
{%- elif message.role == "tool" %}
|
| 69 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 70 |
+
{{- '<|im_start|>user' }}
|
| 71 |
+
{%- endif %}
|
| 72 |
+
{{- '\n<tool_response>\n' }}
|
| 73 |
+
{{- message.content }}
|
| 74 |
+
{{- '\n</tool_response>' }}
|
| 75 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 76 |
+
{{- '<|im_end|>\n' }}
|
| 77 |
+
{%- endif %}
|
| 78 |
+
{%- endif %}
|
| 79 |
+
{%- endfor %}
|
| 80 |
+
{%- if add_generation_prompt %}
|
| 81 |
+
{{- '<|im_start|>assistant\n' }}
|
| 82 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 83 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 84 |
+
{%- endif %}
|
| 85 |
+
{%- endif %}
|
config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 151643,
|
| 8 |
+
"eos_token_id": 151643,
|
| 9 |
+
"head_dim": 128,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"max_position_embeddings": 32768,
|
| 15 |
+
"max_window_layers": 28,
|
| 16 |
+
"model_type": "qwen3",
|
| 17 |
+
"num_attention_heads": 16,
|
| 18 |
+
"num_hidden_layers": 28,
|
| 19 |
+
"num_key_value_heads": 8,
|
| 20 |
+
"rms_norm_eps": 1e-06,
|
| 21 |
+
"rope_scaling": null,
|
| 22 |
+
"rope_theta": 1000000,
|
| 23 |
+
"sliding_window": null,
|
| 24 |
+
"tie_word_embeddings": true,
|
| 25 |
+
"torch_dtype": "float32",
|
| 26 |
+
"transformers_version": "4.52.2",
|
| 27 |
+
"use_cache": false,
|
| 28 |
+
"use_sliding_window": false,
|
| 29 |
+
"vocab_size": 151936
|
| 30 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151643,
|
| 4 |
+
"max_new_tokens": 2048,
|
| 5 |
+
"transformers_version": "4.52.2"
|
| 6 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6e0424ac7c869526b8b9cd0ebccf227436f0112201b2dbb224d0e0be6b17d0f
|
| 3 |
+
size 2384234968
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65a3cafd5fa5184bf559d3fd1ed533167cf4441dbb05d590896396eb47c5dc9f
|
| 3 |
+
size 4768662910
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f32c25f8ac5f19710a6ed752bf6968914f1573258beefeff5964234566a45c0a
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b82b000dfaca85efd02b61694f290dd72f3fe9bd03590b8a9d86848be2428b5
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|im_start|>",
|
| 4 |
+
"<|im_end|>",
|
| 5 |
+
"<|object_ref_start|>",
|
| 6 |
+
"<|object_ref_end|>",
|
| 7 |
+
"<|box_start|>",
|
| 8 |
+
"<|box_end|>",
|
| 9 |
+
"<|quad_start|>",
|
| 10 |
+
"<|quad_end|>",
|
| 11 |
+
"<|vision_start|>",
|
| 12 |
+
"<|vision_end|>",
|
| 13 |
+
"<|vision_pad|>",
|
| 14 |
+
"<|image_pad|>",
|
| 15 |
+
"<|video_pad|>"
|
| 16 |
+
],
|
| 17 |
+
"eos_token": {
|
| 18 |
+
"content": "<|endoftext|>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
},
|
| 24 |
+
"pad_token": {
|
| 25 |
+
"content": "<|endoftext|>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": false,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
}
|
| 31 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
|
| 3 |
+
size 11422654
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,239 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
}
|
| 213 |
+
},
|
| 214 |
+
"additional_special_tokens": [
|
| 215 |
+
"<|im_start|>",
|
| 216 |
+
"<|im_end|>",
|
| 217 |
+
"<|object_ref_start|>",
|
| 218 |
+
"<|object_ref_end|>",
|
| 219 |
+
"<|box_start|>",
|
| 220 |
+
"<|box_end|>",
|
| 221 |
+
"<|quad_start|>",
|
| 222 |
+
"<|quad_end|>",
|
| 223 |
+
"<|vision_start|>",
|
| 224 |
+
"<|vision_end|>",
|
| 225 |
+
"<|vision_pad|>",
|
| 226 |
+
"<|image_pad|>",
|
| 227 |
+
"<|video_pad|>"
|
| 228 |
+
],
|
| 229 |
+
"bos_token": null,
|
| 230 |
+
"clean_up_tokenization_spaces": false,
|
| 231 |
+
"eos_token": "<|endoftext|>",
|
| 232 |
+
"errors": "replace",
|
| 233 |
+
"extra_special_tokens": {},
|
| 234 |
+
"model_max_length": 131072,
|
| 235 |
+
"pad_token": "<|endoftext|>",
|
| 236 |
+
"split_special_tokens": false,
|
| 237 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 238 |
+
"unk_token": null
|
| 239 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2642 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 1449,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.006904289289721239,
|
| 14 |
+
"grad_norm": 117.87319946289062,
|
| 15 |
+
"learning_rate": 4.968944099378881e-07,
|
| 16 |
+
"logits/chosen": -1.6597025394439697,
|
| 17 |
+
"logits/rejected": -1.4006378650665283,
|
| 18 |
+
"logps/chosen": -401.00640869140625,
|
| 19 |
+
"logps/rejected": -376.5096130371094,
|
| 20 |
+
"loss": 0.698,
|
| 21 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 22 |
+
"rewards/chosen": -0.004771101754158735,
|
| 23 |
+
"rewards/margins": -0.0065481411293148994,
|
| 24 |
+
"rewards/rejected": 0.001777039491571486,
|
| 25 |
+
"step": 10
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.013808578579442479,
|
| 29 |
+
"grad_norm": 90.4378433227539,
|
| 30 |
+
"learning_rate": 4.934437543133195e-07,
|
| 31 |
+
"logits/chosen": NaN,
|
| 32 |
+
"logits/rejected": NaN,
|
| 33 |
+
"logps/chosen": -391.04168701171875,
|
| 34 |
+
"logps/rejected": -373.75286865234375,
|
| 35 |
+
"loss": 0.6922,
|
| 36 |
+
"rewards/accuracies": 0.518750011920929,
|
| 37 |
+
"rewards/chosen": 0.012870313599705696,
|
| 38 |
+
"rewards/margins": 0.003891288535669446,
|
| 39 |
+
"rewards/rejected": 0.00897902436554432,
|
| 40 |
+
"step": 20
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.02071286786916372,
|
| 44 |
+
"grad_norm": 107.49934387207031,
|
| 45 |
+
"learning_rate": 4.899930986887508e-07,
|
| 46 |
+
"logits/chosen": -1.2349631786346436,
|
| 47 |
+
"logits/rejected": -1.1878811120986938,
|
| 48 |
+
"logps/chosen": -379.464599609375,
|
| 49 |
+
"logps/rejected": -380.815185546875,
|
| 50 |
+
"loss": 0.6913,
|
| 51 |
+
"rewards/accuracies": 0.518750011920929,
|
| 52 |
+
"rewards/chosen": 0.016194406896829605,
|
| 53 |
+
"rewards/margins": 0.0065515548922121525,
|
| 54 |
+
"rewards/rejected": 0.00964285247027874,
|
| 55 |
+
"step": 30
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.027617157158884957,
|
| 59 |
+
"grad_norm": 103.74991607666016,
|
| 60 |
+
"learning_rate": 4.865424430641822e-07,
|
| 61 |
+
"logits/chosen": -1.433064579963684,
|
| 62 |
+
"logits/rejected": -1.22975492477417,
|
| 63 |
+
"logps/chosen": -411.400634765625,
|
| 64 |
+
"logps/rejected": -369.98309326171875,
|
| 65 |
+
"loss": 0.6929,
|
| 66 |
+
"rewards/accuracies": 0.512499988079071,
|
| 67 |
+
"rewards/chosen": 0.012084421701729298,
|
| 68 |
+
"rewards/margins": 0.003508642315864563,
|
| 69 |
+
"rewards/rejected": 0.008575777523219585,
|
| 70 |
+
"step": 40
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.0345214464486062,
|
| 74 |
+
"grad_norm": 122.35696411132812,
|
| 75 |
+
"learning_rate": 4.830917874396135e-07,
|
| 76 |
+
"logits/chosen": -1.6217113733291626,
|
| 77 |
+
"logits/rejected": -1.3987983465194702,
|
| 78 |
+
"logps/chosen": -398.4120178222656,
|
| 79 |
+
"logps/rejected": -421.19915771484375,
|
| 80 |
+
"loss": 0.6945,
|
| 81 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 82 |
+
"rewards/chosen": 0.026384565979242325,
|
| 83 |
+
"rewards/margins": 0.0009822694119066,
|
| 84 |
+
"rewards/rejected": 0.02540229819715023,
|
| 85 |
+
"step": 50
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.0345214464486062,
|
| 89 |
+
"eval_logits/chosen": NaN,
|
| 90 |
+
"eval_logits/rejected": -1.1866018772125244,
|
| 91 |
+
"eval_logps/chosen": -413.12347412109375,
|
| 92 |
+
"eval_logps/rejected": -386.6858825683594,
|
| 93 |
+
"eval_loss": 0.6903728246688843,
|
| 94 |
+
"eval_rewards/accuracies": 0.534844696521759,
|
| 95 |
+
"eval_rewards/chosen": 0.034061625599861145,
|
| 96 |
+
"eval_rewards/margins": 0.008847548626363277,
|
| 97 |
+
"eval_rewards/rejected": 0.025214076042175293,
|
| 98 |
+
"eval_runtime": 174.7974,
|
| 99 |
+
"eval_samples_per_second": 6.814,
|
| 100 |
+
"eval_steps_per_second": 6.814,
|
| 101 |
+
"step": 50
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.04142573573832744,
|
| 105 |
+
"grad_norm": 93.54374694824219,
|
| 106 |
+
"learning_rate": 4.796411318150448e-07,
|
| 107 |
+
"logits/chosen": -1.348595142364502,
|
| 108 |
+
"logits/rejected": -1.2194318771362305,
|
| 109 |
+
"logps/chosen": -402.88262939453125,
|
| 110 |
+
"logps/rejected": -371.10443115234375,
|
| 111 |
+
"loss": 0.6856,
|
| 112 |
+
"rewards/accuracies": 0.574999988079071,
|
| 113 |
+
"rewards/chosen": 0.04310911148786545,
|
| 114 |
+
"rewards/margins": 0.018718212842941284,
|
| 115 |
+
"rewards/rejected": 0.024390896782279015,
|
| 116 |
+
"step": 60
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04833002502804867,
|
| 120 |
+
"grad_norm": 107.30802917480469,
|
| 121 |
+
"learning_rate": 4.761904761904761e-07,
|
| 122 |
+
"logits/chosen": -1.6407556533813477,
|
| 123 |
+
"logits/rejected": -1.3553545475006104,
|
| 124 |
+
"logps/chosen": -373.2159729003906,
|
| 125 |
+
"logps/rejected": -349.29217529296875,
|
| 126 |
+
"loss": 0.6962,
|
| 127 |
+
"rewards/accuracies": 0.512499988079071,
|
| 128 |
+
"rewards/chosen": 0.04530588164925575,
|
| 129 |
+
"rewards/margins": -0.002548071090131998,
|
| 130 |
+
"rewards/rejected": 0.04785395413637161,
|
| 131 |
+
"step": 70
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.055234314317769914,
|
| 135 |
+
"grad_norm": 107.91950988769531,
|
| 136 |
+
"learning_rate": 4.727398205659075e-07,
|
| 137 |
+
"logits/chosen": -1.5417871475219727,
|
| 138 |
+
"logits/rejected": -1.3636682033538818,
|
| 139 |
+
"logps/chosen": -411.05743408203125,
|
| 140 |
+
"logps/rejected": -385.8003845214844,
|
| 141 |
+
"loss": 0.6882,
|
| 142 |
+
"rewards/accuracies": 0.5625,
|
| 143 |
+
"rewards/chosen": 0.08122040331363678,
|
| 144 |
+
"rewards/margins": 0.014340770430862904,
|
| 145 |
+
"rewards/rejected": 0.06687963753938675,
|
| 146 |
+
"step": 80
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.062138603607491155,
|
| 150 |
+
"grad_norm": 91.66450500488281,
|
| 151 |
+
"learning_rate": 4.6928916494133887e-07,
|
| 152 |
+
"logits/chosen": -1.541373610496521,
|
| 153 |
+
"logits/rejected": -1.3865256309509277,
|
| 154 |
+
"logps/chosen": -416.6394958496094,
|
| 155 |
+
"logps/rejected": -386.05023193359375,
|
| 156 |
+
"loss": 0.6845,
|
| 157 |
+
"rewards/accuracies": 0.5625,
|
| 158 |
+
"rewards/chosen": 0.10198626667261124,
|
| 159 |
+
"rewards/margins": 0.02344668284058571,
|
| 160 |
+
"rewards/rejected": 0.07853958010673523,
|
| 161 |
+
"step": 90
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.0690428928972124,
|
| 165 |
+
"grad_norm": 84.30016326904297,
|
| 166 |
+
"learning_rate": 4.6583850931677014e-07,
|
| 167 |
+
"logits/chosen": -1.452531337738037,
|
| 168 |
+
"logits/rejected": -1.2693792581558228,
|
| 169 |
+
"logps/chosen": -432.15301513671875,
|
| 170 |
+
"logps/rejected": -382.9742126464844,
|
| 171 |
+
"loss": 0.6756,
|
| 172 |
+
"rewards/accuracies": 0.5625,
|
| 173 |
+
"rewards/chosen": 0.12693843245506287,
|
| 174 |
+
"rewards/margins": 0.04367733746767044,
|
| 175 |
+
"rewards/rejected": 0.08326110243797302,
|
| 176 |
+
"step": 100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.0690428928972124,
|
| 180 |
+
"eval_logits/chosen": NaN,
|
| 181 |
+
"eval_logits/rejected": -1.1412357091903687,
|
| 182 |
+
"eval_logps/chosen": -412.13031005859375,
|
| 183 |
+
"eval_logps/rejected": -385.9539794921875,
|
| 184 |
+
"eval_loss": 0.6796301603317261,
|
| 185 |
+
"eval_rewards/accuracies": 0.5751469135284424,
|
| 186 |
+
"eval_rewards/chosen": 0.13337667286396027,
|
| 187 |
+
"eval_rewards/margins": 0.034973062574863434,
|
| 188 |
+
"eval_rewards/rejected": 0.09840361773967743,
|
| 189 |
+
"eval_runtime": 174.9705,
|
| 190 |
+
"eval_samples_per_second": 6.807,
|
| 191 |
+
"eval_steps_per_second": 6.807,
|
| 192 |
+
"step": 100
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.07594718218693364,
|
| 196 |
+
"grad_norm": 106.82269287109375,
|
| 197 |
+
"learning_rate": 4.623878536922015e-07,
|
| 198 |
+
"logits/chosen": -1.4040501117706299,
|
| 199 |
+
"logits/rejected": -1.3599127531051636,
|
| 200 |
+
"logps/chosen": -406.8678283691406,
|
| 201 |
+
"logps/rejected": -398.7535095214844,
|
| 202 |
+
"loss": 0.6888,
|
| 203 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 204 |
+
"rewards/chosen": 0.14022807776927948,
|
| 205 |
+
"rewards/margins": 0.01770814135670662,
|
| 206 |
+
"rewards/rejected": 0.12251994758844376,
|
| 207 |
+
"step": 110
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.08285147147665488,
|
| 211 |
+
"grad_norm": 115.4521713256836,
|
| 212 |
+
"learning_rate": 4.5893719806763283e-07,
|
| 213 |
+
"logits/chosen": -1.555271863937378,
|
| 214 |
+
"logits/rejected": -1.3029377460479736,
|
| 215 |
+
"logps/chosen": -389.0620422363281,
|
| 216 |
+
"logps/rejected": -369.9129943847656,
|
| 217 |
+
"loss": 0.685,
|
| 218 |
+
"rewards/accuracies": 0.53125,
|
| 219 |
+
"rewards/chosen": 0.14878275990486145,
|
| 220 |
+
"rewards/margins": 0.026980455964803696,
|
| 221 |
+
"rewards/rejected": 0.12180230766534805,
|
| 222 |
+
"step": 120
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.0897557607663761,
|
| 226 |
+
"grad_norm": 117.05790710449219,
|
| 227 |
+
"learning_rate": 4.5548654244306415e-07,
|
| 228 |
+
"logits/chosen": -1.5591048002243042,
|
| 229 |
+
"logits/rejected": -1.329421877861023,
|
| 230 |
+
"logps/chosen": -406.03240966796875,
|
| 231 |
+
"logps/rejected": -395.0228271484375,
|
| 232 |
+
"loss": 0.683,
|
| 233 |
+
"rewards/accuracies": 0.543749988079071,
|
| 234 |
+
"rewards/chosen": 0.17079384624958038,
|
| 235 |
+
"rewards/margins": 0.033137861639261246,
|
| 236 |
+
"rewards/rejected": 0.13765597343444824,
|
| 237 |
+
"step": 130
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.09666005005609735,
|
| 241 |
+
"grad_norm": 95.98699951171875,
|
| 242 |
+
"learning_rate": 4.520358868184955e-07,
|
| 243 |
+
"logits/chosen": -1.3720393180847168,
|
| 244 |
+
"logits/rejected": -1.1874357461929321,
|
| 245 |
+
"logps/chosen": -386.15765380859375,
|
| 246 |
+
"logps/rejected": -347.72686767578125,
|
| 247 |
+
"loss": 0.6762,
|
| 248 |
+
"rewards/accuracies": 0.606249988079071,
|
| 249 |
+
"rewards/chosen": 0.14972324669361115,
|
| 250 |
+
"rewards/margins": 0.04324805364012718,
|
| 251 |
+
"rewards/rejected": 0.10647518932819366,
|
| 252 |
+
"step": 140
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 0.10356433934581859,
|
| 256 |
+
"grad_norm": 93.65694427490234,
|
| 257 |
+
"learning_rate": 4.4858523119392685e-07,
|
| 258 |
+
"logits/chosen": -1.3812742233276367,
|
| 259 |
+
"logits/rejected": -1.2242071628570557,
|
| 260 |
+
"logps/chosen": -407.5322570800781,
|
| 261 |
+
"logps/rejected": -380.6641540527344,
|
| 262 |
+
"loss": 0.6806,
|
| 263 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 264 |
+
"rewards/chosen": 0.1826440393924713,
|
| 265 |
+
"rewards/margins": 0.03333988040685654,
|
| 266 |
+
"rewards/rejected": 0.14930415153503418,
|
| 267 |
+
"step": 150
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.10356433934581859,
|
| 271 |
+
"eval_logits/chosen": NaN,
|
| 272 |
+
"eval_logits/rejected": -1.134230613708496,
|
| 273 |
+
"eval_logps/chosen": -411.6000671386719,
|
| 274 |
+
"eval_logps/rejected": -385.556640625,
|
| 275 |
+
"eval_loss": 0.6752930283546448,
|
| 276 |
+
"eval_rewards/accuracies": 0.5894206762313843,
|
| 277 |
+
"eval_rewards/chosen": 0.1863998919725418,
|
| 278 |
+
"eval_rewards/margins": 0.04826309531927109,
|
| 279 |
+
"eval_rewards/rejected": 0.13813677430152893,
|
| 280 |
+
"eval_runtime": 175.5391,
|
| 281 |
+
"eval_samples_per_second": 6.785,
|
| 282 |
+
"eval_steps_per_second": 6.785,
|
| 283 |
+
"step": 150
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.11046862863553983,
|
| 287 |
+
"grad_norm": 88.74897766113281,
|
| 288 |
+
"learning_rate": 4.4513457556935817e-07,
|
| 289 |
+
"logits/chosen": -1.3776233196258545,
|
| 290 |
+
"logits/rejected": -1.0611753463745117,
|
| 291 |
+
"logps/chosen": -353.5108337402344,
|
| 292 |
+
"logps/rejected": -313.82452392578125,
|
| 293 |
+
"loss": 0.6647,
|
| 294 |
+
"rewards/accuracies": 0.637499988079071,
|
| 295 |
+
"rewards/chosen": 0.16904719173908234,
|
| 296 |
+
"rewards/margins": 0.07107650488615036,
|
| 297 |
+
"rewards/rejected": 0.09797067940235138,
|
| 298 |
+
"step": 160
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.11737291792526107,
|
| 302 |
+
"grad_norm": 88.10498809814453,
|
| 303 |
+
"learning_rate": 4.416839199447895e-07,
|
| 304 |
+
"logits/chosen": -1.5680519342422485,
|
| 305 |
+
"logits/rejected": -1.441646933555603,
|
| 306 |
+
"logps/chosen": -388.49346923828125,
|
| 307 |
+
"logps/rejected": -345.51287841796875,
|
| 308 |
+
"loss": 0.6763,
|
| 309 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 310 |
+
"rewards/chosen": 0.2311728447675705,
|
| 311 |
+
"rewards/margins": 0.05381856486201286,
|
| 312 |
+
"rewards/rejected": 0.17735427618026733,
|
| 313 |
+
"step": 170
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.12427720721498231,
|
| 317 |
+
"grad_norm": 116.68927764892578,
|
| 318 |
+
"learning_rate": 4.3823326432022087e-07,
|
| 319 |
+
"logits/chosen": -1.4206693172454834,
|
| 320 |
+
"logits/rejected": -1.175628900527954,
|
| 321 |
+
"logps/chosen": -421.7391052246094,
|
| 322 |
+
"logps/rejected": -380.89300537109375,
|
| 323 |
+
"loss": 0.6777,
|
| 324 |
+
"rewards/accuracies": 0.581250011920929,
|
| 325 |
+
"rewards/chosen": 0.25413596630096436,
|
| 326 |
+
"rewards/margins": 0.04794498533010483,
|
| 327 |
+
"rewards/rejected": 0.20619097352027893,
|
| 328 |
+
"step": 180
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 0.13118149650470354,
|
| 332 |
+
"grad_norm": 100.38077545166016,
|
| 333 |
+
"learning_rate": 4.3478260869565214e-07,
|
| 334 |
+
"logits/chosen": -1.4726192951202393,
|
| 335 |
+
"logits/rejected": -1.368575096130371,
|
| 336 |
+
"logps/chosen": -413.97216796875,
|
| 337 |
+
"logps/rejected": -382.13116455078125,
|
| 338 |
+
"loss": 0.6751,
|
| 339 |
+
"rewards/accuracies": 0.5625,
|
| 340 |
+
"rewards/chosen": 0.2780989110469818,
|
| 341 |
+
"rewards/margins": 0.06077272817492485,
|
| 342 |
+
"rewards/rejected": 0.21732616424560547,
|
| 343 |
+
"step": 190
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 0.1380857857944248,
|
| 347 |
+
"grad_norm": 102.20967102050781,
|
| 348 |
+
"learning_rate": 4.313319530710835e-07,
|
| 349 |
+
"logits/chosen": -1.1611400842666626,
|
| 350 |
+
"logits/rejected": -0.9503771662712097,
|
| 351 |
+
"logps/chosen": -409.99951171875,
|
| 352 |
+
"logps/rejected": -378.935546875,
|
| 353 |
+
"loss": 0.6825,
|
| 354 |
+
"rewards/accuracies": 0.518750011920929,
|
| 355 |
+
"rewards/chosen": 0.25466907024383545,
|
| 356 |
+
"rewards/margins": 0.03956790268421173,
|
| 357 |
+
"rewards/rejected": 0.21510115265846252,
|
| 358 |
+
"step": 200
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 0.1380857857944248,
|
| 362 |
+
"eval_logits/chosen": NaN,
|
| 363 |
+
"eval_logits/rejected": -1.082338571548462,
|
| 364 |
+
"eval_logps/chosen": -410.69976806640625,
|
| 365 |
+
"eval_logps/rejected": -384.83831787109375,
|
| 366 |
+
"eval_loss": 0.670859694480896,
|
| 367 |
+
"eval_rewards/accuracies": 0.5835432410240173,
|
| 368 |
+
"eval_rewards/chosen": 0.27643224596977234,
|
| 369 |
+
"eval_rewards/margins": 0.06646192073822021,
|
| 370 |
+
"eval_rewards/rejected": 0.20997031033039093,
|
| 371 |
+
"eval_runtime": 175.4317,
|
| 372 |
+
"eval_samples_per_second": 6.789,
|
| 373 |
+
"eval_steps_per_second": 6.789,
|
| 374 |
+
"step": 200
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.14499007508414602,
|
| 378 |
+
"grad_norm": 90.83344268798828,
|
| 379 |
+
"learning_rate": 4.278812974465148e-07,
|
| 380 |
+
"logits/chosen": -1.4185205698013306,
|
| 381 |
+
"logits/rejected": -1.2403172254562378,
|
| 382 |
+
"logps/chosen": -382.5022888183594,
|
| 383 |
+
"logps/rejected": -358.65582275390625,
|
| 384 |
+
"loss": 0.6733,
|
| 385 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 386 |
+
"rewards/chosen": 0.2773120105266571,
|
| 387 |
+
"rewards/margins": 0.06518922746181488,
|
| 388 |
+
"rewards/rejected": 0.21212276816368103,
|
| 389 |
+
"step": 210
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.15189436437386727,
|
| 393 |
+
"grad_norm": 107.23486328125,
|
| 394 |
+
"learning_rate": 4.2443064182194615e-07,
|
| 395 |
+
"logits/chosen": -1.3033957481384277,
|
| 396 |
+
"logits/rejected": -1.0996795892715454,
|
| 397 |
+
"logps/chosen": -407.9463195800781,
|
| 398 |
+
"logps/rejected": -421.99310302734375,
|
| 399 |
+
"loss": 0.6744,
|
| 400 |
+
"rewards/accuracies": 0.543749988079071,
|
| 401 |
+
"rewards/chosen": 0.3146436810493469,
|
| 402 |
+
"rewards/margins": 0.056582093238830566,
|
| 403 |
+
"rewards/rejected": 0.25806158781051636,
|
| 404 |
+
"step": 220
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.1587986536635885,
|
| 408 |
+
"grad_norm": 110.84426879882812,
|
| 409 |
+
"learning_rate": 4.209799861973775e-07,
|
| 410 |
+
"logits/chosen": -1.305482268333435,
|
| 411 |
+
"logits/rejected": -1.1424553394317627,
|
| 412 |
+
"logps/chosen": -403.65643310546875,
|
| 413 |
+
"logps/rejected": -390.69427490234375,
|
| 414 |
+
"loss": 0.6616,
|
| 415 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 416 |
+
"rewards/chosen": 0.32808226346969604,
|
| 417 |
+
"rewards/margins": 0.0847150906920433,
|
| 418 |
+
"rewards/rejected": 0.24336715042591095,
|
| 419 |
+
"step": 230
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.16570294295330976,
|
| 423 |
+
"grad_norm": 101.36908721923828,
|
| 424 |
+
"learning_rate": 4.175293305728088e-07,
|
| 425 |
+
"logits/chosen": -1.1226942539215088,
|
| 426 |
+
"logits/rejected": -0.9961945414543152,
|
| 427 |
+
"logps/chosen": -400.3923034667969,
|
| 428 |
+
"logps/rejected": -379.03253173828125,
|
| 429 |
+
"loss": 0.6891,
|
| 430 |
+
"rewards/accuracies": 0.543749988079071,
|
| 431 |
+
"rewards/chosen": 0.2640329897403717,
|
| 432 |
+
"rewards/margins": 0.02379865013062954,
|
| 433 |
+
"rewards/rejected": 0.24023433029651642,
|
| 434 |
+
"step": 240
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.17260723224303098,
|
| 438 |
+
"grad_norm": 111.8565902709961,
|
| 439 |
+
"learning_rate": 4.1407867494824017e-07,
|
| 440 |
+
"logits/chosen": -1.4368869066238403,
|
| 441 |
+
"logits/rejected": -1.1330538988113403,
|
| 442 |
+
"logps/chosen": -404.5845642089844,
|
| 443 |
+
"logps/rejected": -387.0537414550781,
|
| 444 |
+
"loss": 0.6547,
|
| 445 |
+
"rewards/accuracies": 0.59375,
|
| 446 |
+
"rewards/chosen": 0.3308050334453583,
|
| 447 |
+
"rewards/margins": 0.10226805508136749,
|
| 448 |
+
"rewards/rejected": 0.22853699326515198,
|
| 449 |
+
"step": 250
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.17260723224303098,
|
| 453 |
+
"eval_logits/chosen": NaN,
|
| 454 |
+
"eval_logits/rejected": -1.0943695306777954,
|
| 455 |
+
"eval_logps/chosen": -410.5004577636719,
|
| 456 |
+
"eval_logps/rejected": -384.76116943359375,
|
| 457 |
+
"eval_loss": 0.6669625639915466,
|
| 458 |
+
"eval_rewards/accuracies": 0.5894206762313843,
|
| 459 |
+
"eval_rewards/chosen": 0.2963607609272003,
|
| 460 |
+
"eval_rewards/margins": 0.07867557555437088,
|
| 461 |
+
"eval_rewards/rejected": 0.21768519282341003,
|
| 462 |
+
"eval_runtime": 174.8859,
|
| 463 |
+
"eval_samples_per_second": 6.81,
|
| 464 |
+
"eval_steps_per_second": 6.81,
|
| 465 |
+
"step": 250
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.1795115215327522,
|
| 469 |
+
"grad_norm": 83.30036926269531,
|
| 470 |
+
"learning_rate": 4.106280193236715e-07,
|
| 471 |
+
"logits/chosen": -1.3349380493164062,
|
| 472 |
+
"logits/rejected": -1.1268976926803589,
|
| 473 |
+
"logps/chosen": -408.28631591796875,
|
| 474 |
+
"logps/rejected": -409.681884765625,
|
| 475 |
+
"loss": 0.6812,
|
| 476 |
+
"rewards/accuracies": 0.625,
|
| 477 |
+
"rewards/chosen": 0.2973114848136902,
|
| 478 |
+
"rewards/margins": 0.0491001270711422,
|
| 479 |
+
"rewards/rejected": 0.2482113093137741,
|
| 480 |
+
"step": 260
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.18641581082247347,
|
| 484 |
+
"grad_norm": 109.46472930908203,
|
| 485 |
+
"learning_rate": 4.071773636991028e-07,
|
| 486 |
+
"logits/chosen": -1.2926867008209229,
|
| 487 |
+
"logits/rejected": -1.175357460975647,
|
| 488 |
+
"logps/chosen": -437.421142578125,
|
| 489 |
+
"logps/rejected": -408.375,
|
| 490 |
+
"loss": 0.6695,
|
| 491 |
+
"rewards/accuracies": 0.574999988079071,
|
| 492 |
+
"rewards/chosen": 0.28953924775123596,
|
| 493 |
+
"rewards/margins": 0.07461558282375336,
|
| 494 |
+
"rewards/rejected": 0.2149236649274826,
|
| 495 |
+
"step": 270
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.1933201001121947,
|
| 499 |
+
"grad_norm": 107.54530334472656,
|
| 500 |
+
"learning_rate": 4.0372670807453413e-07,
|
| 501 |
+
"logits/chosen": -1.3313913345336914,
|
| 502 |
+
"logits/rejected": -1.216606855392456,
|
| 503 |
+
"logps/chosen": -390.5809326171875,
|
| 504 |
+
"logps/rejected": -385.1927795410156,
|
| 505 |
+
"loss": 0.6815,
|
| 506 |
+
"rewards/accuracies": 0.550000011920929,
|
| 507 |
+
"rewards/chosen": 0.2860502302646637,
|
| 508 |
+
"rewards/margins": 0.05519815534353256,
|
| 509 |
+
"rewards/rejected": 0.23085205256938934,
|
| 510 |
+
"step": 280
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 0.20022438940191595,
|
| 514 |
+
"grad_norm": 100.78239440917969,
|
| 515 |
+
"learning_rate": 4.002760524499655e-07,
|
| 516 |
+
"logits/chosen": -1.4627584218978882,
|
| 517 |
+
"logits/rejected": -1.2909877300262451,
|
| 518 |
+
"logps/chosen": -370.5834045410156,
|
| 519 |
+
"logps/rejected": -371.1207275390625,
|
| 520 |
+
"loss": 0.6749,
|
| 521 |
+
"rewards/accuracies": 0.606249988079071,
|
| 522 |
+
"rewards/chosen": 0.29502958059310913,
|
| 523 |
+
"rewards/margins": 0.05848982185125351,
|
| 524 |
+
"rewards/rejected": 0.23653972148895264,
|
| 525 |
+
"step": 290
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.20712867869163717,
|
| 529 |
+
"grad_norm": 92.07577514648438,
|
| 530 |
+
"learning_rate": 3.968253968253968e-07,
|
| 531 |
+
"logits/chosen": -1.3123762607574463,
|
| 532 |
+
"logits/rejected": -1.1357301473617554,
|
| 533 |
+
"logps/chosen": -360.8210144042969,
|
| 534 |
+
"logps/rejected": -342.74444580078125,
|
| 535 |
+
"loss": 0.6613,
|
| 536 |
+
"rewards/accuracies": 0.625,
|
| 537 |
+
"rewards/chosen": 0.2599900960922241,
|
| 538 |
+
"rewards/margins": 0.08143080025911331,
|
| 539 |
+
"rewards/rejected": 0.17855927348136902,
|
| 540 |
+
"step": 300
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 0.20712867869163717,
|
| 544 |
+
"eval_logits/chosen": NaN,
|
| 545 |
+
"eval_logits/rejected": -1.1052350997924805,
|
| 546 |
+
"eval_logps/chosen": -410.4856872558594,
|
| 547 |
+
"eval_logps/rejected": -384.8339538574219,
|
| 548 |
+
"eval_loss": 0.6639109253883362,
|
| 549 |
+
"eval_rewards/accuracies": 0.5994962453842163,
|
| 550 |
+
"eval_rewards/chosen": 0.2978414297103882,
|
| 551 |
+
"eval_rewards/margins": 0.08743705600500107,
|
| 552 |
+
"eval_rewards/rejected": 0.2104043960571289,
|
| 553 |
+
"eval_runtime": 175.082,
|
| 554 |
+
"eval_samples_per_second": 6.803,
|
| 555 |
+
"eval_steps_per_second": 6.803,
|
| 556 |
+
"step": 300
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.21403296798135843,
|
| 560 |
+
"grad_norm": 96.54574584960938,
|
| 561 |
+
"learning_rate": 3.9337474120082815e-07,
|
| 562 |
+
"logits/chosen": -1.4912551641464233,
|
| 563 |
+
"logits/rejected": -1.2566022872924805,
|
| 564 |
+
"logps/chosen": -447.9432067871094,
|
| 565 |
+
"logps/rejected": -382.5755920410156,
|
| 566 |
+
"loss": 0.6313,
|
| 567 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 568 |
+
"rewards/chosen": 0.3629879057407379,
|
| 569 |
+
"rewards/margins": 0.16798029839992523,
|
| 570 |
+
"rewards/rejected": 0.19500760734081268,
|
| 571 |
+
"step": 310
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.22093725727107966,
|
| 575 |
+
"grad_norm": 88.83377075195312,
|
| 576 |
+
"learning_rate": 3.8992408557625947e-07,
|
| 577 |
+
"logits/chosen": NaN,
|
| 578 |
+
"logits/rejected": -1.2359377145767212,
|
| 579 |
+
"logps/chosen": -397.00042724609375,
|
| 580 |
+
"logps/rejected": -368.5748596191406,
|
| 581 |
+
"loss": 0.6573,
|
| 582 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 583 |
+
"rewards/chosen": 0.37342214584350586,
|
| 584 |
+
"rewards/margins": 0.10562543570995331,
|
| 585 |
+
"rewards/rejected": 0.26779669523239136,
|
| 586 |
+
"step": 320
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 0.22784154656080088,
|
| 590 |
+
"grad_norm": 92.1037368774414,
|
| 591 |
+
"learning_rate": 3.864734299516908e-07,
|
| 592 |
+
"logits/chosen": -1.4307029247283936,
|
| 593 |
+
"logits/rejected": -1.1998441219329834,
|
| 594 |
+
"logps/chosen": -365.99481201171875,
|
| 595 |
+
"logps/rejected": -349.28411865234375,
|
| 596 |
+
"loss": 0.6768,
|
| 597 |
+
"rewards/accuracies": 0.5625,
|
| 598 |
+
"rewards/chosen": 0.30214032530784607,
|
| 599 |
+
"rewards/margins": 0.06596145778894424,
|
| 600 |
+
"rewards/rejected": 0.23617887496948242,
|
| 601 |
+
"step": 330
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 0.23474583585052214,
|
| 605 |
+
"grad_norm": 102.44023895263672,
|
| 606 |
+
"learning_rate": 3.8302277432712217e-07,
|
| 607 |
+
"logits/chosen": -1.5167787075042725,
|
| 608 |
+
"logits/rejected": -1.3095741271972656,
|
| 609 |
+
"logps/chosen": -430.99078369140625,
|
| 610 |
+
"logps/rejected": -414.8297424316406,
|
| 611 |
+
"loss": 0.6712,
|
| 612 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 613 |
+
"rewards/chosen": 0.34007564187049866,
|
| 614 |
+
"rewards/margins": 0.0786903128027916,
|
| 615 |
+
"rewards/rejected": 0.26138535141944885,
|
| 616 |
+
"step": 340
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.24165012514024337,
|
| 620 |
+
"grad_norm": 99.54808044433594,
|
| 621 |
+
"learning_rate": 3.7957211870255344e-07,
|
| 622 |
+
"logits/chosen": -1.3461697101593018,
|
| 623 |
+
"logits/rejected": -1.3173012733459473,
|
| 624 |
+
"logps/chosen": -348.45855712890625,
|
| 625 |
+
"logps/rejected": -328.32958984375,
|
| 626 |
+
"loss": 0.6956,
|
| 627 |
+
"rewards/accuracies": 0.53125,
|
| 628 |
+
"rewards/chosen": 0.2547549903392792,
|
| 629 |
+
"rewards/margins": 0.02365240827202797,
|
| 630 |
+
"rewards/rejected": 0.23110255599021912,
|
| 631 |
+
"step": 350
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.24165012514024337,
|
| 635 |
+
"eval_logits/chosen": NaN,
|
| 636 |
+
"eval_logits/rejected": -1.1432876586914062,
|
| 637 |
+
"eval_logps/chosen": -410.524658203125,
|
| 638 |
+
"eval_logps/rejected": -384.90447998046875,
|
| 639 |
+
"eval_loss": 0.6644947528839111,
|
| 640 |
+
"eval_rewards/accuracies": 0.6087321639060974,
|
| 641 |
+
"eval_rewards/chosen": 0.29394418001174927,
|
| 642 |
+
"eval_rewards/margins": 0.09058759361505508,
|
| 643 |
+
"eval_rewards/rejected": 0.20335662364959717,
|
| 644 |
+
"eval_runtime": 174.1765,
|
| 645 |
+
"eval_samples_per_second": 6.838,
|
| 646 |
+
"eval_steps_per_second": 6.838,
|
| 647 |
+
"step": 350
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.24855441442996462,
|
| 651 |
+
"grad_norm": 95.7793960571289,
|
| 652 |
+
"learning_rate": 3.761214630779848e-07,
|
| 653 |
+
"logits/chosen": -1.4167577028274536,
|
| 654 |
+
"logits/rejected": -1.2274904251098633,
|
| 655 |
+
"logps/chosen": -398.9777526855469,
|
| 656 |
+
"logps/rejected": -375.252685546875,
|
| 657 |
+
"loss": 0.6706,
|
| 658 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 659 |
+
"rewards/chosen": 0.2948678433895111,
|
| 660 |
+
"rewards/margins": 0.07842870056629181,
|
| 661 |
+
"rewards/rejected": 0.2164391577243805,
|
| 662 |
+
"step": 360
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.2554587037196859,
|
| 666 |
+
"grad_norm": 115.86205291748047,
|
| 667 |
+
"learning_rate": 3.7267080745341613e-07,
|
| 668 |
+
"logits/chosen": -1.475426435470581,
|
| 669 |
+
"logits/rejected": -1.3543269634246826,
|
| 670 |
+
"logps/chosen": -379.80987548828125,
|
| 671 |
+
"logps/rejected": -393.3304138183594,
|
| 672 |
+
"loss": 0.6962,
|
| 673 |
+
"rewards/accuracies": 0.550000011920929,
|
| 674 |
+
"rewards/chosen": 0.2728275656700134,
|
| 675 |
+
"rewards/margins": 0.02714722231030464,
|
| 676 |
+
"rewards/rejected": 0.2456803321838379,
|
| 677 |
+
"step": 370
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.2623629930094071,
|
| 681 |
+
"grad_norm": 94.25291442871094,
|
| 682 |
+
"learning_rate": 3.6922015182884745e-07,
|
| 683 |
+
"logits/chosen": -1.3522151708602905,
|
| 684 |
+
"logits/rejected": -1.1761189699172974,
|
| 685 |
+
"logps/chosen": -410.3907165527344,
|
| 686 |
+
"logps/rejected": -408.0509948730469,
|
| 687 |
+
"loss": 0.6491,
|
| 688 |
+
"rewards/accuracies": 0.581250011920929,
|
| 689 |
+
"rewards/chosen": 0.3047269582748413,
|
| 690 |
+
"rewards/margins": 0.1313890665769577,
|
| 691 |
+
"rewards/rejected": 0.17333786189556122,
|
| 692 |
+
"step": 380
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.26926728229912833,
|
| 696 |
+
"grad_norm": 87.04241943359375,
|
| 697 |
+
"learning_rate": 3.657694962042788e-07,
|
| 698 |
+
"logits/chosen": -1.4277892112731934,
|
| 699 |
+
"logits/rejected": -1.2214381694793701,
|
| 700 |
+
"logps/chosen": -401.0085754394531,
|
| 701 |
+
"logps/rejected": -370.23626708984375,
|
| 702 |
+
"loss": 0.6507,
|
| 703 |
+
"rewards/accuracies": 0.625,
|
| 704 |
+
"rewards/chosen": 0.30590149760246277,
|
| 705 |
+
"rewards/margins": 0.11139049381017685,
|
| 706 |
+
"rewards/rejected": 0.1945110410451889,
|
| 707 |
+
"step": 390
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.2761715715888496,
|
| 711 |
+
"grad_norm": 92.03490447998047,
|
| 712 |
+
"learning_rate": 3.6231884057971015e-07,
|
| 713 |
+
"logits/chosen": -1.3768854141235352,
|
| 714 |
+
"logits/rejected": -1.2312657833099365,
|
| 715 |
+
"logps/chosen": -381.2527770996094,
|
| 716 |
+
"logps/rejected": -368.76116943359375,
|
| 717 |
+
"loss": 0.6524,
|
| 718 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 719 |
+
"rewards/chosen": 0.30460983514785767,
|
| 720 |
+
"rewards/margins": 0.1114836111664772,
|
| 721 |
+
"rewards/rejected": 0.19312623143196106,
|
| 722 |
+
"step": 400
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.2761715715888496,
|
| 726 |
+
"eval_logits/chosen": NaN,
|
| 727 |
+
"eval_logits/rejected": -1.1455719470977783,
|
| 728 |
+
"eval_logps/chosen": -410.375,
|
| 729 |
+
"eval_logps/rejected": -384.88275146484375,
|
| 730 |
+
"eval_loss": 0.6600396037101746,
|
| 731 |
+
"eval_rewards/accuracies": 0.6112510561943054,
|
| 732 |
+
"eval_rewards/chosen": 0.30890780687332153,
|
| 733 |
+
"eval_rewards/margins": 0.10338202118873596,
|
| 734 |
+
"eval_rewards/rejected": 0.20552580058574677,
|
| 735 |
+
"eval_runtime": 173.7586,
|
| 736 |
+
"eval_samples_per_second": 6.854,
|
| 737 |
+
"eval_steps_per_second": 6.854,
|
| 738 |
+
"step": 400
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.28307586087857084,
|
| 742 |
+
"grad_norm": 84.42977905273438,
|
| 743 |
+
"learning_rate": 3.5886818495514147e-07,
|
| 744 |
+
"logits/chosen": -1.5309593677520752,
|
| 745 |
+
"logits/rejected": -1.3086028099060059,
|
| 746 |
+
"logps/chosen": -412.37579345703125,
|
| 747 |
+
"logps/rejected": -371.64691162109375,
|
| 748 |
+
"loss": 0.6613,
|
| 749 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 750 |
+
"rewards/chosen": 0.30601978302001953,
|
| 751 |
+
"rewards/margins": 0.11061513423919678,
|
| 752 |
+
"rewards/rejected": 0.19540461897850037,
|
| 753 |
+
"step": 410
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.28998015016829204,
|
| 757 |
+
"grad_norm": 78.36170959472656,
|
| 758 |
+
"learning_rate": 3.554175293305728e-07,
|
| 759 |
+
"logits/chosen": -1.5359389781951904,
|
| 760 |
+
"logits/rejected": NaN,
|
| 761 |
+
"logps/chosen": -388.64056396484375,
|
| 762 |
+
"logps/rejected": -351.0355529785156,
|
| 763 |
+
"loss": 0.6426,
|
| 764 |
+
"rewards/accuracies": 0.625,
|
| 765 |
+
"rewards/chosen": 0.3550390601158142,
|
| 766 |
+
"rewards/margins": 0.1474730670452118,
|
| 767 |
+
"rewards/rejected": 0.20756597816944122,
|
| 768 |
+
"step": 420
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 0.2968844394580133,
|
| 772 |
+
"grad_norm": 115.095458984375,
|
| 773 |
+
"learning_rate": 3.5196687370600417e-07,
|
| 774 |
+
"logits/chosen": -1.5559413433074951,
|
| 775 |
+
"logits/rejected": -1.3967716693878174,
|
| 776 |
+
"logps/chosen": -410.49249267578125,
|
| 777 |
+
"logps/rejected": -402.0940246582031,
|
| 778 |
+
"loss": 0.6704,
|
| 779 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 780 |
+
"rewards/chosen": 0.3497038781642914,
|
| 781 |
+
"rewards/margins": 0.09456796944141388,
|
| 782 |
+
"rewards/rejected": 0.2551359534263611,
|
| 783 |
+
"step": 430
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.30378872874773455,
|
| 787 |
+
"grad_norm": 90.03668212890625,
|
| 788 |
+
"learning_rate": 3.4851621808143543e-07,
|
| 789 |
+
"logits/chosen": -1.3578782081604004,
|
| 790 |
+
"logits/rejected": -1.1276133060455322,
|
| 791 |
+
"logps/chosen": -388.08343505859375,
|
| 792 |
+
"logps/rejected": -370.1001281738281,
|
| 793 |
+
"loss": 0.6763,
|
| 794 |
+
"rewards/accuracies": 0.581250011920929,
|
| 795 |
+
"rewards/chosen": 0.33041852712631226,
|
| 796 |
+
"rewards/margins": 0.0695219561457634,
|
| 797 |
+
"rewards/rejected": 0.26089656352996826,
|
| 798 |
+
"step": 440
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.31069301803745575,
|
| 802 |
+
"grad_norm": 69.30516052246094,
|
| 803 |
+
"learning_rate": 3.450655624568668e-07,
|
| 804 |
+
"logits/chosen": -1.3265705108642578,
|
| 805 |
+
"logits/rejected": -1.1799354553222656,
|
| 806 |
+
"logps/chosen": -370.85882568359375,
|
| 807 |
+
"logps/rejected": -352.0668029785156,
|
| 808 |
+
"loss": 0.668,
|
| 809 |
+
"rewards/accuracies": 0.574999988079071,
|
| 810 |
+
"rewards/chosen": 0.32953041791915894,
|
| 811 |
+
"rewards/margins": 0.09510184824466705,
|
| 812 |
+
"rewards/rejected": 0.2344285249710083,
|
| 813 |
+
"step": 450
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.31069301803745575,
|
| 817 |
+
"eval_logits/chosen": NaN,
|
| 818 |
+
"eval_logits/rejected": -1.1227965354919434,
|
| 819 |
+
"eval_logps/chosen": -409.8273620605469,
|
| 820 |
+
"eval_logps/rejected": -384.4886474609375,
|
| 821 |
+
"eval_loss": 0.6568603515625,
|
| 822 |
+
"eval_rewards/accuracies": 0.6120907068252563,
|
| 823 |
+
"eval_rewards/chosen": 0.36367088556289673,
|
| 824 |
+
"eval_rewards/margins": 0.1187346950173378,
|
| 825 |
+
"eval_rewards/rejected": 0.24493615329265594,
|
| 826 |
+
"eval_runtime": 173.6419,
|
| 827 |
+
"eval_samples_per_second": 6.859,
|
| 828 |
+
"eval_steps_per_second": 6.859,
|
| 829 |
+
"step": 450
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.317597307327177,
|
| 833 |
+
"grad_norm": 78.89984893798828,
|
| 834 |
+
"learning_rate": 3.416149068322981e-07,
|
| 835 |
+
"logits/chosen": NaN,
|
| 836 |
+
"logits/rejected": -1.305570363998413,
|
| 837 |
+
"logps/chosen": -416.13726806640625,
|
| 838 |
+
"logps/rejected": -372.0033264160156,
|
| 839 |
+
"loss": 0.6481,
|
| 840 |
+
"rewards/accuracies": 0.612500011920929,
|
| 841 |
+
"rewards/chosen": 0.4147353172302246,
|
| 842 |
+
"rewards/margins": 0.1446826159954071,
|
| 843 |
+
"rewards/rejected": 0.2700527012348175,
|
| 844 |
+
"step": 460
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.32450159661689826,
|
| 848 |
+
"grad_norm": 103.23446655273438,
|
| 849 |
+
"learning_rate": 3.3816425120772945e-07,
|
| 850 |
+
"logits/chosen": -1.5534855127334595,
|
| 851 |
+
"logits/rejected": NaN,
|
| 852 |
+
"logps/chosen": -368.0318603515625,
|
| 853 |
+
"logps/rejected": -359.0155944824219,
|
| 854 |
+
"loss": 0.6966,
|
| 855 |
+
"rewards/accuracies": 0.53125,
|
| 856 |
+
"rewards/chosen": 0.3369843661785126,
|
| 857 |
+
"rewards/margins": 0.03579873591661453,
|
| 858 |
+
"rewards/rejected": 0.30118563771247864,
|
| 859 |
+
"step": 470
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.3314058859066195,
|
| 863 |
+
"grad_norm": 78.23011016845703,
|
| 864 |
+
"learning_rate": 3.347135955831608e-07,
|
| 865 |
+
"logits/chosen": -1.5334765911102295,
|
| 866 |
+
"logits/rejected": -1.240417242050171,
|
| 867 |
+
"logps/chosen": -399.98333740234375,
|
| 868 |
+
"logps/rejected": -373.2457580566406,
|
| 869 |
+
"loss": 0.6489,
|
| 870 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 871 |
+
"rewards/chosen": 0.35515138506889343,
|
| 872 |
+
"rewards/margins": 0.13600310683250427,
|
| 873 |
+
"rewards/rejected": 0.21914830803871155,
|
| 874 |
+
"step": 480
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.3383101751963407,
|
| 878 |
+
"grad_norm": 101.94794464111328,
|
| 879 |
+
"learning_rate": 3.312629399585921e-07,
|
| 880 |
+
"logits/chosen": -1.6701557636260986,
|
| 881 |
+
"logits/rejected": -1.3532097339630127,
|
| 882 |
+
"logps/chosen": -370.28363037109375,
|
| 883 |
+
"logps/rejected": -324.91845703125,
|
| 884 |
+
"loss": 0.6381,
|
| 885 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 886 |
+
"rewards/chosen": 0.3905836045742035,
|
| 887 |
+
"rewards/margins": 0.1694364845752716,
|
| 888 |
+
"rewards/rejected": 0.22114713490009308,
|
| 889 |
+
"step": 490
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.34521446448606197,
|
| 893 |
+
"grad_norm": 133.3175811767578,
|
| 894 |
+
"learning_rate": 3.2781228433402347e-07,
|
| 895 |
+
"logits/chosen": -1.3769439458847046,
|
| 896 |
+
"logits/rejected": -1.239466905593872,
|
| 897 |
+
"logps/chosen": -426.37969970703125,
|
| 898 |
+
"logps/rejected": -407.44061279296875,
|
| 899 |
+
"loss": 0.6857,
|
| 900 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 901 |
+
"rewards/chosen": 0.31460708379745483,
|
| 902 |
+
"rewards/margins": 0.06561081111431122,
|
| 903 |
+
"rewards/rejected": 0.24899625778198242,
|
| 904 |
+
"step": 500
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.34521446448606197,
|
| 908 |
+
"eval_logits/chosen": NaN,
|
| 909 |
+
"eval_logits/rejected": -1.135837197303772,
|
| 910 |
+
"eval_logps/chosen": -409.88800048828125,
|
| 911 |
+
"eval_logps/rejected": -384.6177978515625,
|
| 912 |
+
"eval_loss": 0.6550039052963257,
|
| 913 |
+
"eval_rewards/accuracies": 0.6045340299606323,
|
| 914 |
+
"eval_rewards/chosen": 0.3576072156429291,
|
| 915 |
+
"eval_rewards/margins": 0.1255844682455063,
|
| 916 |
+
"eval_rewards/rejected": 0.2320227324962616,
|
| 917 |
+
"eval_runtime": 174.8216,
|
| 918 |
+
"eval_samples_per_second": 6.813,
|
| 919 |
+
"eval_steps_per_second": 6.813,
|
| 920 |
+
"step": 500
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.3521187537757832,
|
| 924 |
+
"grad_norm": 80.95104217529297,
|
| 925 |
+
"learning_rate": 3.243616287094548e-07,
|
| 926 |
+
"logits/chosen": -1.50025475025177,
|
| 927 |
+
"logits/rejected": -1.2558667659759521,
|
| 928 |
+
"logps/chosen": -384.44903564453125,
|
| 929 |
+
"logps/rejected": -405.1640930175781,
|
| 930 |
+
"loss": 0.6756,
|
| 931 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 932 |
+
"rewards/chosen": 0.32988351583480835,
|
| 933 |
+
"rewards/margins": 0.08733965456485748,
|
| 934 |
+
"rewards/rejected": 0.24254386126995087,
|
| 935 |
+
"step": 510
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.3590230430655044,
|
| 939 |
+
"grad_norm": 92.58390045166016,
|
| 940 |
+
"learning_rate": 3.209109730848861e-07,
|
| 941 |
+
"logits/chosen": -1.4795461893081665,
|
| 942 |
+
"logits/rejected": -1.2118721008300781,
|
| 943 |
+
"logps/chosen": -381.4693298339844,
|
| 944 |
+
"logps/rejected": -409.9659729003906,
|
| 945 |
+
"loss": 0.6841,
|
| 946 |
+
"rewards/accuracies": 0.574999988079071,
|
| 947 |
+
"rewards/chosen": 0.2862910032272339,
|
| 948 |
+
"rewards/margins": 0.05146180838346481,
|
| 949 |
+
"rewards/rejected": 0.23482923209667206,
|
| 950 |
+
"step": 520
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.3659273323552257,
|
| 954 |
+
"grad_norm": 91.9266128540039,
|
| 955 |
+
"learning_rate": 3.1746031746031743e-07,
|
| 956 |
+
"logits/chosen": -1.3780999183654785,
|
| 957 |
+
"logits/rejected": -1.1586534976959229,
|
| 958 |
+
"logps/chosen": -426.71636962890625,
|
| 959 |
+
"logps/rejected": -420.63250732421875,
|
| 960 |
+
"loss": 0.6447,
|
| 961 |
+
"rewards/accuracies": 0.637499988079071,
|
| 962 |
+
"rewards/chosen": 0.3207171559333801,
|
| 963 |
+
"rewards/margins": 0.14687559008598328,
|
| 964 |
+
"rewards/rejected": 0.17384156584739685,
|
| 965 |
+
"step": 530
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 0.37283162164494693,
|
| 969 |
+
"grad_norm": 107.24482727050781,
|
| 970 |
+
"learning_rate": 3.140096618357488e-07,
|
| 971 |
+
"logits/chosen": -1.5469199419021606,
|
| 972 |
+
"logits/rejected": -1.3309047222137451,
|
| 973 |
+
"logps/chosen": -423.83251953125,
|
| 974 |
+
"logps/rejected": -390.2005310058594,
|
| 975 |
+
"loss": 0.649,
|
| 976 |
+
"rewards/accuracies": 0.59375,
|
| 977 |
+
"rewards/chosen": 0.35180041193962097,
|
| 978 |
+
"rewards/margins": 0.13371917605400085,
|
| 979 |
+
"rewards/rejected": 0.2180812656879425,
|
| 980 |
+
"step": 540
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.3797359109346682,
|
| 984 |
+
"grad_norm": 90.34844207763672,
|
| 985 |
+
"learning_rate": 3.105590062111801e-07,
|
| 986 |
+
"logits/chosen": -1.437620759010315,
|
| 987 |
+
"logits/rejected": -1.2692636251449585,
|
| 988 |
+
"logps/chosen": -375.0314025878906,
|
| 989 |
+
"logps/rejected": -353.4309997558594,
|
| 990 |
+
"loss": 0.6666,
|
| 991 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 992 |
+
"rewards/chosen": 0.305647611618042,
|
| 993 |
+
"rewards/margins": 0.09037254005670547,
|
| 994 |
+
"rewards/rejected": 0.2152750939130783,
|
| 995 |
+
"step": 550
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 0.3797359109346682,
|
| 999 |
+
"eval_logits/chosen": NaN,
|
| 1000 |
+
"eval_logits/rejected": -1.1454120874404907,
|
| 1001 |
+
"eval_logps/chosen": -410.1690673828125,
|
| 1002 |
+
"eval_logps/rejected": -384.90673828125,
|
| 1003 |
+
"eval_loss": 0.6544755697250366,
|
| 1004 |
+
"eval_rewards/accuracies": 0.6129302978515625,
|
| 1005 |
+
"eval_rewards/chosen": 0.3294997811317444,
|
| 1006 |
+
"eval_rewards/margins": 0.12637175619602203,
|
| 1007 |
+
"eval_rewards/rejected": 0.20312805473804474,
|
| 1008 |
+
"eval_runtime": 173.6702,
|
| 1009 |
+
"eval_samples_per_second": 6.858,
|
| 1010 |
+
"eval_steps_per_second": 6.858,
|
| 1011 |
+
"step": 550
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.3866402002243894,
|
| 1015 |
+
"grad_norm": 85.39889526367188,
|
| 1016 |
+
"learning_rate": 3.0710835058661145e-07,
|
| 1017 |
+
"logits/chosen": -1.47105872631073,
|
| 1018 |
+
"logits/rejected": -1.2406635284423828,
|
| 1019 |
+
"logps/chosen": -343.89788818359375,
|
| 1020 |
+
"logps/rejected": -328.2047119140625,
|
| 1021 |
+
"loss": 0.6619,
|
| 1022 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1023 |
+
"rewards/chosen": 0.2944934368133545,
|
| 1024 |
+
"rewards/margins": 0.11027739197015762,
|
| 1025 |
+
"rewards/rejected": 0.18421605229377747,
|
| 1026 |
+
"step": 560
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 0.39354448951411064,
|
| 1030 |
+
"grad_norm": 96.35163116455078,
|
| 1031 |
+
"learning_rate": 3.036576949620428e-07,
|
| 1032 |
+
"logits/chosen": -1.636962890625,
|
| 1033 |
+
"logits/rejected": -1.3636503219604492,
|
| 1034 |
+
"logps/chosen": -404.2951965332031,
|
| 1035 |
+
"logps/rejected": -391.49615478515625,
|
| 1036 |
+
"loss": 0.6572,
|
| 1037 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1038 |
+
"rewards/chosen": 0.3543318808078766,
|
| 1039 |
+
"rewards/margins": 0.14378124475479126,
|
| 1040 |
+
"rewards/rejected": 0.21055062115192413,
|
| 1041 |
+
"step": 570
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.4004487788038319,
|
| 1045 |
+
"grad_norm": 99.60575866699219,
|
| 1046 |
+
"learning_rate": 3.002070393374741e-07,
|
| 1047 |
+
"logits/chosen": -1.3705114126205444,
|
| 1048 |
+
"logits/rejected": -1.0337412357330322,
|
| 1049 |
+
"logps/chosen": -356.34637451171875,
|
| 1050 |
+
"logps/rejected": -346.79766845703125,
|
| 1051 |
+
"loss": 0.6537,
|
| 1052 |
+
"rewards/accuracies": 0.59375,
|
| 1053 |
+
"rewards/chosen": 0.2836061120033264,
|
| 1054 |
+
"rewards/margins": 0.13429149985313416,
|
| 1055 |
+
"rewards/rejected": 0.14931461215019226,
|
| 1056 |
+
"step": 580
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 0.4073530680935531,
|
| 1060 |
+
"grad_norm": 96.50724029541016,
|
| 1061 |
+
"learning_rate": 2.9675638371290547e-07,
|
| 1062 |
+
"logits/chosen": -1.4598503112792969,
|
| 1063 |
+
"logits/rejected": -1.2610881328582764,
|
| 1064 |
+
"logps/chosen": -410.27117919921875,
|
| 1065 |
+
"logps/rejected": -395.2956237792969,
|
| 1066 |
+
"loss": 0.6455,
|
| 1067 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1068 |
+
"rewards/chosen": 0.379031240940094,
|
| 1069 |
+
"rewards/margins": 0.15479955077171326,
|
| 1070 |
+
"rewards/rejected": 0.22423171997070312,
|
| 1071 |
+
"step": 590
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.41425735738327435,
|
| 1075 |
+
"grad_norm": 104.0567626953125,
|
| 1076 |
+
"learning_rate": 2.9330572808833673e-07,
|
| 1077 |
+
"logits/chosen": -1.329871654510498,
|
| 1078 |
+
"logits/rejected": -1.1407099962234497,
|
| 1079 |
+
"logps/chosen": -380.8253173828125,
|
| 1080 |
+
"logps/rejected": -364.3226013183594,
|
| 1081 |
+
"loss": 0.6657,
|
| 1082 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1083 |
+
"rewards/chosen": 0.31459444761276245,
|
| 1084 |
+
"rewards/margins": 0.11110667884349823,
|
| 1085 |
+
"rewards/rejected": 0.20348772406578064,
|
| 1086 |
+
"step": 600
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 0.41425735738327435,
|
| 1090 |
+
"eval_logits/chosen": NaN,
|
| 1091 |
+
"eval_logits/rejected": -1.137957215309143,
|
| 1092 |
+
"eval_logps/chosen": -410.0342102050781,
|
| 1093 |
+
"eval_logps/rejected": -384.80487060546875,
|
| 1094 |
+
"eval_loss": 0.6545748114585876,
|
| 1095 |
+
"eval_rewards/accuracies": 0.6179680824279785,
|
| 1096 |
+
"eval_rewards/chosen": 0.34298890829086304,
|
| 1097 |
+
"eval_rewards/margins": 0.1296708583831787,
|
| 1098 |
+
"eval_rewards/rejected": 0.21331804990768433,
|
| 1099 |
+
"eval_runtime": 173.6763,
|
| 1100 |
+
"eval_samples_per_second": 6.858,
|
| 1101 |
+
"eval_steps_per_second": 6.858,
|
| 1102 |
+
"step": 600
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.4211616466729956,
|
| 1106 |
+
"grad_norm": 100.69998168945312,
|
| 1107 |
+
"learning_rate": 2.898550724637681e-07,
|
| 1108 |
+
"logits/chosen": -1.3926723003387451,
|
| 1109 |
+
"logits/rejected": -1.302227258682251,
|
| 1110 |
+
"logps/chosen": -382.15863037109375,
|
| 1111 |
+
"logps/rejected": -377.8441467285156,
|
| 1112 |
+
"loss": 0.6654,
|
| 1113 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1114 |
+
"rewards/chosen": 0.3457278609275818,
|
| 1115 |
+
"rewards/margins": 0.10164386034011841,
|
| 1116 |
+
"rewards/rejected": 0.24408404529094696,
|
| 1117 |
+
"step": 610
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.42806593596271686,
|
| 1121 |
+
"grad_norm": 90.8844985961914,
|
| 1122 |
+
"learning_rate": 2.8640441683919943e-07,
|
| 1123 |
+
"logits/chosen": -1.4987343549728394,
|
| 1124 |
+
"logits/rejected": -1.3479670286178589,
|
| 1125 |
+
"logps/chosen": -374.33746337890625,
|
| 1126 |
+
"logps/rejected": -344.2901916503906,
|
| 1127 |
+
"loss": 0.6639,
|
| 1128 |
+
"rewards/accuracies": 0.59375,
|
| 1129 |
+
"rewards/chosen": 0.32695019245147705,
|
| 1130 |
+
"rewards/margins": 0.105133056640625,
|
| 1131 |
+
"rewards/rejected": 0.22181710600852966,
|
| 1132 |
+
"step": 620
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.43497022525243806,
|
| 1136 |
+
"grad_norm": 100.92221069335938,
|
| 1137 |
+
"learning_rate": 2.8295376121463075e-07,
|
| 1138 |
+
"logits/chosen": -1.6069577932357788,
|
| 1139 |
+
"logits/rejected": -1.357591986656189,
|
| 1140 |
+
"logps/chosen": -391.8199157714844,
|
| 1141 |
+
"logps/rejected": -361.18756103515625,
|
| 1142 |
+
"loss": 0.6445,
|
| 1143 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1144 |
+
"rewards/chosen": 0.3474247455596924,
|
| 1145 |
+
"rewards/margins": 0.15058517456054688,
|
| 1146 |
+
"rewards/rejected": 0.1968395859003067,
|
| 1147 |
+
"step": 630
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.4418745145421593,
|
| 1151 |
+
"grad_norm": 90.86432647705078,
|
| 1152 |
+
"learning_rate": 2.7950310559006207e-07,
|
| 1153 |
+
"logits/chosen": -1.4821056127548218,
|
| 1154 |
+
"logits/rejected": -1.32998526096344,
|
| 1155 |
+
"logps/chosen": -390.75848388671875,
|
| 1156 |
+
"logps/rejected": -352.44219970703125,
|
| 1157 |
+
"loss": 0.6924,
|
| 1158 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1159 |
+
"rewards/chosen": 0.2883882522583008,
|
| 1160 |
+
"rewards/margins": 0.04232420399785042,
|
| 1161 |
+
"rewards/rejected": 0.24606403708457947,
|
| 1162 |
+
"step": 640
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.44877880383188057,
|
| 1166 |
+
"grad_norm": 99.2775650024414,
|
| 1167 |
+
"learning_rate": 2.7605244996549345e-07,
|
| 1168 |
+
"logits/chosen": -1.398462176322937,
|
| 1169 |
+
"logits/rejected": NaN,
|
| 1170 |
+
"logps/chosen": -411.4620056152344,
|
| 1171 |
+
"logps/rejected": -401.382080078125,
|
| 1172 |
+
"loss": 0.6576,
|
| 1173 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 1174 |
+
"rewards/chosen": 0.3147468864917755,
|
| 1175 |
+
"rewards/margins": 0.1398230642080307,
|
| 1176 |
+
"rewards/rejected": 0.1749238222837448,
|
| 1177 |
+
"step": 650
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 0.44877880383188057,
|
| 1181 |
+
"eval_logits/chosen": NaN,
|
| 1182 |
+
"eval_logits/rejected": -1.155210018157959,
|
| 1183 |
+
"eval_logps/chosen": -410.21453857421875,
|
| 1184 |
+
"eval_logps/rejected": -385.0280456542969,
|
| 1185 |
+
"eval_loss": 0.6529302000999451,
|
| 1186 |
+
"eval_rewards/accuracies": 0.6179680824279785,
|
| 1187 |
+
"eval_rewards/chosen": 0.3249528408050537,
|
| 1188 |
+
"eval_rewards/margins": 0.13395507633686066,
|
| 1189 |
+
"eval_rewards/rejected": 0.19099776446819305,
|
| 1190 |
+
"eval_runtime": 173.1301,
|
| 1191 |
+
"eval_samples_per_second": 6.879,
|
| 1192 |
+
"eval_steps_per_second": 6.879,
|
| 1193 |
+
"step": 650
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.45568309312160177,
|
| 1197 |
+
"grad_norm": 92.97110748291016,
|
| 1198 |
+
"learning_rate": 2.7260179434092477e-07,
|
| 1199 |
+
"logits/chosen": -1.43343985080719,
|
| 1200 |
+
"logits/rejected": -1.3144687414169312,
|
| 1201 |
+
"logps/chosen": -399.0343933105469,
|
| 1202 |
+
"logps/rejected": -371.15106201171875,
|
| 1203 |
+
"loss": 0.6415,
|
| 1204 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1205 |
+
"rewards/chosen": 0.3595283031463623,
|
| 1206 |
+
"rewards/margins": 0.15820467472076416,
|
| 1207 |
+
"rewards/rejected": 0.20132365822792053,
|
| 1208 |
+
"step": 660
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.462587382411323,
|
| 1212 |
+
"grad_norm": 103.37547302246094,
|
| 1213 |
+
"learning_rate": 2.691511387163561e-07,
|
| 1214 |
+
"logits/chosen": -1.4777318239212036,
|
| 1215 |
+
"logits/rejected": -1.2895824909210205,
|
| 1216 |
+
"logps/chosen": -414.9844665527344,
|
| 1217 |
+
"logps/rejected": -380.60174560546875,
|
| 1218 |
+
"loss": 0.66,
|
| 1219 |
+
"rewards/accuracies": 0.625,
|
| 1220 |
+
"rewards/chosen": 0.34182053804397583,
|
| 1221 |
+
"rewards/margins": 0.12233030796051025,
|
| 1222 |
+
"rewards/rejected": 0.21949021518230438,
|
| 1223 |
+
"step": 670
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 0.4694916717010443,
|
| 1227 |
+
"grad_norm": 96.86497497558594,
|
| 1228 |
+
"learning_rate": 2.6570048309178746e-07,
|
| 1229 |
+
"logits/chosen": -1.3753479719161987,
|
| 1230 |
+
"logits/rejected": -1.2074251174926758,
|
| 1231 |
+
"logps/chosen": -373.104248046875,
|
| 1232 |
+
"logps/rejected": -358.5914001464844,
|
| 1233 |
+
"loss": 0.6344,
|
| 1234 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1235 |
+
"rewards/chosen": 0.32691770792007446,
|
| 1236 |
+
"rewards/margins": 0.17708414793014526,
|
| 1237 |
+
"rewards/rejected": 0.1498335748910904,
|
| 1238 |
+
"step": 680
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.47639596099076553,
|
| 1242 |
+
"grad_norm": 101.70399475097656,
|
| 1243 |
+
"learning_rate": 2.6224982746721873e-07,
|
| 1244 |
+
"logits/chosen": -1.4386012554168701,
|
| 1245 |
+
"logits/rejected": -1.2293193340301514,
|
| 1246 |
+
"logps/chosen": -370.0113220214844,
|
| 1247 |
+
"logps/rejected": -367.55242919921875,
|
| 1248 |
+
"loss": 0.6806,
|
| 1249 |
+
"rewards/accuracies": 0.59375,
|
| 1250 |
+
"rewards/chosen": 0.3327386975288391,
|
| 1251 |
+
"rewards/margins": 0.07743687927722931,
|
| 1252 |
+
"rewards/rejected": 0.2553017735481262,
|
| 1253 |
+
"step": 690
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 0.48330025028048673,
|
| 1257 |
+
"grad_norm": 99.53569030761719,
|
| 1258 |
+
"learning_rate": 2.587991718426501e-07,
|
| 1259 |
+
"logits/chosen": -1.4941202402114868,
|
| 1260 |
+
"logits/rejected": -1.3218339681625366,
|
| 1261 |
+
"logps/chosen": -389.9266357421875,
|
| 1262 |
+
"logps/rejected": -346.59326171875,
|
| 1263 |
+
"loss": 0.6406,
|
| 1264 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1265 |
+
"rewards/chosen": 0.3879491984844208,
|
| 1266 |
+
"rewards/margins": 0.1433873325586319,
|
| 1267 |
+
"rewards/rejected": 0.24456188082695007,
|
| 1268 |
+
"step": 700
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.48330025028048673,
|
| 1272 |
+
"eval_logits/chosen": NaN,
|
| 1273 |
+
"eval_logits/rejected": -1.1317625045776367,
|
| 1274 |
+
"eval_logps/chosen": -409.8088684082031,
|
| 1275 |
+
"eval_logps/rejected": -384.7127685546875,
|
| 1276 |
+
"eval_loss": 0.6519591212272644,
|
| 1277 |
+
"eval_rewards/accuracies": 0.6196473836898804,
|
| 1278 |
+
"eval_rewards/chosen": 0.36551961302757263,
|
| 1279 |
+
"eval_rewards/margins": 0.14299476146697998,
|
| 1280 |
+
"eval_rewards/rejected": 0.22252482175827026,
|
| 1281 |
+
"eval_runtime": 173.3348,
|
| 1282 |
+
"eval_samples_per_second": 6.871,
|
| 1283 |
+
"eval_steps_per_second": 6.871,
|
| 1284 |
+
"step": 700
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.490204539570208,
|
| 1288 |
+
"grad_norm": 97.43363189697266,
|
| 1289 |
+
"learning_rate": 2.553485162180814e-07,
|
| 1290 |
+
"logits/chosen": -1.5055333375930786,
|
| 1291 |
+
"logits/rejected": -1.2466411590576172,
|
| 1292 |
+
"logps/chosen": -424.33074951171875,
|
| 1293 |
+
"logps/rejected": -419.7621154785156,
|
| 1294 |
+
"loss": 0.6469,
|
| 1295 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1296 |
+
"rewards/chosen": 0.43039917945861816,
|
| 1297 |
+
"rewards/margins": 0.15889059007167816,
|
| 1298 |
+
"rewards/rejected": 0.2715086042881012,
|
| 1299 |
+
"step": 710
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.49710882885992924,
|
| 1303 |
+
"grad_norm": 100.15609741210938,
|
| 1304 |
+
"learning_rate": 2.5189786059351275e-07,
|
| 1305 |
+
"logits/chosen": NaN,
|
| 1306 |
+
"logits/rejected": -1.3018251657485962,
|
| 1307 |
+
"logps/chosen": -383.4604187011719,
|
| 1308 |
+
"logps/rejected": -386.2461853027344,
|
| 1309 |
+
"loss": 0.6626,
|
| 1310 |
+
"rewards/accuracies": 0.59375,
|
| 1311 |
+
"rewards/chosen": 0.3409530222415924,
|
| 1312 |
+
"rewards/margins": 0.13614806532859802,
|
| 1313 |
+
"rewards/rejected": 0.20480497181415558,
|
| 1314 |
+
"step": 720
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.5040131181496504,
|
| 1318 |
+
"grad_norm": 104.17349243164062,
|
| 1319 |
+
"learning_rate": 2.4844720496894407e-07,
|
| 1320 |
+
"logits/chosen": -1.3319756984710693,
|
| 1321 |
+
"logits/rejected": -1.1958295106887817,
|
| 1322 |
+
"logps/chosen": -379.02313232421875,
|
| 1323 |
+
"logps/rejected": -340.41943359375,
|
| 1324 |
+
"loss": 0.689,
|
| 1325 |
+
"rewards/accuracies": 0.53125,
|
| 1326 |
+
"rewards/chosen": 0.29082250595092773,
|
| 1327 |
+
"rewards/margins": 0.06135740876197815,
|
| 1328 |
+
"rewards/rejected": 0.22946509718894958,
|
| 1329 |
+
"step": 730
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.5109174074393718,
|
| 1333 |
+
"grad_norm": 100.09748077392578,
|
| 1334 |
+
"learning_rate": 2.449965493443754e-07,
|
| 1335 |
+
"logits/chosen": -1.4177885055541992,
|
| 1336 |
+
"logits/rejected": -1.2871652841567993,
|
| 1337 |
+
"logps/chosen": -377.2490234375,
|
| 1338 |
+
"logps/rejected": -341.6965637207031,
|
| 1339 |
+
"loss": 0.6714,
|
| 1340 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1341 |
+
"rewards/chosen": 0.4028010368347168,
|
| 1342 |
+
"rewards/margins": 0.11562231928110123,
|
| 1343 |
+
"rewards/rejected": 0.28717875480651855,
|
| 1344 |
+
"step": 740
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 0.517821696729093,
|
| 1348 |
+
"grad_norm": 103.06417846679688,
|
| 1349 |
+
"learning_rate": 2.4154589371980677e-07,
|
| 1350 |
+
"logits/chosen": -1.4369862079620361,
|
| 1351 |
+
"logits/rejected": -1.2680126428604126,
|
| 1352 |
+
"logps/chosen": -375.4537658691406,
|
| 1353 |
+
"logps/rejected": -380.01666259765625,
|
| 1354 |
+
"loss": 0.654,
|
| 1355 |
+
"rewards/accuracies": 0.65625,
|
| 1356 |
+
"rewards/chosen": 0.34813031554222107,
|
| 1357 |
+
"rewards/margins": 0.13279417157173157,
|
| 1358 |
+
"rewards/rejected": 0.2153361290693283,
|
| 1359 |
+
"step": 750
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 0.517821696729093,
|
| 1363 |
+
"eval_logits/chosen": NaN,
|
| 1364 |
+
"eval_logits/rejected": -1.128820538520813,
|
| 1365 |
+
"eval_logps/chosen": -409.6957092285156,
|
| 1366 |
+
"eval_logps/rejected": -384.63482666015625,
|
| 1367 |
+
"eval_loss": 0.651150643825531,
|
| 1368 |
+
"eval_rewards/accuracies": 0.6062132716178894,
|
| 1369 |
+
"eval_rewards/chosen": 0.37683534622192383,
|
| 1370 |
+
"eval_rewards/margins": 0.14651274681091309,
|
| 1371 |
+
"eval_rewards/rejected": 0.23032261431217194,
|
| 1372 |
+
"eval_runtime": 173.301,
|
| 1373 |
+
"eval_samples_per_second": 6.872,
|
| 1374 |
+
"eval_steps_per_second": 6.872,
|
| 1375 |
+
"step": 750
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.5247259860188142,
|
| 1379 |
+
"grad_norm": 109.10247802734375,
|
| 1380 |
+
"learning_rate": 2.3809523809523806e-07,
|
| 1381 |
+
"logits/chosen": -1.5034782886505127,
|
| 1382 |
+
"logits/rejected": -1.2673709392547607,
|
| 1383 |
+
"logps/chosen": -394.3305969238281,
|
| 1384 |
+
"logps/rejected": -371.8570251464844,
|
| 1385 |
+
"loss": 0.6226,
|
| 1386 |
+
"rewards/accuracies": 0.65625,
|
| 1387 |
+
"rewards/chosen": 0.3872641921043396,
|
| 1388 |
+
"rewards/margins": 0.20504610240459442,
|
| 1389 |
+
"rewards/rejected": 0.18221807479858398,
|
| 1390 |
+
"step": 760
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 0.5316302753085355,
|
| 1394 |
+
"grad_norm": 96.4645767211914,
|
| 1395 |
+
"learning_rate": 2.3464458247066943e-07,
|
| 1396 |
+
"logits/chosen": -1.4274332523345947,
|
| 1397 |
+
"logits/rejected": -1.2091481685638428,
|
| 1398 |
+
"logps/chosen": -425.9805603027344,
|
| 1399 |
+
"logps/rejected": -407.9385681152344,
|
| 1400 |
+
"loss": 0.6542,
|
| 1401 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1402 |
+
"rewards/chosen": 0.40589994192123413,
|
| 1403 |
+
"rewards/margins": 0.16390803456306458,
|
| 1404 |
+
"rewards/rejected": 0.24199192225933075,
|
| 1405 |
+
"step": 770
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 0.5385345645982567,
|
| 1409 |
+
"grad_norm": 111.85659790039062,
|
| 1410 |
+
"learning_rate": 2.3119392684610076e-07,
|
| 1411 |
+
"logits/chosen": NaN,
|
| 1412 |
+
"logits/rejected": NaN,
|
| 1413 |
+
"logps/chosen": -409.7381896972656,
|
| 1414 |
+
"logps/rejected": -390.08599853515625,
|
| 1415 |
+
"loss": 0.6571,
|
| 1416 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1417 |
+
"rewards/chosen": 0.4092329442501068,
|
| 1418 |
+
"rewards/margins": 0.14268992841243744,
|
| 1419 |
+
"rewards/rejected": 0.2665430009365082,
|
| 1420 |
+
"step": 780
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.5454388538879779,
|
| 1424 |
+
"grad_norm": 107.36064910888672,
|
| 1425 |
+
"learning_rate": 2.2774327122153208e-07,
|
| 1426 |
+
"logits/chosen": -1.2583667039871216,
|
| 1427 |
+
"logits/rejected": -1.238249659538269,
|
| 1428 |
+
"logps/chosen": -377.5832214355469,
|
| 1429 |
+
"logps/rejected": -362.2899169921875,
|
| 1430 |
+
"loss": 0.6579,
|
| 1431 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1432 |
+
"rewards/chosen": 0.3105475902557373,
|
| 1433 |
+
"rewards/margins": 0.12654425203800201,
|
| 1434 |
+
"rewards/rejected": 0.1840033382177353,
|
| 1435 |
+
"step": 790
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 0.5523431431776992,
|
| 1439 |
+
"grad_norm": 96.03706359863281,
|
| 1440 |
+
"learning_rate": 2.2429261559696343e-07,
|
| 1441 |
+
"logits/chosen": -1.3471505641937256,
|
| 1442 |
+
"logits/rejected": -1.0923124551773071,
|
| 1443 |
+
"logps/chosen": -367.17803955078125,
|
| 1444 |
+
"logps/rejected": -356.579833984375,
|
| 1445 |
+
"loss": 0.6456,
|
| 1446 |
+
"rewards/accuracies": 0.625,
|
| 1447 |
+
"rewards/chosen": 0.3641411364078522,
|
| 1448 |
+
"rewards/margins": 0.1728210747241974,
|
| 1449 |
+
"rewards/rejected": 0.1913200318813324,
|
| 1450 |
+
"step": 800
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 0.5523431431776992,
|
| 1454 |
+
"eval_logits/chosen": NaN,
|
| 1455 |
+
"eval_logits/rejected": -1.1347825527191162,
|
| 1456 |
+
"eval_logps/chosen": -409.65533447265625,
|
| 1457 |
+
"eval_logps/rejected": -384.67364501953125,
|
| 1458 |
+
"eval_loss": 0.6488075256347656,
|
| 1459 |
+
"eval_rewards/accuracies": 0.6146095991134644,
|
| 1460 |
+
"eval_rewards/chosen": 0.3808766305446625,
|
| 1461 |
+
"eval_rewards/margins": 0.15444040298461914,
|
| 1462 |
+
"eval_rewards/rejected": 0.22643625736236572,
|
| 1463 |
+
"eval_runtime": 173.4914,
|
| 1464 |
+
"eval_samples_per_second": 6.865,
|
| 1465 |
+
"eval_steps_per_second": 6.865,
|
| 1466 |
+
"step": 800
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.5592474324674204,
|
| 1470 |
+
"grad_norm": 106.62773132324219,
|
| 1471 |
+
"learning_rate": 2.2084195997239475e-07,
|
| 1472 |
+
"logits/chosen": -1.6777063608169556,
|
| 1473 |
+
"logits/rejected": -1.4368447065353394,
|
| 1474 |
+
"logps/chosen": -391.08135986328125,
|
| 1475 |
+
"logps/rejected": -382.46142578125,
|
| 1476 |
+
"loss": 0.6622,
|
| 1477 |
+
"rewards/accuracies": 0.5625,
|
| 1478 |
+
"rewards/chosen": 0.38126569986343384,
|
| 1479 |
+
"rewards/margins": 0.11523087322711945,
|
| 1480 |
+
"rewards/rejected": 0.2660347819328308,
|
| 1481 |
+
"step": 810
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 0.5661517217571417,
|
| 1485 |
+
"grad_norm": 83.2169418334961,
|
| 1486 |
+
"learning_rate": 2.1739130434782607e-07,
|
| 1487 |
+
"logits/chosen": -1.4315296411514282,
|
| 1488 |
+
"logits/rejected": -1.306997537612915,
|
| 1489 |
+
"logps/chosen": -373.8775329589844,
|
| 1490 |
+
"logps/rejected": -349.00909423828125,
|
| 1491 |
+
"loss": 0.6473,
|
| 1492 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1493 |
+
"rewards/chosen": 0.3956912159919739,
|
| 1494 |
+
"rewards/margins": 0.17349565029144287,
|
| 1495 |
+
"rewards/rejected": 0.2221955806016922,
|
| 1496 |
+
"step": 820
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.5730560110468629,
|
| 1500 |
+
"grad_norm": 121.26054382324219,
|
| 1501 |
+
"learning_rate": 2.139406487232574e-07,
|
| 1502 |
+
"logits/chosen": -1.4633328914642334,
|
| 1503 |
+
"logits/rejected": -1.3243625164031982,
|
| 1504 |
+
"logps/chosen": -400.9346008300781,
|
| 1505 |
+
"logps/rejected": -395.54705810546875,
|
| 1506 |
+
"loss": 0.6725,
|
| 1507 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1508 |
+
"rewards/chosen": 0.36117586493492126,
|
| 1509 |
+
"rewards/margins": 0.12074120342731476,
|
| 1510 |
+
"rewards/rejected": 0.2404346913099289,
|
| 1511 |
+
"step": 830
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 0.5799603003365841,
|
| 1515 |
+
"grad_norm": 98.78184509277344,
|
| 1516 |
+
"learning_rate": 2.1048999309868874e-07,
|
| 1517 |
+
"logits/chosen": -1.5927917957305908,
|
| 1518 |
+
"logits/rejected": -1.3326520919799805,
|
| 1519 |
+
"logps/chosen": -395.4635009765625,
|
| 1520 |
+
"logps/rejected": -390.7279357910156,
|
| 1521 |
+
"loss": 0.7039,
|
| 1522 |
+
"rewards/accuracies": 0.543749988079071,
|
| 1523 |
+
"rewards/chosen": 0.2971932291984558,
|
| 1524 |
+
"rewards/margins": 0.03939943015575409,
|
| 1525 |
+
"rewards/rejected": 0.2577938139438629,
|
| 1526 |
+
"step": 840
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.5868645896263054,
|
| 1530 |
+
"grad_norm": 83.3376235961914,
|
| 1531 |
+
"learning_rate": 2.0703933747412008e-07,
|
| 1532 |
+
"logits/chosen": -1.4211909770965576,
|
| 1533 |
+
"logits/rejected": -1.2685062885284424,
|
| 1534 |
+
"logps/chosen": -404.7445983886719,
|
| 1535 |
+
"logps/rejected": -383.61993408203125,
|
| 1536 |
+
"loss": 0.634,
|
| 1537 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1538 |
+
"rewards/chosen": 0.37307268381118774,
|
| 1539 |
+
"rewards/margins": 0.16583478450775146,
|
| 1540 |
+
"rewards/rejected": 0.2072378695011139,
|
| 1541 |
+
"step": 850
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 0.5868645896263054,
|
| 1545 |
+
"eval_logits/chosen": NaN,
|
| 1546 |
+
"eval_logits/rejected": -1.1369266510009766,
|
| 1547 |
+
"eval_logps/chosen": -409.7511901855469,
|
| 1548 |
+
"eval_logps/rejected": -384.7900695800781,
|
| 1549 |
+
"eval_loss": 0.6476391553878784,
|
| 1550 |
+
"eval_rewards/accuracies": 0.6087321639060974,
|
| 1551 |
+
"eval_rewards/chosen": 0.3712875247001648,
|
| 1552 |
+
"eval_rewards/margins": 0.15649177134037018,
|
| 1553 |
+
"eval_rewards/rejected": 0.2147957682609558,
|
| 1554 |
+
"eval_runtime": 173.685,
|
| 1555 |
+
"eval_samples_per_second": 6.857,
|
| 1556 |
+
"eval_steps_per_second": 6.857,
|
| 1557 |
+
"step": 850
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.5937688789160266,
|
| 1561 |
+
"grad_norm": 91.06554412841797,
|
| 1562 |
+
"learning_rate": 2.035886818495514e-07,
|
| 1563 |
+
"logits/chosen": -1.365419864654541,
|
| 1564 |
+
"logits/rejected": -1.2031277418136597,
|
| 1565 |
+
"logps/chosen": -363.187744140625,
|
| 1566 |
+
"logps/rejected": -333.2200622558594,
|
| 1567 |
+
"loss": 0.6489,
|
| 1568 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1569 |
+
"rewards/chosen": 0.3162587881088257,
|
| 1570 |
+
"rewards/margins": 0.14883024990558624,
|
| 1571 |
+
"rewards/rejected": 0.16742852330207825,
|
| 1572 |
+
"step": 860
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.6006731682057478,
|
| 1576 |
+
"grad_norm": 95.876220703125,
|
| 1577 |
+
"learning_rate": 2.0013802622498275e-07,
|
| 1578 |
+
"logits/chosen": NaN,
|
| 1579 |
+
"logits/rejected": -1.3413794040679932,
|
| 1580 |
+
"logps/chosen": -389.54595947265625,
|
| 1581 |
+
"logps/rejected": -358.7523498535156,
|
| 1582 |
+
"loss": 0.6445,
|
| 1583 |
+
"rewards/accuracies": 0.625,
|
| 1584 |
+
"rewards/chosen": 0.3688567280769348,
|
| 1585 |
+
"rewards/margins": 0.15345630049705505,
|
| 1586 |
+
"rewards/rejected": 0.21540041267871857,
|
| 1587 |
+
"step": 870
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.6075774574954691,
|
| 1591 |
+
"grad_norm": 99.45269012451172,
|
| 1592 |
+
"learning_rate": 1.9668737060041408e-07,
|
| 1593 |
+
"logits/chosen": -1.4373767375946045,
|
| 1594 |
+
"logits/rejected": NaN,
|
| 1595 |
+
"logps/chosen": -378.1737365722656,
|
| 1596 |
+
"logps/rejected": -353.63250732421875,
|
| 1597 |
+
"loss": 0.6638,
|
| 1598 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1599 |
+
"rewards/chosen": 0.33772993087768555,
|
| 1600 |
+
"rewards/margins": 0.11045277118682861,
|
| 1601 |
+
"rewards/rejected": 0.22727715969085693,
|
| 1602 |
+
"step": 880
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 0.6144817467851903,
|
| 1606 |
+
"grad_norm": 94.84121704101562,
|
| 1607 |
+
"learning_rate": 1.932367149758454e-07,
|
| 1608 |
+
"logits/chosen": -1.5210530757904053,
|
| 1609 |
+
"logits/rejected": -1.3680551052093506,
|
| 1610 |
+
"logps/chosen": -404.3708190917969,
|
| 1611 |
+
"logps/rejected": -369.14068603515625,
|
| 1612 |
+
"loss": 0.6429,
|
| 1613 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1614 |
+
"rewards/chosen": 0.40372776985168457,
|
| 1615 |
+
"rewards/margins": 0.16238902509212494,
|
| 1616 |
+
"rewards/rejected": 0.2413388043642044,
|
| 1617 |
+
"step": 890
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.6213860360749115,
|
| 1621 |
+
"grad_norm": 90.55352783203125,
|
| 1622 |
+
"learning_rate": 1.8978605935127672e-07,
|
| 1623 |
+
"logits/chosen": -1.4078751802444458,
|
| 1624 |
+
"logits/rejected": -1.2929816246032715,
|
| 1625 |
+
"logps/chosen": -353.96453857421875,
|
| 1626 |
+
"logps/rejected": -348.5416564941406,
|
| 1627 |
+
"loss": 0.66,
|
| 1628 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1629 |
+
"rewards/chosen": 0.34200939536094666,
|
| 1630 |
+
"rewards/margins": 0.11792133003473282,
|
| 1631 |
+
"rewards/rejected": 0.22408807277679443,
|
| 1632 |
+
"step": 900
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 0.6213860360749115,
|
| 1636 |
+
"eval_logits/chosen": NaN,
|
| 1637 |
+
"eval_logits/rejected": -1.139318823814392,
|
| 1638 |
+
"eval_logps/chosen": -409.7491760253906,
|
| 1639 |
+
"eval_logps/rejected": -384.7567138671875,
|
| 1640 |
+
"eval_loss": 0.6495311856269836,
|
| 1641 |
+
"eval_rewards/accuracies": 0.6204869747161865,
|
| 1642 |
+
"eval_rewards/chosen": 0.3714897930622101,
|
| 1643 |
+
"eval_rewards/margins": 0.15335862338542938,
|
| 1644 |
+
"eval_rewards/rejected": 0.2181311994791031,
|
| 1645 |
+
"eval_runtime": 173.8621,
|
| 1646 |
+
"eval_samples_per_second": 6.85,
|
| 1647 |
+
"eval_steps_per_second": 6.85,
|
| 1648 |
+
"step": 900
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.6282903253646328,
|
| 1652 |
+
"grad_norm": 84.32146453857422,
|
| 1653 |
+
"learning_rate": 1.8633540372670807e-07,
|
| 1654 |
+
"logits/chosen": NaN,
|
| 1655 |
+
"logits/rejected": -1.269124984741211,
|
| 1656 |
+
"logps/chosen": -383.15533447265625,
|
| 1657 |
+
"logps/rejected": -359.85833740234375,
|
| 1658 |
+
"loss": 0.6471,
|
| 1659 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1660 |
+
"rewards/chosen": 0.38757017254829407,
|
| 1661 |
+
"rewards/margins": 0.15741722285747528,
|
| 1662 |
+
"rewards/rejected": 0.230152890086174,
|
| 1663 |
+
"step": 910
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 0.635194614654354,
|
| 1667 |
+
"grad_norm": 109.78142547607422,
|
| 1668 |
+
"learning_rate": 1.828847481021394e-07,
|
| 1669 |
+
"logits/chosen": -1.4139900207519531,
|
| 1670 |
+
"logits/rejected": -1.3312866687774658,
|
| 1671 |
+
"logps/chosen": -381.28277587890625,
|
| 1672 |
+
"logps/rejected": -390.725830078125,
|
| 1673 |
+
"loss": 0.6325,
|
| 1674 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1675 |
+
"rewards/chosen": 0.3692389130592346,
|
| 1676 |
+
"rewards/margins": 0.17626893520355225,
|
| 1677 |
+
"rewards/rejected": 0.19296997785568237,
|
| 1678 |
+
"step": 920
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.6420989039440752,
|
| 1682 |
+
"grad_norm": 107.46936798095703,
|
| 1683 |
+
"learning_rate": 1.7943409247757073e-07,
|
| 1684 |
+
"logits/chosen": -1.4537097215652466,
|
| 1685 |
+
"logits/rejected": -1.2186715602874756,
|
| 1686 |
+
"logps/chosen": -396.77392578125,
|
| 1687 |
+
"logps/rejected": -374.40618896484375,
|
| 1688 |
+
"loss": 0.6333,
|
| 1689 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1690 |
+
"rewards/chosen": 0.4124184548854828,
|
| 1691 |
+
"rewards/margins": 0.21722407639026642,
|
| 1692 |
+
"rewards/rejected": 0.19519445300102234,
|
| 1693 |
+
"step": 930
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 0.6490031932337965,
|
| 1697 |
+
"grad_norm": 87.2940444946289,
|
| 1698 |
+
"learning_rate": 1.7598343685300208e-07,
|
| 1699 |
+
"logits/chosen": -1.4950411319732666,
|
| 1700 |
+
"logits/rejected": NaN,
|
| 1701 |
+
"logps/chosen": -382.4498596191406,
|
| 1702 |
+
"logps/rejected": -320.4793701171875,
|
| 1703 |
+
"loss": 0.6422,
|
| 1704 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1705 |
+
"rewards/chosen": 0.36269697546958923,
|
| 1706 |
+
"rewards/margins": 0.15182676911354065,
|
| 1707 |
+
"rewards/rejected": 0.2108701765537262,
|
| 1708 |
+
"step": 940
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 0.6559074825235177,
|
| 1712 |
+
"grad_norm": 86.76087951660156,
|
| 1713 |
+
"learning_rate": 1.725327812284334e-07,
|
| 1714 |
+
"logits/chosen": -1.2916945219039917,
|
| 1715 |
+
"logits/rejected": -1.1377825736999512,
|
| 1716 |
+
"logps/chosen": -370.7120666503906,
|
| 1717 |
+
"logps/rejected": -355.99212646484375,
|
| 1718 |
+
"loss": 0.6847,
|
| 1719 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1720 |
+
"rewards/chosen": 0.3327692747116089,
|
| 1721 |
+
"rewards/margins": 0.07058731466531754,
|
| 1722 |
+
"rewards/rejected": 0.26218199729919434,
|
| 1723 |
+
"step": 950
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 0.6559074825235177,
|
| 1727 |
+
"eval_logits/chosen": NaN,
|
| 1728 |
+
"eval_logits/rejected": -1.1284286975860596,
|
| 1729 |
+
"eval_logps/chosen": -409.7359313964844,
|
| 1730 |
+
"eval_logps/rejected": -384.813720703125,
|
| 1731 |
+
"eval_loss": 0.647310197353363,
|
| 1732 |
+
"eval_rewards/accuracies": 0.6078925132751465,
|
| 1733 |
+
"eval_rewards/chosen": 0.37281379103660583,
|
| 1734 |
+
"eval_rewards/margins": 0.16038304567337036,
|
| 1735 |
+
"eval_rewards/rejected": 0.21243073046207428,
|
| 1736 |
+
"eval_runtime": 173.875,
|
| 1737 |
+
"eval_samples_per_second": 6.85,
|
| 1738 |
+
"eval_steps_per_second": 6.85,
|
| 1739 |
+
"step": 950
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 0.662811771813239,
|
| 1743 |
+
"grad_norm": 84.70791625976562,
|
| 1744 |
+
"learning_rate": 1.6908212560386473e-07,
|
| 1745 |
+
"logits/chosen": -1.2661709785461426,
|
| 1746 |
+
"logits/rejected": -1.1103928089141846,
|
| 1747 |
+
"logps/chosen": -381.80291748046875,
|
| 1748 |
+
"logps/rejected": -328.13555908203125,
|
| 1749 |
+
"loss": 0.6615,
|
| 1750 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1751 |
+
"rewards/chosen": 0.31213271617889404,
|
| 1752 |
+
"rewards/margins": 0.13148072361946106,
|
| 1753 |
+
"rewards/rejected": 0.18065199255943298,
|
| 1754 |
+
"step": 960
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.6697160611029602,
|
| 1758 |
+
"grad_norm": 97.82742309570312,
|
| 1759 |
+
"learning_rate": 1.6563146997929605e-07,
|
| 1760 |
+
"logits/chosen": -1.4364001750946045,
|
| 1761 |
+
"logits/rejected": -1.3133249282836914,
|
| 1762 |
+
"logps/chosen": -417.947998046875,
|
| 1763 |
+
"logps/rejected": -395.54681396484375,
|
| 1764 |
+
"loss": 0.6598,
|
| 1765 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1766 |
+
"rewards/chosen": 0.40047353506088257,
|
| 1767 |
+
"rewards/margins": 0.12428589165210724,
|
| 1768 |
+
"rewards/rejected": 0.2761876583099365,
|
| 1769 |
+
"step": 970
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.6766203503926814,
|
| 1773 |
+
"grad_norm": 85.0315170288086,
|
| 1774 |
+
"learning_rate": 1.621808143547274e-07,
|
| 1775 |
+
"logits/chosen": -1.5771453380584717,
|
| 1776 |
+
"logits/rejected": -1.3503344058990479,
|
| 1777 |
+
"logps/chosen": -363.5293884277344,
|
| 1778 |
+
"logps/rejected": -334.8780212402344,
|
| 1779 |
+
"loss": 0.6364,
|
| 1780 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1781 |
+
"rewards/chosen": 0.3424316346645355,
|
| 1782 |
+
"rewards/margins": 0.15654315054416656,
|
| 1783 |
+
"rewards/rejected": 0.18588848412036896,
|
| 1784 |
+
"step": 980
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.6835246396824027,
|
| 1788 |
+
"grad_norm": 126.81350708007812,
|
| 1789 |
+
"learning_rate": 1.5873015873015872e-07,
|
| 1790 |
+
"logits/chosen": -1.4660173654556274,
|
| 1791 |
+
"logits/rejected": -1.261212706565857,
|
| 1792 |
+
"logps/chosen": -408.24102783203125,
|
| 1793 |
+
"logps/rejected": -408.9168701171875,
|
| 1794 |
+
"loss": 0.6551,
|
| 1795 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1796 |
+
"rewards/chosen": 0.3640829920768738,
|
| 1797 |
+
"rewards/margins": 0.13493572175502777,
|
| 1798 |
+
"rewards/rejected": 0.2291472852230072,
|
| 1799 |
+
"step": 990
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 0.6904289289721239,
|
| 1803 |
+
"grad_norm": 101.64756774902344,
|
| 1804 |
+
"learning_rate": 1.5527950310559004e-07,
|
| 1805 |
+
"logits/chosen": -1.4593558311462402,
|
| 1806 |
+
"logits/rejected": -1.284190058708191,
|
| 1807 |
+
"logps/chosen": -428.12188720703125,
|
| 1808 |
+
"logps/rejected": -404.9797668457031,
|
| 1809 |
+
"loss": 0.6473,
|
| 1810 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1811 |
+
"rewards/chosen": 0.4170054495334625,
|
| 1812 |
+
"rewards/margins": 0.16244842112064362,
|
| 1813 |
+
"rewards/rejected": 0.2545570433139801,
|
| 1814 |
+
"step": 1000
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.6904289289721239,
|
| 1818 |
+
"eval_logits/chosen": NaN,
|
| 1819 |
+
"eval_logits/rejected": -1.1269783973693848,
|
| 1820 |
+
"eval_logps/chosen": -409.7304382324219,
|
| 1821 |
+
"eval_logps/rejected": -384.82025146484375,
|
| 1822 |
+
"eval_loss": 0.6481216549873352,
|
| 1823 |
+
"eval_rewards/accuracies": 0.6078925132751465,
|
| 1824 |
+
"eval_rewards/chosen": 0.3733593821525574,
|
| 1825 |
+
"eval_rewards/margins": 0.16157987713813782,
|
| 1826 |
+
"eval_rewards/rejected": 0.21177950501441956,
|
| 1827 |
+
"eval_runtime": 173.8531,
|
| 1828 |
+
"eval_samples_per_second": 6.851,
|
| 1829 |
+
"eval_steps_per_second": 6.851,
|
| 1830 |
+
"step": 1000
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 0.6973332182618451,
|
| 1834 |
+
"grad_norm": 79.46784973144531,
|
| 1835 |
+
"learning_rate": 1.518288474810214e-07,
|
| 1836 |
+
"logits/chosen": -1.3219681978225708,
|
| 1837 |
+
"logits/rejected": -1.1236135959625244,
|
| 1838 |
+
"logps/chosen": -415.108642578125,
|
| 1839 |
+
"logps/rejected": -416.01629638671875,
|
| 1840 |
+
"loss": 0.636,
|
| 1841 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1842 |
+
"rewards/chosen": 0.3638695180416107,
|
| 1843 |
+
"rewards/margins": 0.19566866755485535,
|
| 1844 |
+
"rewards/rejected": 0.16820085048675537,
|
| 1845 |
+
"step": 1010
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.7042375075515664,
|
| 1849 |
+
"grad_norm": 98.8138198852539,
|
| 1850 |
+
"learning_rate": 1.4837819185645273e-07,
|
| 1851 |
+
"logits/chosen": -1.5049182176589966,
|
| 1852 |
+
"logits/rejected": -1.3138519525527954,
|
| 1853 |
+
"logps/chosen": -380.9891052246094,
|
| 1854 |
+
"logps/rejected": -373.53729248046875,
|
| 1855 |
+
"loss": 0.6605,
|
| 1856 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1857 |
+
"rewards/chosen": 0.35565224289894104,
|
| 1858 |
+
"rewards/margins": 0.1293681114912033,
|
| 1859 |
+
"rewards/rejected": 0.22628410160541534,
|
| 1860 |
+
"step": 1020
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.7111417968412876,
|
| 1864 |
+
"grad_norm": 96.57266235351562,
|
| 1865 |
+
"learning_rate": 1.4492753623188405e-07,
|
| 1866 |
+
"logits/chosen": -1.3988029956817627,
|
| 1867 |
+
"logits/rejected": NaN,
|
| 1868 |
+
"logps/chosen": -406.1911926269531,
|
| 1869 |
+
"logps/rejected": -391.26300048828125,
|
| 1870 |
+
"loss": 0.6774,
|
| 1871 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1872 |
+
"rewards/chosen": 0.3720816671848297,
|
| 1873 |
+
"rewards/margins": 0.09232926368713379,
|
| 1874 |
+
"rewards/rejected": 0.2797524034976959,
|
| 1875 |
+
"step": 1030
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 0.7180460861310088,
|
| 1879 |
+
"grad_norm": 104.09722900390625,
|
| 1880 |
+
"learning_rate": 1.4147688060731538e-07,
|
| 1881 |
+
"logits/chosen": -1.5075652599334717,
|
| 1882 |
+
"logits/rejected": -1.2237682342529297,
|
| 1883 |
+
"logps/chosen": -392.244873046875,
|
| 1884 |
+
"logps/rejected": -347.1589660644531,
|
| 1885 |
+
"loss": 0.6265,
|
| 1886 |
+
"rewards/accuracies": 0.6875,
|
| 1887 |
+
"rewards/chosen": 0.3723699748516083,
|
| 1888 |
+
"rewards/margins": 0.1985810399055481,
|
| 1889 |
+
"rewards/rejected": 0.17378894984722137,
|
| 1890 |
+
"step": 1040
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 0.7249503754207302,
|
| 1894 |
+
"grad_norm": 95.77185821533203,
|
| 1895 |
+
"learning_rate": 1.3802622498274672e-07,
|
| 1896 |
+
"logits/chosen": -1.3212801218032837,
|
| 1897 |
+
"logits/rejected": -1.1590490341186523,
|
| 1898 |
+
"logps/chosen": -377.89013671875,
|
| 1899 |
+
"logps/rejected": -363.0843811035156,
|
| 1900 |
+
"loss": 0.6494,
|
| 1901 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1902 |
+
"rewards/chosen": 0.33318403363227844,
|
| 1903 |
+
"rewards/margins": 0.15657344460487366,
|
| 1904 |
+
"rewards/rejected": 0.17661058902740479,
|
| 1905 |
+
"step": 1050
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.7249503754207302,
|
| 1909 |
+
"eval_logits/chosen": NaN,
|
| 1910 |
+
"eval_logits/rejected": -1.1044600009918213,
|
| 1911 |
+
"eval_logps/chosen": -409.47552490234375,
|
| 1912 |
+
"eval_logps/rejected": -384.6523742675781,
|
| 1913 |
+
"eval_loss": 0.6458028554916382,
|
| 1914 |
+
"eval_rewards/accuracies": 0.6087321639060974,
|
| 1915 |
+
"eval_rewards/chosen": 0.39885976910591125,
|
| 1916 |
+
"eval_rewards/margins": 0.17029373347759247,
|
| 1917 |
+
"eval_rewards/rejected": 0.22856605052947998,
|
| 1918 |
+
"eval_runtime": 173.7412,
|
| 1919 |
+
"eval_samples_per_second": 6.855,
|
| 1920 |
+
"eval_steps_per_second": 6.855,
|
| 1921 |
+
"step": 1050
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 0.7318546647104514,
|
| 1925 |
+
"grad_norm": 82.92915344238281,
|
| 1926 |
+
"learning_rate": 1.3457556935817804e-07,
|
| 1927 |
+
"logits/chosen": -1.450514316558838,
|
| 1928 |
+
"logits/rejected": -1.1329553127288818,
|
| 1929 |
+
"logps/chosen": -370.6721496582031,
|
| 1930 |
+
"logps/rejected": -376.0997314453125,
|
| 1931 |
+
"loss": 0.6139,
|
| 1932 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1933 |
+
"rewards/chosen": 0.4179357886314392,
|
| 1934 |
+
"rewards/margins": 0.23517921566963196,
|
| 1935 |
+
"rewards/rejected": 0.18275660276412964,
|
| 1936 |
+
"step": 1060
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.7387589540001726,
|
| 1940 |
+
"grad_norm": 95.6248779296875,
|
| 1941 |
+
"learning_rate": 1.3112491373360937e-07,
|
| 1942 |
+
"logits/chosen": -1.4916682243347168,
|
| 1943 |
+
"logits/rejected": NaN,
|
| 1944 |
+
"logps/chosen": -399.62060546875,
|
| 1945 |
+
"logps/rejected": -378.38043212890625,
|
| 1946 |
+
"loss": 0.6628,
|
| 1947 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1948 |
+
"rewards/chosen": 0.4076872766017914,
|
| 1949 |
+
"rewards/margins": 0.12090712785720825,
|
| 1950 |
+
"rewards/rejected": 0.2867801785469055,
|
| 1951 |
+
"step": 1070
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.7456632432898939,
|
| 1955 |
+
"grad_norm": 87.97006225585938,
|
| 1956 |
+
"learning_rate": 1.276742581090407e-07,
|
| 1957 |
+
"logits/chosen": -1.2332895994186401,
|
| 1958 |
+
"logits/rejected": -1.0287469625473022,
|
| 1959 |
+
"logps/chosen": -379.1185302734375,
|
| 1960 |
+
"logps/rejected": -376.4650573730469,
|
| 1961 |
+
"loss": 0.6665,
|
| 1962 |
+
"rewards/accuracies": 0.59375,
|
| 1963 |
+
"rewards/chosen": 0.3581882119178772,
|
| 1964 |
+
"rewards/margins": 0.11071789264678955,
|
| 1965 |
+
"rewards/rejected": 0.24747030436992645,
|
| 1966 |
+
"step": 1080
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 0.7525675325796151,
|
| 1970 |
+
"grad_norm": 88.0757064819336,
|
| 1971 |
+
"learning_rate": 1.2422360248447204e-07,
|
| 1972 |
+
"logits/chosen": -1.3982305526733398,
|
| 1973 |
+
"logits/rejected": -1.2792531251907349,
|
| 1974 |
+
"logps/chosen": -404.583984375,
|
| 1975 |
+
"logps/rejected": -371.7993469238281,
|
| 1976 |
+
"loss": 0.655,
|
| 1977 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1978 |
+
"rewards/chosen": 0.36442241072654724,
|
| 1979 |
+
"rewards/margins": 0.13779766857624054,
|
| 1980 |
+
"rewards/rejected": 0.2266247272491455,
|
| 1981 |
+
"step": 1090
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 0.7594718218693364,
|
| 1985 |
+
"grad_norm": 94.24323272705078,
|
| 1986 |
+
"learning_rate": 1.2077294685990338e-07,
|
| 1987 |
+
"logits/chosen": -1.3688199520111084,
|
| 1988 |
+
"logits/rejected": -1.2200040817260742,
|
| 1989 |
+
"logps/chosen": -402.47320556640625,
|
| 1990 |
+
"logps/rejected": -378.7218933105469,
|
| 1991 |
+
"loss": 0.6456,
|
| 1992 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1993 |
+
"rewards/chosen": 0.3827332854270935,
|
| 1994 |
+
"rewards/margins": 0.14937618374824524,
|
| 1995 |
+
"rewards/rejected": 0.23335710167884827,
|
| 1996 |
+
"step": 1100
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 0.7594718218693364,
|
| 2000 |
+
"eval_logits/chosen": NaN,
|
| 2001 |
+
"eval_logits/rejected": -1.0987216234207153,
|
| 2002 |
+
"eval_logps/chosen": -409.48699951171875,
|
| 2003 |
+
"eval_logps/rejected": -384.6772766113281,
|
| 2004 |
+
"eval_loss": 0.6459550857543945,
|
| 2005 |
+
"eval_rewards/accuracies": 0.6204869747161865,
|
| 2006 |
+
"eval_rewards/chosen": 0.39770370721817017,
|
| 2007 |
+
"eval_rewards/margins": 0.1716250628232956,
|
| 2008 |
+
"eval_rewards/rejected": 0.22607862949371338,
|
| 2009 |
+
"eval_runtime": 173.2214,
|
| 2010 |
+
"eval_samples_per_second": 6.876,
|
| 2011 |
+
"eval_steps_per_second": 6.876,
|
| 2012 |
+
"step": 1100
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.7663761111590576,
|
| 2016 |
+
"grad_norm": 97.47419738769531,
|
| 2017 |
+
"learning_rate": 1.1732229123533472e-07,
|
| 2018 |
+
"logits/chosen": -1.2847949266433716,
|
| 2019 |
+
"logits/rejected": -1.2520158290863037,
|
| 2020 |
+
"logps/chosen": -397.53900146484375,
|
| 2021 |
+
"logps/rejected": -373.2650451660156,
|
| 2022 |
+
"loss": 0.6825,
|
| 2023 |
+
"rewards/accuracies": 0.581250011920929,
|
| 2024 |
+
"rewards/chosen": 0.35880032181739807,
|
| 2025 |
+
"rewards/margins": 0.0934886708855629,
|
| 2026 |
+
"rewards/rejected": 0.26531165838241577,
|
| 2027 |
+
"step": 1110
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.7732804004487788,
|
| 2031 |
+
"grad_norm": 118.96842956542969,
|
| 2032 |
+
"learning_rate": 1.1387163561076604e-07,
|
| 2033 |
+
"logits/chosen": -1.404904842376709,
|
| 2034 |
+
"logits/rejected": NaN,
|
| 2035 |
+
"logps/chosen": -397.9188537597656,
|
| 2036 |
+
"logps/rejected": -357.46038818359375,
|
| 2037 |
+
"loss": 0.6383,
|
| 2038 |
+
"rewards/accuracies": 0.625,
|
| 2039 |
+
"rewards/chosen": 0.3993021249771118,
|
| 2040 |
+
"rewards/margins": 0.18723976612091064,
|
| 2041 |
+
"rewards/rejected": 0.21206238865852356,
|
| 2042 |
+
"step": 1120
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.7801846897385001,
|
| 2046 |
+
"grad_norm": 92.28319549560547,
|
| 2047 |
+
"learning_rate": 1.1042097998619737e-07,
|
| 2048 |
+
"logits/chosen": -1.4132840633392334,
|
| 2049 |
+
"logits/rejected": -1.1864253282546997,
|
| 2050 |
+
"logps/chosen": -376.830810546875,
|
| 2051 |
+
"logps/rejected": -378.5279846191406,
|
| 2052 |
+
"loss": 0.6474,
|
| 2053 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2054 |
+
"rewards/chosen": 0.38667160272598267,
|
| 2055 |
+
"rewards/margins": 0.1643599271774292,
|
| 2056 |
+
"rewards/rejected": 0.22231166064739227,
|
| 2057 |
+
"step": 1130
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 0.7870889790282213,
|
| 2061 |
+
"grad_norm": 83.31596374511719,
|
| 2062 |
+
"learning_rate": 1.069703243616287e-07,
|
| 2063 |
+
"logits/chosen": -1.4797419309616089,
|
| 2064 |
+
"logits/rejected": -1.30462646484375,
|
| 2065 |
+
"logps/chosen": -376.4138488769531,
|
| 2066 |
+
"logps/rejected": -348.89801025390625,
|
| 2067 |
+
"loss": 0.661,
|
| 2068 |
+
"rewards/accuracies": 0.637499988079071,
|
| 2069 |
+
"rewards/chosen": 0.40063542127609253,
|
| 2070 |
+
"rewards/margins": 0.1296340525150299,
|
| 2071 |
+
"rewards/rejected": 0.27100133895874023,
|
| 2072 |
+
"step": 1140
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.7939932683179425,
|
| 2076 |
+
"grad_norm": 103.56004333496094,
|
| 2077 |
+
"learning_rate": 1.0351966873706004e-07,
|
| 2078 |
+
"logits/chosen": -1.5351804494857788,
|
| 2079 |
+
"logits/rejected": -1.3304837942123413,
|
| 2080 |
+
"logps/chosen": -399.09283447265625,
|
| 2081 |
+
"logps/rejected": -388.9432678222656,
|
| 2082 |
+
"loss": 0.6282,
|
| 2083 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2084 |
+
"rewards/chosen": 0.4186945855617523,
|
| 2085 |
+
"rewards/margins": 0.2270217388868332,
|
| 2086 |
+
"rewards/rejected": 0.19167284667491913,
|
| 2087 |
+
"step": 1150
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 0.7939932683179425,
|
| 2091 |
+
"eval_logits/chosen": NaN,
|
| 2092 |
+
"eval_logits/rejected": -1.1050151586532593,
|
| 2093 |
+
"eval_logps/chosen": -409.507080078125,
|
| 2094 |
+
"eval_logps/rejected": -384.7345886230469,
|
| 2095 |
+
"eval_loss": 0.6440198421478271,
|
| 2096 |
+
"eval_rewards/accuracies": 0.6120907068252563,
|
| 2097 |
+
"eval_rewards/chosen": 0.3957003355026245,
|
| 2098 |
+
"eval_rewards/margins": 0.17536155879497528,
|
| 2099 |
+
"eval_rewards/rejected": 0.22033876180648804,
|
| 2100 |
+
"eval_runtime": 173.6491,
|
| 2101 |
+
"eval_samples_per_second": 6.859,
|
| 2102 |
+
"eval_steps_per_second": 6.859,
|
| 2103 |
+
"step": 1150
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.8008975576076638,
|
| 2107 |
+
"grad_norm": 90.4014663696289,
|
| 2108 |
+
"learning_rate": 1.0006901311249138e-07,
|
| 2109 |
+
"logits/chosen": -1.5077835321426392,
|
| 2110 |
+
"logits/rejected": -1.2734102010726929,
|
| 2111 |
+
"logps/chosen": -379.27410888671875,
|
| 2112 |
+
"logps/rejected": -351.67279052734375,
|
| 2113 |
+
"loss": 0.6381,
|
| 2114 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2115 |
+
"rewards/chosen": 0.38737237453460693,
|
| 2116 |
+
"rewards/margins": 0.174201101064682,
|
| 2117 |
+
"rewards/rejected": 0.21317127346992493,
|
| 2118 |
+
"step": 1160
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.807801846897385,
|
| 2122 |
+
"grad_norm": 66.83627319335938,
|
| 2123 |
+
"learning_rate": 9.66183574879227e-08,
|
| 2124 |
+
"logits/chosen": -1.3416311740875244,
|
| 2125 |
+
"logits/rejected": -1.2346522808074951,
|
| 2126 |
+
"logps/chosen": -397.44805908203125,
|
| 2127 |
+
"logps/rejected": -378.79144287109375,
|
| 2128 |
+
"loss": 0.6569,
|
| 2129 |
+
"rewards/accuracies": 0.612500011920929,
|
| 2130 |
+
"rewards/chosen": 0.3850553333759308,
|
| 2131 |
+
"rewards/margins": 0.15491603314876556,
|
| 2132 |
+
"rewards/rejected": 0.23013930022716522,
|
| 2133 |
+
"step": 1170
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.8147061361871062,
|
| 2137 |
+
"grad_norm": 85.14424896240234,
|
| 2138 |
+
"learning_rate": 9.316770186335403e-08,
|
| 2139 |
+
"logits/chosen": -1.2797658443450928,
|
| 2140 |
+
"logits/rejected": -1.1257115602493286,
|
| 2141 |
+
"logps/chosen": -405.3745422363281,
|
| 2142 |
+
"logps/rejected": -366.5975036621094,
|
| 2143 |
+
"loss": 0.654,
|
| 2144 |
+
"rewards/accuracies": 0.581250011920929,
|
| 2145 |
+
"rewards/chosen": 0.35339251160621643,
|
| 2146 |
+
"rewards/margins": 0.14307786524295807,
|
| 2147 |
+
"rewards/rejected": 0.21031466126441956,
|
| 2148 |
+
"step": 1180
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 0.8216104254768275,
|
| 2152 |
+
"grad_norm": 102.25609588623047,
|
| 2153 |
+
"learning_rate": 8.971704623878537e-08,
|
| 2154 |
+
"logits/chosen": -1.6227967739105225,
|
| 2155 |
+
"logits/rejected": -1.4342666864395142,
|
| 2156 |
+
"logps/chosen": -377.0340881347656,
|
| 2157 |
+
"logps/rejected": -383.2848815917969,
|
| 2158 |
+
"loss": 0.6524,
|
| 2159 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2160 |
+
"rewards/chosen": 0.3992713987827301,
|
| 2161 |
+
"rewards/margins": 0.14228931069374084,
|
| 2162 |
+
"rewards/rejected": 0.25698205828666687,
|
| 2163 |
+
"step": 1190
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.8285147147665487,
|
| 2167 |
+
"grad_norm": 86.08372497558594,
|
| 2168 |
+
"learning_rate": 8.62663906142167e-08,
|
| 2169 |
+
"logits/chosen": -1.4563744068145752,
|
| 2170 |
+
"logits/rejected": -1.352936029434204,
|
| 2171 |
+
"logps/chosen": -397.0251770019531,
|
| 2172 |
+
"logps/rejected": -369.11883544921875,
|
| 2173 |
+
"loss": 0.657,
|
| 2174 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2175 |
+
"rewards/chosen": 0.38390326499938965,
|
| 2176 |
+
"rewards/margins": 0.14529408514499664,
|
| 2177 |
+
"rewards/rejected": 0.2386091649532318,
|
| 2178 |
+
"step": 1200
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 0.8285147147665487,
|
| 2182 |
+
"eval_logits/chosen": NaN,
|
| 2183 |
+
"eval_logits/rejected": -1.105219841003418,
|
| 2184 |
+
"eval_logps/chosen": -409.50103759765625,
|
| 2185 |
+
"eval_logps/rejected": -384.6767883300781,
|
| 2186 |
+
"eval_loss": 0.6475593447685242,
|
| 2187 |
+
"eval_rewards/accuracies": 0.6095718145370483,
|
| 2188 |
+
"eval_rewards/chosen": 0.396301805973053,
|
| 2189 |
+
"eval_rewards/margins": 0.17017905414104462,
|
| 2190 |
+
"eval_rewards/rejected": 0.22612272202968597,
|
| 2191 |
+
"eval_runtime": 173.6541,
|
| 2192 |
+
"eval_samples_per_second": 6.858,
|
| 2193 |
+
"eval_steps_per_second": 6.858,
|
| 2194 |
+
"step": 1200
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 0.83541900405627,
|
| 2198 |
+
"grad_norm": 103.7150650024414,
|
| 2199 |
+
"learning_rate": 8.281573498964802e-08,
|
| 2200 |
+
"logits/chosen": NaN,
|
| 2201 |
+
"logits/rejected": -1.2163610458374023,
|
| 2202 |
+
"logps/chosen": -372.20989990234375,
|
| 2203 |
+
"logps/rejected": -364.71295166015625,
|
| 2204 |
+
"loss": 0.7005,
|
| 2205 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 2206 |
+
"rewards/chosen": 0.3188936710357666,
|
| 2207 |
+
"rewards/margins": 0.05130113288760185,
|
| 2208 |
+
"rewards/rejected": 0.26759251952171326,
|
| 2209 |
+
"step": 1210
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.8423232933459912,
|
| 2213 |
+
"grad_norm": 94.80174255371094,
|
| 2214 |
+
"learning_rate": 7.936507936507936e-08,
|
| 2215 |
+
"logits/chosen": -1.2603299617767334,
|
| 2216 |
+
"logits/rejected": -1.2094240188598633,
|
| 2217 |
+
"logps/chosen": -381.99761962890625,
|
| 2218 |
+
"logps/rejected": -394.0637512207031,
|
| 2219 |
+
"loss": 0.6938,
|
| 2220 |
+
"rewards/accuracies": 0.53125,
|
| 2221 |
+
"rewards/chosen": 0.3386909067630768,
|
| 2222 |
+
"rewards/margins": 0.05538243055343628,
|
| 2223 |
+
"rewards/rejected": 0.2833084464073181,
|
| 2224 |
+
"step": 1220
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 0.8492275826357124,
|
| 2228 |
+
"grad_norm": 96.21173858642578,
|
| 2229 |
+
"learning_rate": 7.59144237405107e-08,
|
| 2230 |
+
"logits/chosen": -1.2958943843841553,
|
| 2231 |
+
"logits/rejected": -1.0766475200653076,
|
| 2232 |
+
"logps/chosen": -389.5054626464844,
|
| 2233 |
+
"logps/rejected": -359.45501708984375,
|
| 2234 |
+
"loss": 0.6571,
|
| 2235 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2236 |
+
"rewards/chosen": 0.37096503376960754,
|
| 2237 |
+
"rewards/margins": 0.14743277430534363,
|
| 2238 |
+
"rewards/rejected": 0.22353224456310272,
|
| 2239 |
+
"step": 1230
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 0.8561318719254337,
|
| 2243 |
+
"grad_norm": 111.18427276611328,
|
| 2244 |
+
"learning_rate": 7.246376811594203e-08,
|
| 2245 |
+
"logits/chosen": -1.4039971828460693,
|
| 2246 |
+
"logits/rejected": NaN,
|
| 2247 |
+
"logps/chosen": -395.5658264160156,
|
| 2248 |
+
"logps/rejected": -356.8921813964844,
|
| 2249 |
+
"loss": 0.6967,
|
| 2250 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 2251 |
+
"rewards/chosen": 0.3431479334831238,
|
| 2252 |
+
"rewards/margins": 0.04705189913511276,
|
| 2253 |
+
"rewards/rejected": 0.2960960268974304,
|
| 2254 |
+
"step": 1240
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 0.8630361612151549,
|
| 2258 |
+
"grad_norm": 85.78289031982422,
|
| 2259 |
+
"learning_rate": 6.901311249137336e-08,
|
| 2260 |
+
"logits/chosen": -1.4105193614959717,
|
| 2261 |
+
"logits/rejected": -1.2953484058380127,
|
| 2262 |
+
"logps/chosen": -370.317626953125,
|
| 2263 |
+
"logps/rejected": -343.9225158691406,
|
| 2264 |
+
"loss": 0.6743,
|
| 2265 |
+
"rewards/accuracies": 0.581250011920929,
|
| 2266 |
+
"rewards/chosen": 0.3631276488304138,
|
| 2267 |
+
"rewards/margins": 0.11619863659143448,
|
| 2268 |
+
"rewards/rejected": 0.24692897498607635,
|
| 2269 |
+
"step": 1250
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 0.8630361612151549,
|
| 2273 |
+
"eval_logits/chosen": NaN,
|
| 2274 |
+
"eval_logits/rejected": -1.1084688901901245,
|
| 2275 |
+
"eval_logps/chosen": -409.56561279296875,
|
| 2276 |
+
"eval_logps/rejected": -384.7563171386719,
|
| 2277 |
+
"eval_loss": 0.6463225483894348,
|
| 2278 |
+
"eval_rewards/accuracies": 0.6146095991134644,
|
| 2279 |
+
"eval_rewards/chosen": 0.3898475766181946,
|
| 2280 |
+
"eval_rewards/margins": 0.17167842388153076,
|
| 2281 |
+
"eval_rewards/rejected": 0.2181691974401474,
|
| 2282 |
+
"eval_runtime": 173.5928,
|
| 2283 |
+
"eval_samples_per_second": 6.861,
|
| 2284 |
+
"eval_steps_per_second": 6.861,
|
| 2285 |
+
"step": 1250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.8699404505048761,
|
| 2289 |
+
"grad_norm": 88.3895492553711,
|
| 2290 |
+
"learning_rate": 6.556245686680468e-08,
|
| 2291 |
+
"logits/chosen": -1.4868072271347046,
|
| 2292 |
+
"logits/rejected": -1.3154561519622803,
|
| 2293 |
+
"logps/chosen": -389.04620361328125,
|
| 2294 |
+
"logps/rejected": -379.40338134765625,
|
| 2295 |
+
"loss": 0.6299,
|
| 2296 |
+
"rewards/accuracies": 0.637499988079071,
|
| 2297 |
+
"rewards/chosen": 0.3888487219810486,
|
| 2298 |
+
"rewards/margins": 0.20886433124542236,
|
| 2299 |
+
"rewards/rejected": 0.1799844205379486,
|
| 2300 |
+
"step": 1260
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.8768447397945974,
|
| 2304 |
+
"grad_norm": 110.88898468017578,
|
| 2305 |
+
"learning_rate": 6.211180124223602e-08,
|
| 2306 |
+
"logits/chosen": -1.2930047512054443,
|
| 2307 |
+
"logits/rejected": NaN,
|
| 2308 |
+
"logps/chosen": -417.73126220703125,
|
| 2309 |
+
"logps/rejected": -387.33563232421875,
|
| 2310 |
+
"loss": 0.6686,
|
| 2311 |
+
"rewards/accuracies": 0.59375,
|
| 2312 |
+
"rewards/chosen": 0.3636248707771301,
|
| 2313 |
+
"rewards/margins": 0.12059865146875381,
|
| 2314 |
+
"rewards/rejected": 0.24302616715431213,
|
| 2315 |
+
"step": 1270
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 0.8837490290843186,
|
| 2319 |
+
"grad_norm": 89.85807800292969,
|
| 2320 |
+
"learning_rate": 5.866114561766736e-08,
|
| 2321 |
+
"logits/chosen": -1.390554428100586,
|
| 2322 |
+
"logits/rejected": -1.328037977218628,
|
| 2323 |
+
"logps/chosen": -413.32244873046875,
|
| 2324 |
+
"logps/rejected": -387.0220947265625,
|
| 2325 |
+
"loss": 0.6563,
|
| 2326 |
+
"rewards/accuracies": 0.574999988079071,
|
| 2327 |
+
"rewards/chosen": 0.41810736060142517,
|
| 2328 |
+
"rewards/margins": 0.14793208241462708,
|
| 2329 |
+
"rewards/rejected": 0.2701753079891205,
|
| 2330 |
+
"step": 1280
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.8906533183740398,
|
| 2334 |
+
"grad_norm": 102.83927917480469,
|
| 2335 |
+
"learning_rate": 5.5210489993098687e-08,
|
| 2336 |
+
"logits/chosen": -1.3396317958831787,
|
| 2337 |
+
"logits/rejected": -1.0842339992523193,
|
| 2338 |
+
"logps/chosen": -372.3958740234375,
|
| 2339 |
+
"logps/rejected": -378.26324462890625,
|
| 2340 |
+
"loss": 0.6655,
|
| 2341 |
+
"rewards/accuracies": 0.612500011920929,
|
| 2342 |
+
"rewards/chosen": 0.367808997631073,
|
| 2343 |
+
"rewards/margins": 0.13876868784427643,
|
| 2344 |
+
"rewards/rejected": 0.22904033958911896,
|
| 2345 |
+
"step": 1290
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 0.8975576076637611,
|
| 2349 |
+
"grad_norm": 108.53854370117188,
|
| 2350 |
+
"learning_rate": 5.175983436853002e-08,
|
| 2351 |
+
"logits/chosen": -1.4352149963378906,
|
| 2352 |
+
"logits/rejected": -1.2435885667800903,
|
| 2353 |
+
"logps/chosen": -431.7513732910156,
|
| 2354 |
+
"logps/rejected": -421.14215087890625,
|
| 2355 |
+
"loss": 0.6502,
|
| 2356 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2357 |
+
"rewards/chosen": 0.40972501039505005,
|
| 2358 |
+
"rewards/margins": 0.16878488659858704,
|
| 2359 |
+
"rewards/rejected": 0.2409401386976242,
|
| 2360 |
+
"step": 1300
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 0.8975576076637611,
|
| 2364 |
+
"eval_logits/chosen": NaN,
|
| 2365 |
+
"eval_logits/rejected": -1.114261269569397,
|
| 2366 |
+
"eval_logps/chosen": -409.6363220214844,
|
| 2367 |
+
"eval_logps/rejected": -384.7745056152344,
|
| 2368 |
+
"eval_loss": 0.6483097076416016,
|
| 2369 |
+
"eval_rewards/accuracies": 0.6120907068252563,
|
| 2370 |
+
"eval_rewards/chosen": 0.3827739357948303,
|
| 2371 |
+
"eval_rewards/margins": 0.16641728579998016,
|
| 2372 |
+
"eval_rewards/rejected": 0.21635663509368896,
|
| 2373 |
+
"eval_runtime": 173.3443,
|
| 2374 |
+
"eval_samples_per_second": 6.871,
|
| 2375 |
+
"eval_steps_per_second": 6.871,
|
| 2376 |
+
"step": 1300
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.9044618969534823,
|
| 2380 |
+
"grad_norm": 91.35237121582031,
|
| 2381 |
+
"learning_rate": 4.830917874396135e-08,
|
| 2382 |
+
"logits/chosen": -1.2966923713684082,
|
| 2383 |
+
"logits/rejected": -1.116370439529419,
|
| 2384 |
+
"logps/chosen": -388.6182556152344,
|
| 2385 |
+
"logps/rejected": -388.0231628417969,
|
| 2386 |
+
"loss": 0.6535,
|
| 2387 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2388 |
+
"rewards/chosen": 0.41747522354125977,
|
| 2389 |
+
"rewards/margins": 0.14344492554664612,
|
| 2390 |
+
"rewards/rejected": 0.27403026819229126,
|
| 2391 |
+
"step": 1310
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.9113661862432035,
|
| 2395 |
+
"grad_norm": 122.95155334472656,
|
| 2396 |
+
"learning_rate": 4.4858523119392684e-08,
|
| 2397 |
+
"logits/chosen": -1.486899971961975,
|
| 2398 |
+
"logits/rejected": -1.4145976305007935,
|
| 2399 |
+
"logps/chosen": -389.6207580566406,
|
| 2400 |
+
"logps/rejected": -373.4767150878906,
|
| 2401 |
+
"loss": 0.6954,
|
| 2402 |
+
"rewards/accuracies": 0.606249988079071,
|
| 2403 |
+
"rewards/chosen": 0.37160056829452515,
|
| 2404 |
+
"rewards/margins": 0.07465855777263641,
|
| 2405 |
+
"rewards/rejected": 0.29694199562072754,
|
| 2406 |
+
"step": 1320
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 0.9182704755329248,
|
| 2410 |
+
"grad_norm": 101.79618072509766,
|
| 2411 |
+
"learning_rate": 4.140786749482401e-08,
|
| 2412 |
+
"logits/chosen": -1.3558008670806885,
|
| 2413 |
+
"logits/rejected": -1.236149549484253,
|
| 2414 |
+
"logps/chosen": -407.06085205078125,
|
| 2415 |
+
"logps/rejected": -349.31304931640625,
|
| 2416 |
+
"loss": 0.6597,
|
| 2417 |
+
"rewards/accuracies": 0.637499988079071,
|
| 2418 |
+
"rewards/chosen": 0.34769049286842346,
|
| 2419 |
+
"rewards/margins": 0.1324794739484787,
|
| 2420 |
+
"rewards/rejected": 0.21521100401878357,
|
| 2421 |
+
"step": 1330
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.925174764822646,
|
| 2425 |
+
"grad_norm": 81.09689331054688,
|
| 2426 |
+
"learning_rate": 3.795721187025535e-08,
|
| 2427 |
+
"logits/chosen": -1.5392885208129883,
|
| 2428 |
+
"logits/rejected": -1.367064118385315,
|
| 2429 |
+
"logps/chosen": -356.09393310546875,
|
| 2430 |
+
"logps/rejected": -319.89520263671875,
|
| 2431 |
+
"loss": 0.6443,
|
| 2432 |
+
"rewards/accuracies": 0.581250011920929,
|
| 2433 |
+
"rewards/chosen": 0.3483354449272156,
|
| 2434 |
+
"rewards/margins": 0.16505715250968933,
|
| 2435 |
+
"rewards/rejected": 0.18327829241752625,
|
| 2436 |
+
"step": 1340
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 0.9320790541123674,
|
| 2440 |
+
"grad_norm": 100.64744567871094,
|
| 2441 |
+
"learning_rate": 3.450655624568668e-08,
|
| 2442 |
+
"logits/chosen": -1.4479949474334717,
|
| 2443 |
+
"logits/rejected": -1.1448233127593994,
|
| 2444 |
+
"logps/chosen": -375.9516296386719,
|
| 2445 |
+
"logps/rejected": -373.54180908203125,
|
| 2446 |
+
"loss": 0.627,
|
| 2447 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 2448 |
+
"rewards/chosen": 0.3978195786476135,
|
| 2449 |
+
"rewards/margins": 0.2024550437927246,
|
| 2450 |
+
"rewards/rejected": 0.1953645646572113,
|
| 2451 |
+
"step": 1350
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 0.9320790541123674,
|
| 2455 |
+
"eval_logits/chosen": NaN,
|
| 2456 |
+
"eval_logits/rejected": -1.1154719591140747,
|
| 2457 |
+
"eval_logps/chosen": -409.63140869140625,
|
| 2458 |
+
"eval_logps/rejected": -384.8327941894531,
|
| 2459 |
+
"eval_loss": 0.6454894542694092,
|
| 2460 |
+
"eval_rewards/accuracies": 0.6104114055633545,
|
| 2461 |
+
"eval_rewards/chosen": 0.3832691013813019,
|
| 2462 |
+
"eval_rewards/margins": 0.1727461814880371,
|
| 2463 |
+
"eval_rewards/rejected": 0.21052293479442596,
|
| 2464 |
+
"eval_runtime": 173.3889,
|
| 2465 |
+
"eval_samples_per_second": 6.869,
|
| 2466 |
+
"eval_steps_per_second": 6.869,
|
| 2467 |
+
"step": 1350
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.9389833434020886,
|
| 2471 |
+
"grad_norm": 94.91573333740234,
|
| 2472 |
+
"learning_rate": 3.105590062111801e-08,
|
| 2473 |
+
"logits/chosen": -1.3383309841156006,
|
| 2474 |
+
"logits/rejected": -1.2590186595916748,
|
| 2475 |
+
"logps/chosen": -361.2745666503906,
|
| 2476 |
+
"logps/rejected": -342.0213623046875,
|
| 2477 |
+
"loss": 0.6599,
|
| 2478 |
+
"rewards/accuracies": 0.625,
|
| 2479 |
+
"rewards/chosen": 0.33325955271720886,
|
| 2480 |
+
"rewards/margins": 0.1604005992412567,
|
| 2481 |
+
"rewards/rejected": 0.17285892367362976,
|
| 2482 |
+
"step": 1360
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.9458876326918098,
|
| 2486 |
+
"grad_norm": 87.90577697753906,
|
| 2487 |
+
"learning_rate": 2.7605244996549343e-08,
|
| 2488 |
+
"logits/chosen": -1.5077590942382812,
|
| 2489 |
+
"logits/rejected": -1.2809985876083374,
|
| 2490 |
+
"logps/chosen": -388.0267333984375,
|
| 2491 |
+
"logps/rejected": -360.5313720703125,
|
| 2492 |
+
"loss": 0.6284,
|
| 2493 |
+
"rewards/accuracies": 0.675000011920929,
|
| 2494 |
+
"rewards/chosen": 0.392368346452713,
|
| 2495 |
+
"rewards/margins": 0.18990589678287506,
|
| 2496 |
+
"rewards/rejected": 0.20246246457099915,
|
| 2497 |
+
"step": 1370
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 0.9527919219815311,
|
| 2501 |
+
"grad_norm": 99.08484649658203,
|
| 2502 |
+
"learning_rate": 2.4154589371980675e-08,
|
| 2503 |
+
"logits/chosen": -1.5224372148513794,
|
| 2504 |
+
"logits/rejected": -1.38528311252594,
|
| 2505 |
+
"logps/chosen": -414.42901611328125,
|
| 2506 |
+
"logps/rejected": -400.0823059082031,
|
| 2507 |
+
"loss": 0.6683,
|
| 2508 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 2509 |
+
"rewards/chosen": 0.3946588933467865,
|
| 2510 |
+
"rewards/margins": 0.11731584370136261,
|
| 2511 |
+
"rewards/rejected": 0.2773430645465851,
|
| 2512 |
+
"step": 1380
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 0.9596962112712523,
|
| 2516 |
+
"grad_norm": 109.33934020996094,
|
| 2517 |
+
"learning_rate": 2.0703933747412006e-08,
|
| 2518 |
+
"logits/chosen": -1.316396951675415,
|
| 2519 |
+
"logits/rejected": -1.102836012840271,
|
| 2520 |
+
"logps/chosen": -387.43194580078125,
|
| 2521 |
+
"logps/rejected": -370.46539306640625,
|
| 2522 |
+
"loss": 0.6383,
|
| 2523 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2524 |
+
"rewards/chosen": 0.37366634607315063,
|
| 2525 |
+
"rewards/margins": 0.18556134402751923,
|
| 2526 |
+
"rewards/rejected": 0.18810497224330902,
|
| 2527 |
+
"step": 1390
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 0.9666005005609735,
|
| 2531 |
+
"grad_norm": 113.74722290039062,
|
| 2532 |
+
"learning_rate": 1.725327812284334e-08,
|
| 2533 |
+
"logits/chosen": -1.435141682624817,
|
| 2534 |
+
"logits/rejected": NaN,
|
| 2535 |
+
"logps/chosen": -382.6386413574219,
|
| 2536 |
+
"logps/rejected": -358.0811767578125,
|
| 2537 |
+
"loss": 0.6845,
|
| 2538 |
+
"rewards/accuracies": 0.574999988079071,
|
| 2539 |
+
"rewards/chosen": 0.303692489862442,
|
| 2540 |
+
"rewards/margins": 0.08099554479122162,
|
| 2541 |
+
"rewards/rejected": 0.2226969450712204,
|
| 2542 |
+
"step": 1400
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 0.9666005005609735,
|
| 2546 |
+
"eval_logits/chosen": NaN,
|
| 2547 |
+
"eval_logits/rejected": -1.1203211545944214,
|
| 2548 |
+
"eval_logps/chosen": -409.6943664550781,
|
| 2549 |
+
"eval_logps/rejected": -384.8702392578125,
|
| 2550 |
+
"eval_loss": 0.6463683247566223,
|
| 2551 |
+
"eval_rewards/accuracies": 0.6070529222488403,
|
| 2552 |
+
"eval_rewards/chosen": 0.3769714832305908,
|
| 2553 |
+
"eval_rewards/margins": 0.17019246518611908,
|
| 2554 |
+
"eval_rewards/rejected": 0.20677906274795532,
|
| 2555 |
+
"eval_runtime": 173.6978,
|
| 2556 |
+
"eval_samples_per_second": 6.857,
|
| 2557 |
+
"eval_steps_per_second": 6.857,
|
| 2558 |
+
"step": 1400
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.9735047898506948,
|
| 2562 |
+
"grad_norm": 75.73210906982422,
|
| 2563 |
+
"learning_rate": 1.3802622498274672e-08,
|
| 2564 |
+
"logits/chosen": -1.4460961818695068,
|
| 2565 |
+
"logits/rejected": -1.2193866968154907,
|
| 2566 |
+
"logps/chosen": -385.35186767578125,
|
| 2567 |
+
"logps/rejected": -381.22784423828125,
|
| 2568 |
+
"loss": 0.6737,
|
| 2569 |
+
"rewards/accuracies": 0.5625,
|
| 2570 |
+
"rewards/chosen": 0.37897366285324097,
|
| 2571 |
+
"rewards/margins": 0.09797494113445282,
|
| 2572 |
+
"rewards/rejected": 0.28099876642227173,
|
| 2573 |
+
"step": 1410
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.980409079140416,
|
| 2577 |
+
"grad_norm": 95.66902923583984,
|
| 2578 |
+
"learning_rate": 1.0351966873706003e-08,
|
| 2579 |
+
"logits/chosen": -1.525224208831787,
|
| 2580 |
+
"logits/rejected": -1.349152684211731,
|
| 2581 |
+
"logps/chosen": -374.0038146972656,
|
| 2582 |
+
"logps/rejected": -349.24755859375,
|
| 2583 |
+
"loss": 0.6468,
|
| 2584 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2585 |
+
"rewards/chosen": 0.37688612937927246,
|
| 2586 |
+
"rewards/margins": 0.1724523901939392,
|
| 2587 |
+
"rewards/rejected": 0.20443375408649445,
|
| 2588 |
+
"step": 1420
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 0.9873133684301372,
|
| 2592 |
+
"grad_norm": 84.95851135253906,
|
| 2593 |
+
"learning_rate": 6.901311249137336e-09,
|
| 2594 |
+
"logits/chosen": -1.6710094213485718,
|
| 2595 |
+
"logits/rejected": -1.3803582191467285,
|
| 2596 |
+
"logps/chosen": -413.56396484375,
|
| 2597 |
+
"logps/rejected": -392.18267822265625,
|
| 2598 |
+
"loss": 0.6433,
|
| 2599 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2600 |
+
"rewards/chosen": 0.3816106915473938,
|
| 2601 |
+
"rewards/margins": 0.18318775296211243,
|
| 2602 |
+
"rewards/rejected": 0.19842293858528137,
|
| 2603 |
+
"step": 1430
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.9942176577198585,
|
| 2607 |
+
"grad_norm": 91.50519561767578,
|
| 2608 |
+
"learning_rate": 3.450655624568668e-09,
|
| 2609 |
+
"logits/chosen": -1.3672393560409546,
|
| 2610 |
+
"logits/rejected": -1.2583658695220947,
|
| 2611 |
+
"logps/chosen": -423.9781799316406,
|
| 2612 |
+
"logps/rejected": -384.3447265625,
|
| 2613 |
+
"loss": 0.6292,
|
| 2614 |
+
"rewards/accuracies": 0.625,
|
| 2615 |
+
"rewards/chosen": 0.40440577268600464,
|
| 2616 |
+
"rewards/margins": 0.20975744724273682,
|
| 2617 |
+
"rewards/rejected": 0.1946483552455902,
|
| 2618 |
+
"step": 1440
|
| 2619 |
+
}
|
| 2620 |
+
],
|
| 2621 |
+
"logging_steps": 10,
|
| 2622 |
+
"max_steps": 1449,
|
| 2623 |
+
"num_input_tokens_seen": 0,
|
| 2624 |
+
"num_train_epochs": 1,
|
| 2625 |
+
"save_steps": 400,
|
| 2626 |
+
"stateful_callbacks": {
|
| 2627 |
+
"TrainerControl": {
|
| 2628 |
+
"args": {
|
| 2629 |
+
"should_epoch_stop": false,
|
| 2630 |
+
"should_evaluate": false,
|
| 2631 |
+
"should_log": false,
|
| 2632 |
+
"should_save": true,
|
| 2633 |
+
"should_training_stop": true
|
| 2634 |
+
},
|
| 2635 |
+
"attributes": {}
|
| 2636 |
+
}
|
| 2637 |
+
},
|
| 2638 |
+
"total_flos": 0.0,
|
| 2639 |
+
"train_batch_size": 1,
|
| 2640 |
+
"trial_name": null,
|
| 2641 |
+
"trial_params": null
|
| 2642 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d5225ac4dcccb2325860244ff0e38ba2840f9cbb684f5f4f6d8441ec98be27a
|
| 3 |
+
size 6136
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|