Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- D_full_emf1_seed22/adapter_config.json +43 -0
- D_full_emf1_seed22/adapter_model.safetensors +3 -0
- D_full_emf1_seed22/chat_template.jinja +54 -0
- D_full_emf1_seed22/tokenizer.json +3 -0
- D_full_emf1_seed22/tokenizer_config.json +29 -0
- D_full_emf1_seed22/training_summary.json +2302 -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 |
+
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
D_full_emf1_seed22/adapter_config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"megatron_config": null,
|
| 23 |
+
"megatron_core": "megatron.core",
|
| 24 |
+
"modules_to_save": null,
|
| 25 |
+
"peft_type": "LORA",
|
| 26 |
+
"peft_version": "0.18.1",
|
| 27 |
+
"qalora_group_size": 16,
|
| 28 |
+
"r": 16,
|
| 29 |
+
"rank_pattern": {},
|
| 30 |
+
"revision": null,
|
| 31 |
+
"target_modules": [
|
| 32 |
+
"k_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"v_proj",
|
| 35 |
+
"q_proj"
|
| 36 |
+
],
|
| 37 |
+
"target_parameters": null,
|
| 38 |
+
"task_type": "CAUSAL_LM",
|
| 39 |
+
"trainable_token_indices": null,
|
| 40 |
+
"use_dora": false,
|
| 41 |
+
"use_qalora": false,
|
| 42 |
+
"use_rslora": false
|
| 43 |
+
}
|
D_full_emf1_seed22/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:305528baa439ac6edb72e7e2297cb0908ab7fb1e7637b8251db36e9ba63b9067
|
| 3 |
+
size 17462432
|
D_full_emf1_seed22/chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
D_full_emf1_seed22/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b4360dd6a184650ffc48056c2569bc603f896c5adfe94b10f1c79f809638aa5
|
| 3 |
+
size 11422166
|
D_full_emf1_seed22/tokenizer_config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": true,
|
| 24 |
+
"model_max_length": 131072,
|
| 25 |
+
"pad_token": "<|endoftext|>",
|
| 26 |
+
"split_special_tokens": false,
|
| 27 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 28 |
+
"unk_token": null
|
| 29 |
+
}
|
D_full_emf1_seed22/training_summary.json
ADDED
|
@@ -0,0 +1,2302 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_name": "Qwen/Qwen2.5-1.5B-Instruct",
|
| 3 |
+
"resolved_model": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 4 |
+
"num_train_records": 10000,
|
| 5 |
+
"num_validation_records_available": 0,
|
| 6 |
+
"training_config": {
|
| 7 |
+
"lora_rank": 16,
|
| 8 |
+
"lora_alpha": 32,
|
| 9 |
+
"learning_rate": 1e-05,
|
| 10 |
+
"num_epochs": 2,
|
| 11 |
+
"per_device_batch_size": 1,
|
| 12 |
+
"per_device_eval_batch_size": 1,
|
| 13 |
+
"gradient_accumulation_steps": 16,
|
| 14 |
+
"beta": 0.1,
|
| 15 |
+
"max_length": 512,
|
| 16 |
+
"max_prompt_length": 384,
|
| 17 |
+
"seed": 22,
|
| 18 |
+
"save_steps": 250,
|
| 19 |
+
"save_total_limit": 2,
|
| 20 |
+
"resume_from_checkpoint": null,
|
| 21 |
+
"precision_dtype": "torch.bfloat16",
|
| 22 |
+
"bf16": true,
|
| 23 |
+
"fp16": false,
|
| 24 |
+
"tokenizer_add_bos_token": false
|
| 25 |
+
},
|
| 26 |
+
"global_step": 1250,
|
| 27 |
+
"best_metric": null,
|
| 28 |
+
"best_model_checkpoint": null,
|
| 29 |
+
"train_metrics": {
|
| 30 |
+
"train_runtime": 5977.8981,
|
| 31 |
+
"train_samples_per_second": 3.346,
|
| 32 |
+
"train_steps_per_second": 0.209,
|
| 33 |
+
"total_flos": 1.3612284112597402e+17,
|
| 34 |
+
"train_loss": 0.45482119541168214
|
| 35 |
+
},
|
| 36 |
+
"log_history": [
|
| 37 |
+
{
|
| 38 |
+
"loss": 0.689018440246582,
|
| 39 |
+
"grad_norm": 0.10905654728412628,
|
| 40 |
+
"learning_rate": 9.928e-06,
|
| 41 |
+
"entropy": 0.8563553406624124,
|
| 42 |
+
"num_tokens": 134743.0,
|
| 43 |
+
"logits/chosen": -0.9788623969289567,
|
| 44 |
+
"logits/rejected": -0.963421240675876,
|
| 45 |
+
"mean_token_accuracy": 0.7335915770381689,
|
| 46 |
+
"rewards/chosen": 0.003118900115805445,
|
| 47 |
+
"rewards/rejected": -0.005566338380720026,
|
| 48 |
+
"rewards/accuracies": 0.53125,
|
| 49 |
+
"rewards/margins": 0.0086852383617952,
|
| 50 |
+
"logps/chosen": -53.44429122209549,
|
| 51 |
+
"logps/rejected": -58.05595805644989,
|
| 52 |
+
"epoch": 0.016,
|
| 53 |
+
"step": 10
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"loss": 0.6828490734100342,
|
| 57 |
+
"grad_norm": 0.11294842511415482,
|
| 58 |
+
"learning_rate": 9.848000000000001e-06,
|
| 59 |
+
"entropy": 0.8075154377147555,
|
| 60 |
+
"num_tokens": 269598.0,
|
| 61 |
+
"logits/chosen": -0.9147015579312413,
|
| 62 |
+
"logits/rejected": -0.9080481267141745,
|
| 63 |
+
"mean_token_accuracy": 0.7330205589532852,
|
| 64 |
+
"rewards/chosen": -0.009720915841876376,
|
| 65 |
+
"rewards/rejected": -0.031031110812728004,
|
| 66 |
+
"rewards/accuracies": 0.675,
|
| 67 |
+
"rewards/margins": 0.02131019511289196,
|
| 68 |
+
"logps/chosen": -55.76358859539032,
|
| 69 |
+
"logps/rejected": -53.03535443544388,
|
| 70 |
+
"epoch": 0.032,
|
| 71 |
+
"step": 20
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"loss": 0.6677823066711426,
|
| 75 |
+
"grad_norm": 0.11828389763832092,
|
| 76 |
+
"learning_rate": 9.768e-06,
|
| 77 |
+
"entropy": 0.8278283050749451,
|
| 78 |
+
"num_tokens": 404333.0,
|
| 79 |
+
"logits/chosen": -1.0811326675861692,
|
| 80 |
+
"logits/rejected": -0.9070872471899184,
|
| 81 |
+
"mean_token_accuracy": 0.7307771431282163,
|
| 82 |
+
"rewards/chosen": -0.03384106803678151,
|
| 83 |
+
"rewards/rejected": -0.08754932412921335,
|
| 84 |
+
"rewards/accuracies": 0.725,
|
| 85 |
+
"rewards/margins": 0.05370825613208581,
|
| 86 |
+
"logps/chosen": -50.51493817567825,
|
| 87 |
+
"logps/rejected": -61.200947272777555,
|
| 88 |
+
"epoch": 0.048,
|
| 89 |
+
"step": 30
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"loss": 0.6462172985076904,
|
| 93 |
+
"grad_norm": 0.11367130279541016,
|
| 94 |
+
"learning_rate": 9.688000000000001e-06,
|
| 95 |
+
"entropy": 0.823192946088966,
|
| 96 |
+
"num_tokens": 538701.0,
|
| 97 |
+
"logits/chosen": -1.0191294171017014,
|
| 98 |
+
"logits/rejected": -0.9379523297738434,
|
| 99 |
+
"mean_token_accuracy": 0.7370131656527519,
|
| 100 |
+
"rewards/chosen": -0.07439912984327748,
|
| 101 |
+
"rewards/rejected": -0.17708719812508206,
|
| 102 |
+
"rewards/accuracies": 0.76875,
|
| 103 |
+
"rewards/margins": 0.10268806891690474,
|
| 104 |
+
"logps/chosen": -50.990779775381085,
|
| 105 |
+
"logps/rejected": -52.61567587852478,
|
| 106 |
+
"epoch": 0.064,
|
| 107 |
+
"step": 40
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"loss": 0.6121170997619629,
|
| 111 |
+
"grad_norm": 0.12841418385505676,
|
| 112 |
+
"learning_rate": 9.608e-06,
|
| 113 |
+
"entropy": 0.8364603637135588,
|
| 114 |
+
"num_tokens": 673524.0,
|
| 115 |
+
"logits/chosen": -0.9610722555180565,
|
| 116 |
+
"logits/rejected": -0.8514297521784684,
|
| 117 |
+
"mean_token_accuracy": 0.7061145938932896,
|
| 118 |
+
"rewards/chosen": -0.14487820940848906,
|
| 119 |
+
"rewards/rejected": -0.332736360195122,
|
| 120 |
+
"rewards/accuracies": 0.78125,
|
| 121 |
+
"rewards/margins": 0.1878581509925425,
|
| 122 |
+
"logps/chosen": -58.89807789325714,
|
| 123 |
+
"logps/rejected": -60.78576712608337,
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"step": 50
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"loss": 0.5916354179382324,
|
| 129 |
+
"grad_norm": 0.11347971856594086,
|
| 130 |
+
"learning_rate": 9.528000000000001e-06,
|
| 131 |
+
"entropy": 0.8045067954575643,
|
| 132 |
+
"num_tokens": 808013.0,
|
| 133 |
+
"logits/chosen": -1.0201351325489214,
|
| 134 |
+
"logits/rejected": -0.8894283035358745,
|
| 135 |
+
"mean_token_accuracy": 0.7225402432493866,
|
| 136 |
+
"rewards/chosen": -0.24984656272717984,
|
| 137 |
+
"rewards/rejected": -0.504135195492563,
|
| 138 |
+
"rewards/accuracies": 0.78125,
|
| 139 |
+
"rewards/margins": 0.25428863214328884,
|
| 140 |
+
"logps/chosen": -52.76062165498733,
|
| 141 |
+
"logps/rejected": -61.86190390586853,
|
| 142 |
+
"epoch": 0.096,
|
| 143 |
+
"step": 60
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"loss": 0.576406192779541,
|
| 147 |
+
"grad_norm": 0.13050724565982819,
|
| 148 |
+
"learning_rate": 9.448e-06,
|
| 149 |
+
"entropy": 0.8006839740090073,
|
| 150 |
+
"num_tokens": 942494.0,
|
| 151 |
+
"logits/chosen": -1.0778371811063052,
|
| 152 |
+
"logits/rejected": -0.9411381606196525,
|
| 153 |
+
"mean_token_accuracy": 0.713543506525457,
|
| 154 |
+
"rewards/chosen": -0.38982224147985106,
|
| 155 |
+
"rewards/rejected": -0.7144045860972256,
|
| 156 |
+
"rewards/accuracies": 0.76875,
|
| 157 |
+
"rewards/margins": 0.32458234429359434,
|
| 158 |
+
"logps/chosen": -60.53339612483978,
|
| 159 |
+
"logps/rejected": -60.80531245470047,
|
| 160 |
+
"epoch": 0.112,
|
| 161 |
+
"step": 70
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"loss": 0.5392527103424072,
|
| 165 |
+
"grad_norm": 0.10742519050836563,
|
| 166 |
+
"learning_rate": 9.368e-06,
|
| 167 |
+
"entropy": 0.8088806970277801,
|
| 168 |
+
"num_tokens": 1076339.0,
|
| 169 |
+
"logits/chosen": -1.1278537035014409,
|
| 170 |
+
"logits/rejected": -0.8774969710621068,
|
| 171 |
+
"mean_token_accuracy": 0.7182694263756275,
|
| 172 |
+
"rewards/chosen": -0.5486542367416405,
|
| 173 |
+
"rewards/rejected": -0.9956708228681237,
|
| 174 |
+
"rewards/accuracies": 0.76875,
|
| 175 |
+
"rewards/margins": 0.4470165837556124,
|
| 176 |
+
"logps/chosen": -52.11761736869812,
|
| 177 |
+
"logps/rejected": -67.0722449183464,
|
| 178 |
+
"epoch": 0.128,
|
| 179 |
+
"step": 80
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"loss": 0.5608308792114258,
|
| 183 |
+
"grad_norm": 0.15674173831939697,
|
| 184 |
+
"learning_rate": 9.288e-06,
|
| 185 |
+
"entropy": 0.768392307573231,
|
| 186 |
+
"num_tokens": 1210662.0,
|
| 187 |
+
"logits/chosen": -1.1788602806606145,
|
| 188 |
+
"logits/rejected": -1.03331966826322,
|
| 189 |
+
"mean_token_accuracy": 0.7145220551639795,
|
| 190 |
+
"rewards/chosen": -0.720596243354521,
|
| 191 |
+
"rewards/rejected": -1.1825086012715473,
|
| 192 |
+
"rewards/accuracies": 0.74375,
|
| 193 |
+
"rewards/margins": 0.46191235817968845,
|
| 194 |
+
"logps/chosen": -60.48848968744278,
|
| 195 |
+
"logps/rejected": -64.15390303134919,
|
| 196 |
+
"epoch": 0.144,
|
| 197 |
+
"step": 90
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"loss": 0.555051851272583,
|
| 201 |
+
"grad_norm": 0.13351617753505707,
|
| 202 |
+
"learning_rate": 9.208e-06,
|
| 203 |
+
"entropy": 0.7862678854435217,
|
| 204 |
+
"num_tokens": 1345120.0,
|
| 205 |
+
"logits/chosen": -1.168248595800457,
|
| 206 |
+
"logits/rejected": -1.0325365206859007,
|
| 207 |
+
"mean_token_accuracy": 0.708121376670897,
|
| 208 |
+
"rewards/chosen": -0.8649011284462176,
|
| 209 |
+
"rewards/rejected": -1.4063857643632218,
|
| 210 |
+
"rewards/accuracies": 0.6875,
|
| 211 |
+
"rewards/margins": 0.5414846379309892,
|
| 212 |
+
"logps/chosen": -61.20340501070022,
|
| 213 |
+
"logps/rejected": -66.58065437078476,
|
| 214 |
+
"epoch": 0.16,
|
| 215 |
+
"step": 100
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"loss": 0.5074045658111572,
|
| 219 |
+
"grad_norm": 0.21781982481479645,
|
| 220 |
+
"learning_rate": 9.128e-06,
|
| 221 |
+
"entropy": 0.7356140260322718,
|
| 222 |
+
"num_tokens": 1479281.0,
|
| 223 |
+
"logits/chosen": -1.3261684939964402,
|
| 224 |
+
"logits/rejected": -1.1306528941057346,
|
| 225 |
+
"mean_token_accuracy": 0.7123186649754644,
|
| 226 |
+
"rewards/chosen": -0.9123384133912623,
|
| 227 |
+
"rewards/rejected": -1.6054315031506121,
|
| 228 |
+
"rewards/accuracies": 0.76875,
|
| 229 |
+
"rewards/margins": 0.6930930894799531,
|
| 230 |
+
"logps/chosen": -59.55553774833679,
|
| 231 |
+
"logps/rejected": -72.37696163654327,
|
| 232 |
+
"epoch": 0.176,
|
| 233 |
+
"step": 110
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"loss": 0.5216456413269043,
|
| 237 |
+
"grad_norm": 0.19241169095039368,
|
| 238 |
+
"learning_rate": 9.048e-06,
|
| 239 |
+
"entropy": 0.7655054951086641,
|
| 240 |
+
"num_tokens": 1613851.0,
|
| 241 |
+
"logits/chosen": -1.3025066533169531,
|
| 242 |
+
"logits/rejected": -1.143896303288123,
|
| 243 |
+
"mean_token_accuracy": 0.7036409357562661,
|
| 244 |
+
"rewards/chosen": -1.0751173802796985,
|
| 245 |
+
"rewards/rejected": -1.776869938801974,
|
| 246 |
+
"rewards/accuracies": 0.75,
|
| 247 |
+
"rewards/margins": 0.7017525571398437,
|
| 248 |
+
"logps/chosen": -66.2069799900055,
|
| 249 |
+
"logps/rejected": -74.56398607492447,
|
| 250 |
+
"epoch": 0.192,
|
| 251 |
+
"step": 120
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"loss": 0.5168225765228271,
|
| 255 |
+
"grad_norm": 0.08994148671627045,
|
| 256 |
+
"learning_rate": 8.968000000000001e-06,
|
| 257 |
+
"entropy": 0.7540512274950742,
|
| 258 |
+
"num_tokens": 1748368.0,
|
| 259 |
+
"logits/chosen": -1.391650497545763,
|
| 260 |
+
"logits/rejected": -1.1460448347696492,
|
| 261 |
+
"mean_token_accuracy": 0.7162771710194648,
|
| 262 |
+
"rewards/chosen": -1.0096850864356384,
|
| 263 |
+
"rewards/rejected": -1.827320344489999,
|
| 264 |
+
"rewards/accuracies": 0.76875,
|
| 265 |
+
"rewards/margins": 0.817635256703943,
|
| 266 |
+
"logps/chosen": -61.82568025588989,
|
| 267 |
+
"logps/rejected": -76.9469052195549,
|
| 268 |
+
"epoch": 0.208,
|
| 269 |
+
"step": 130
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"loss": 0.46869478225708006,
|
| 273 |
+
"grad_norm": 0.13741639256477356,
|
| 274 |
+
"learning_rate": 8.888e-06,
|
| 275 |
+
"entropy": 0.7924248128314503,
|
| 276 |
+
"num_tokens": 1883496.0,
|
| 277 |
+
"logits/chosen": -1.3249019191600873,
|
| 278 |
+
"logits/rejected": -1.1026000772151963,
|
| 279 |
+
"mean_token_accuracy": 0.7087388038635254,
|
| 280 |
+
"rewards/chosen": -1.0496459440822945,
|
| 281 |
+
"rewards/rejected": -2.055653589125723,
|
| 282 |
+
"rewards/accuracies": 0.81875,
|
| 283 |
+
"rewards/margins": 1.0060076432302594,
|
| 284 |
+
"logps/chosen": -64.32736076116562,
|
| 285 |
+
"logps/rejected": -81.53705527782441,
|
| 286 |
+
"epoch": 0.224,
|
| 287 |
+
"step": 140
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"loss": 0.49356565475463865,
|
| 291 |
+
"grad_norm": 0.13059070706367493,
|
| 292 |
+
"learning_rate": 8.808000000000001e-06,
|
| 293 |
+
"entropy": 0.7655621751298896,
|
| 294 |
+
"num_tokens": 2017285.0,
|
| 295 |
+
"logits/chosen": -1.5351193051368675,
|
| 296 |
+
"logits/rejected": -1.235119241285585,
|
| 297 |
+
"mean_token_accuracy": 0.7035089529119432,
|
| 298 |
+
"rewards/chosen": -1.090340746268339,
|
| 299 |
+
"rewards/rejected": -1.9519310850591864,
|
| 300 |
+
"rewards/accuracies": 0.75625,
|
| 301 |
+
"rewards/margins": 0.8615903402678668,
|
| 302 |
+
"logps/chosen": -58.17775413990021,
|
| 303 |
+
"logps/rejected": -74.25504211187362,
|
| 304 |
+
"epoch": 0.24,
|
| 305 |
+
"step": 150
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"loss": 0.5369758605957031,
|
| 309 |
+
"grad_norm": 0.13786712288856506,
|
| 310 |
+
"learning_rate": 8.728e-06,
|
| 311 |
+
"entropy": 0.6862338791484944,
|
| 312 |
+
"num_tokens": 2150045.0,
|
| 313 |
+
"logits/chosen": -1.4912660164800315,
|
| 314 |
+
"logits/rejected": -1.333143902399643,
|
| 315 |
+
"mean_token_accuracy": 0.701057541090995,
|
| 316 |
+
"rewards/chosen": -1.1116661975334865,
|
| 317 |
+
"rewards/rejected": -1.7480126170441508,
|
| 318 |
+
"rewards/accuracies": 0.775,
|
| 319 |
+
"rewards/margins": 0.6363464183639735,
|
| 320 |
+
"logps/chosen": -59.09992996454239,
|
| 321 |
+
"logps/rejected": -60.96899574995041,
|
| 322 |
+
"epoch": 0.256,
|
| 323 |
+
"step": 160
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"loss": 0.5533265113830567,
|
| 327 |
+
"grad_norm": 0.18535780906677246,
|
| 328 |
+
"learning_rate": 8.648000000000001e-06,
|
| 329 |
+
"entropy": 0.7963132435863371,
|
| 330 |
+
"num_tokens": 2284628.0,
|
| 331 |
+
"logits/chosen": -1.4200859032965045,
|
| 332 |
+
"logits/rejected": -1.1475656571055712,
|
| 333 |
+
"mean_token_accuracy": 0.6968366431072355,
|
| 334 |
+
"rewards/chosen": -1.2623908873065375,
|
| 335 |
+
"rewards/rejected": -1.9762427852023392,
|
| 336 |
+
"rewards/accuracies": 0.70625,
|
| 337 |
+
"rewards/margins": 0.7138518976047635,
|
| 338 |
+
"logps/chosen": -65.92677880525589,
|
| 339 |
+
"logps/rejected": -75.31166982650757,
|
| 340 |
+
"epoch": 0.272,
|
| 341 |
+
"step": 170
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"loss": 0.48800249099731446,
|
| 345 |
+
"grad_norm": 0.17086467146873474,
|
| 346 |
+
"learning_rate": 8.568e-06,
|
| 347 |
+
"entropy": 0.7096531906281598,
|
| 348 |
+
"num_tokens": 2418154.0,
|
| 349 |
+
"logits/chosen": -1.4284564529605592,
|
| 350 |
+
"logits/rejected": -1.1960109995497337,
|
| 351 |
+
"mean_token_accuracy": 0.7045122615993022,
|
| 352 |
+
"rewards/chosen": -1.080190175800817,
|
| 353 |
+
"rewards/rejected": -1.9707384748850019,
|
| 354 |
+
"rewards/accuracies": 0.71875,
|
| 355 |
+
"rewards/margins": 0.8905482929199934,
|
| 356 |
+
"logps/chosen": -59.75125551223755,
|
| 357 |
+
"logps/rejected": -70.24758154153824,
|
| 358 |
+
"epoch": 0.288,
|
| 359 |
+
"step": 180
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"loss": 0.46792116165161135,
|
| 363 |
+
"grad_norm": 0.16177140176296234,
|
| 364 |
+
"learning_rate": 8.488e-06,
|
| 365 |
+
"entropy": 0.7515300859929994,
|
| 366 |
+
"num_tokens": 2551816.0,
|
| 367 |
+
"logits/chosen": -1.2414827391616068,
|
| 368 |
+
"logits/rejected": -1.0802036758082083,
|
| 369 |
+
"mean_token_accuracy": 0.7072636014781892,
|
| 370 |
+
"rewards/chosen": -1.0667701240134193,
|
| 371 |
+
"rewards/rejected": -1.9473204111913218,
|
| 372 |
+
"rewards/accuracies": 0.8375,
|
| 373 |
+
"rewards/margins": 0.8805502850795165,
|
| 374 |
+
"logps/chosen": -65.70144573450088,
|
| 375 |
+
"logps/rejected": -68.4914104104042,
|
| 376 |
+
"epoch": 0.304,
|
| 377 |
+
"step": 190
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"loss": 0.5006054878234864,
|
| 381 |
+
"grad_norm": 0.16638650000095367,
|
| 382 |
+
"learning_rate": 8.408e-06,
|
| 383 |
+
"entropy": 0.8327071964275092,
|
| 384 |
+
"num_tokens": 2686525.0,
|
| 385 |
+
"logits/chosen": -1.2503270258894088,
|
| 386 |
+
"logits/rejected": -0.995497852108113,
|
| 387 |
+
"mean_token_accuracy": 0.6908243351615966,
|
| 388 |
+
"rewards/chosen": -1.1853100352571346,
|
| 389 |
+
"rewards/rejected": -2.0631397599121555,
|
| 390 |
+
"rewards/accuracies": 0.76875,
|
| 391 |
+
"rewards/margins": 0.8778297245502472,
|
| 392 |
+
"logps/chosen": -65.22179394960403,
|
| 393 |
+
"logps/rejected": -76.10799474716187,
|
| 394 |
+
"epoch": 0.32,
|
| 395 |
+
"step": 200
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"loss": 0.452646541595459,
|
| 399 |
+
"grad_norm": 0.26452815532684326,
|
| 400 |
+
"learning_rate": 8.328e-06,
|
| 401 |
+
"entropy": 0.8430000780150294,
|
| 402 |
+
"num_tokens": 2821412.0,
|
| 403 |
+
"logits/chosen": -1.1790006368335502,
|
| 404 |
+
"logits/rejected": -0.8779570491578651,
|
| 405 |
+
"mean_token_accuracy": 0.6723773072008044,
|
| 406 |
+
"rewards/chosen": -1.1888155334367183,
|
| 407 |
+
"rewards/rejected": -2.2633378646452913,
|
| 408 |
+
"rewards/accuracies": 0.7375,
|
| 409 |
+
"rewards/margins": 1.074522328004241,
|
| 410 |
+
"logps/chosen": -64.86716014146805,
|
| 411 |
+
"logps/rejected": -79.61585273742676,
|
| 412 |
+
"epoch": 0.336,
|
| 413 |
+
"step": 210
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"loss": 0.5181353569030762,
|
| 417 |
+
"grad_norm": 0.11742082983255386,
|
| 418 |
+
"learning_rate": 8.248e-06,
|
| 419 |
+
"entropy": 0.8045017344535154,
|
| 420 |
+
"num_tokens": 2956627.0,
|
| 421 |
+
"logits/chosen": -1.2154215871610907,
|
| 422 |
+
"logits/rejected": -0.939022998277143,
|
| 423 |
+
"mean_token_accuracy": 0.6996388690546155,
|
| 424 |
+
"rewards/chosen": -1.1623295453668105,
|
| 425 |
+
"rewards/rejected": -2.163627782184631,
|
| 426 |
+
"rewards/accuracies": 0.775,
|
| 427 |
+
"rewards/margins": 1.00129823833704,
|
| 428 |
+
"logps/chosen": -62.80202409029007,
|
| 429 |
+
"logps/rejected": -81.0254952788353,
|
| 430 |
+
"epoch": 0.352,
|
| 431 |
+
"step": 220
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"loss": 0.4743824481964111,
|
| 435 |
+
"grad_norm": 0.16635876893997192,
|
| 436 |
+
"learning_rate": 8.168e-06,
|
| 437 |
+
"entropy": 0.7914588764862855,
|
| 438 |
+
"num_tokens": 3090619.0,
|
| 439 |
+
"logits/chosen": -1.1884228506656345,
|
| 440 |
+
"logits/rejected": -0.9203135491484804,
|
| 441 |
+
"mean_token_accuracy": 0.6981304872781038,
|
| 442 |
+
"rewards/chosen": -1.0808302243589423,
|
| 443 |
+
"rewards/rejected": -2.0085678519913928,
|
| 444 |
+
"rewards/accuracies": 0.78125,
|
| 445 |
+
"rewards/margins": 0.9277376340702176,
|
| 446 |
+
"logps/chosen": -60.4386828660965,
|
| 447 |
+
"logps/rejected": -72.10626875162124,
|
| 448 |
+
"epoch": 0.368,
|
| 449 |
+
"step": 230
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"loss": 0.48973665237426756,
|
| 453 |
+
"grad_norm": 0.1960846185684204,
|
| 454 |
+
"learning_rate": 8.088e-06,
|
| 455 |
+
"entropy": 0.8449207143319655,
|
| 456 |
+
"num_tokens": 3225094.0,
|
| 457 |
+
"logits/chosen": -1.1508734381084937,
|
| 458 |
+
"logits/rejected": -0.9685642621415461,
|
| 459 |
+
"mean_token_accuracy": 0.6950183667242527,
|
| 460 |
+
"rewards/chosen": -1.2113628653900377,
|
| 461 |
+
"rewards/rejected": -2.089532778749708,
|
| 462 |
+
"rewards/accuracies": 0.8,
|
| 463 |
+
"rewards/margins": 0.8781699133105576,
|
| 464 |
+
"logps/chosen": -68.62453348636627,
|
| 465 |
+
"logps/rejected": -75.98600549697876,
|
| 466 |
+
"epoch": 0.384,
|
| 467 |
+
"step": 240
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"loss": 0.5802257061004639,
|
| 471 |
+
"grad_norm": 0.11632820218801498,
|
| 472 |
+
"learning_rate": 8.008e-06,
|
| 473 |
+
"entropy": 0.7761064321442973,
|
| 474 |
+
"num_tokens": 3360182.0,
|
| 475 |
+
"logits/chosen": -1.2831330545177115,
|
| 476 |
+
"logits/rejected": -1.0997321391861137,
|
| 477 |
+
"mean_token_accuracy": 0.6970741396769882,
|
| 478 |
+
"rewards/chosen": -1.1727734387372037,
|
| 479 |
+
"rewards/rejected": -1.8903911848203279,
|
| 480 |
+
"rewards/accuracies": 0.70625,
|
| 481 |
+
"rewards/margins": 0.7176177425310015,
|
| 482 |
+
"logps/chosen": -66.26374975442886,
|
| 483 |
+
"logps/rejected": -72.10199882984162,
|
| 484 |
+
"epoch": 0.4,
|
| 485 |
+
"step": 250
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"loss": 0.43405885696411134,
|
| 489 |
+
"grad_norm": 0.14540572464466095,
|
| 490 |
+
"learning_rate": 7.928e-06,
|
| 491 |
+
"entropy": 0.7629542504204437,
|
| 492 |
+
"num_tokens": 3494544.0,
|
| 493 |
+
"logits/chosen": -1.3185474115706086,
|
| 494 |
+
"logits/rejected": -1.0363903950282343,
|
| 495 |
+
"mean_token_accuracy": 0.69958227686584,
|
| 496 |
+
"rewards/chosen": -0.9348558198704268,
|
| 497 |
+
"rewards/rejected": -1.9087367365602403,
|
| 498 |
+
"rewards/accuracies": 0.8125,
|
| 499 |
+
"rewards/margins": 0.9738809239119292,
|
| 500 |
+
"logps/chosen": -61.2034912109375,
|
| 501 |
+
"logps/rejected": -71.80750354528428,
|
| 502 |
+
"epoch": 0.416,
|
| 503 |
+
"step": 260
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"loss": 0.44985151290893555,
|
| 507 |
+
"grad_norm": 0.14902211725711823,
|
| 508 |
+
"learning_rate": 7.848000000000002e-06,
|
| 509 |
+
"entropy": 0.8110592220822582,
|
| 510 |
+
"num_tokens": 3628021.0,
|
| 511 |
+
"logits/chosen": -1.2687084794247518,
|
| 512 |
+
"logits/rejected": -0.9412073643616663,
|
| 513 |
+
"mean_token_accuracy": 0.6967338371090591,
|
| 514 |
+
"rewards/chosen": -1.0113929210696369,
|
| 515 |
+
"rewards/rejected": -1.9841588545590638,
|
| 516 |
+
"rewards/accuracies": 0.8,
|
| 517 |
+
"rewards/margins": 0.9727659282274544,
|
| 518 |
+
"logps/chosen": -58.82498153448105,
|
| 519 |
+
"logps/rejected": -72.50857379436493,
|
| 520 |
+
"epoch": 0.432,
|
| 521 |
+
"step": 270
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"loss": 0.47422361373901367,
|
| 525 |
+
"grad_norm": 0.14587490260601044,
|
| 526 |
+
"learning_rate": 7.768e-06,
|
| 527 |
+
"entropy": 0.8527051686949563,
|
| 528 |
+
"num_tokens": 3762769.0,
|
| 529 |
+
"logits/chosen": -1.2181775523388327,
|
| 530 |
+
"logits/rejected": -0.8992369459063199,
|
| 531 |
+
"mean_token_accuracy": 0.7088230043649674,
|
| 532 |
+
"rewards/chosen": -1.1207619122928008,
|
| 533 |
+
"rewards/rejected": -2.148672613617964,
|
| 534 |
+
"rewards/accuracies": 0.78125,
|
| 535 |
+
"rewards/margins": 1.0279107017442584,
|
| 536 |
+
"logps/chosen": -61.72244974374771,
|
| 537 |
+
"logps/rejected": -78.88542218208313,
|
| 538 |
+
"epoch": 0.448,
|
| 539 |
+
"step": 280
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"loss": 0.5027151107788086,
|
| 543 |
+
"grad_norm": 0.09337582439184189,
|
| 544 |
+
"learning_rate": 7.688000000000002e-06,
|
| 545 |
+
"entropy": 0.7852856576413615,
|
| 546 |
+
"num_tokens": 3896554.0,
|
| 547 |
+
"logits/chosen": -1.3354775567844146,
|
| 548 |
+
"logits/rejected": -1.1184485921419505,
|
| 549 |
+
"mean_token_accuracy": 0.6952060368843377,
|
| 550 |
+
"rewards/chosen": -1.1932822762464639,
|
| 551 |
+
"rewards/rejected": -2.0500318385427816,
|
| 552 |
+
"rewards/accuracies": 0.75,
|
| 553 |
+
"rewards/margins": 0.8567495625931769,
|
| 554 |
+
"logps/chosen": -66.44852304458618,
|
| 555 |
+
"logps/rejected": -70.7834144949913,
|
| 556 |
+
"epoch": 0.464,
|
| 557 |
+
"step": 290
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"loss": 0.44806575775146484,
|
| 561 |
+
"grad_norm": 0.1337847113609314,
|
| 562 |
+
"learning_rate": 7.608000000000001e-06,
|
| 563 |
+
"entropy": 0.7422546729561873,
|
| 564 |
+
"num_tokens": 4030330.0,
|
| 565 |
+
"logits/chosen": -1.4700220277909535,
|
| 566 |
+
"logits/rejected": -1.1814784580674191,
|
| 567 |
+
"mean_token_accuracy": 0.7079419396817684,
|
| 568 |
+
"rewards/chosen": -1.1185172388635691,
|
| 569 |
+
"rewards/rejected": -2.27528806256596,
|
| 570 |
+
"rewards/accuracies": 0.83125,
|
| 571 |
+
"rewards/margins": 1.1567708211019636,
|
| 572 |
+
"logps/chosen": -63.6253160238266,
|
| 573 |
+
"logps/rejected": -74.7424528837204,
|
| 574 |
+
"epoch": 0.48,
|
| 575 |
+
"step": 300
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"loss": 0.45131826400756836,
|
| 579 |
+
"grad_norm": 0.16172009706497192,
|
| 580 |
+
"learning_rate": 7.528000000000001e-06,
|
| 581 |
+
"entropy": 0.7010916481667664,
|
| 582 |
+
"num_tokens": 4163071.0,
|
| 583 |
+
"logits/chosen": -1.5153129547936302,
|
| 584 |
+
"logits/rejected": -1.209580647062671,
|
| 585 |
+
"mean_token_accuracy": 0.720912242680788,
|
| 586 |
+
"rewards/chosen": -1.1323760432947894,
|
| 587 |
+
"rewards/rejected": -2.123044349125121,
|
| 588 |
+
"rewards/accuracies": 0.825,
|
| 589 |
+
"rewards/margins": 0.9906683029606939,
|
| 590 |
+
"logps/chosen": -61.3173499584198,
|
| 591 |
+
"logps/rejected": -70.07959650754928,
|
| 592 |
+
"epoch": 0.496,
|
| 593 |
+
"step": 310
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"loss": 0.4254595756530762,
|
| 597 |
+
"grad_norm": 0.11975234746932983,
|
| 598 |
+
"learning_rate": 7.4480000000000005e-06,
|
| 599 |
+
"entropy": 0.7769533811253495,
|
| 600 |
+
"num_tokens": 4297650.0,
|
| 601 |
+
"logits/chosen": -1.3886568818841,
|
| 602 |
+
"logits/rejected": -1.13002503495388,
|
| 603 |
+
"mean_token_accuracy": 0.7051799762062728,
|
| 604 |
+
"rewards/chosen": -0.9357022894022521,
|
| 605 |
+
"rewards/rejected": -1.983945999154821,
|
| 606 |
+
"rewards/accuracies": 0.8625,
|
| 607 |
+
"rewards/margins": 1.048243713984266,
|
| 608 |
+
"logps/chosen": -68.33114951848984,
|
| 609 |
+
"logps/rejected": -71.30756684541703,
|
| 610 |
+
"epoch": 0.512,
|
| 611 |
+
"step": 320
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"loss": 0.41445064544677734,
|
| 615 |
+
"grad_norm": 0.2776741087436676,
|
| 616 |
+
"learning_rate": 7.3680000000000004e-06,
|
| 617 |
+
"entropy": 0.7489449862972833,
|
| 618 |
+
"num_tokens": 4431577.0,
|
| 619 |
+
"logits/chosen": -1.6288736615844654,
|
| 620 |
+
"logits/rejected": -1.2485594975833538,
|
| 621 |
+
"mean_token_accuracy": 0.6974780206568539,
|
| 622 |
+
"rewards/chosen": -0.9473370073334081,
|
| 623 |
+
"rewards/rejected": -2.106761427427409,
|
| 624 |
+
"rewards/accuracies": 0.84375,
|
| 625 |
+
"rewards/margins": 1.1594244167208672,
|
| 626 |
+
"logps/chosen": -62.647765040397644,
|
| 627 |
+
"logps/rejected": -75.1746208190918,
|
| 628 |
+
"epoch": 0.528,
|
| 629 |
+
"step": 330
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"loss": 0.4307701587677002,
|
| 633 |
+
"grad_norm": 0.13814318180084229,
|
| 634 |
+
"learning_rate": 7.288e-06,
|
| 635 |
+
"entropy": 0.6795995138236322,
|
| 636 |
+
"num_tokens": 4565362.0,
|
| 637 |
+
"logits/chosen": -1.7014978124379496,
|
| 638 |
+
"logits/rejected": -1.360194546265395,
|
| 639 |
+
"mean_token_accuracy": 0.7122661659494043,
|
| 640 |
+
"rewards/chosen": -1.1425236859824508,
|
| 641 |
+
"rewards/rejected": -2.3793784040724857,
|
| 642 |
+
"rewards/accuracies": 0.8125,
|
| 643 |
+
"rewards/margins": 1.236854719556868,
|
| 644 |
+
"logps/chosen": -61.26923282146454,
|
| 645 |
+
"logps/rejected": -76.55938069820404,
|
| 646 |
+
"epoch": 0.544,
|
| 647 |
+
"step": 340
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"loss": 0.5171660423278809,
|
| 651 |
+
"grad_norm": 0.2421128898859024,
|
| 652 |
+
"learning_rate": 7.208e-06,
|
| 653 |
+
"entropy": 0.736682211542211,
|
| 654 |
+
"num_tokens": 4699278.0,
|
| 655 |
+
"logits/chosen": -1.5375196977562884,
|
| 656 |
+
"logits/rejected": -1.2775026091481378,
|
| 657 |
+
"mean_token_accuracy": 0.7100502958521246,
|
| 658 |
+
"rewards/chosen": -1.234824466597638,
|
| 659 |
+
"rewards/rejected": -2.2794134604977443,
|
| 660 |
+
"rewards/accuracies": 0.78125,
|
| 661 |
+
"rewards/margins": 1.0445889961905777,
|
| 662 |
+
"logps/chosen": -61.170579481124875,
|
| 663 |
+
"logps/rejected": -73.94075057506561,
|
| 664 |
+
"epoch": 0.56,
|
| 665 |
+
"step": 350
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"loss": 0.4048302173614502,
|
| 669 |
+
"grad_norm": 0.13529057800769806,
|
| 670 |
+
"learning_rate": 7.128e-06,
|
| 671 |
+
"entropy": 0.7252036946330918,
|
| 672 |
+
"num_tokens": 4834112.0,
|
| 673 |
+
"logits/chosen": -1.5076449032651227,
|
| 674 |
+
"logits/rejected": -1.3696309019049226,
|
| 675 |
+
"mean_token_accuracy": 0.7079439009539783,
|
| 676 |
+
"rewards/chosen": -1.1786035622935742,
|
| 677 |
+
"rewards/rejected": -2.421615838177968,
|
| 678 |
+
"rewards/accuracies": 0.825,
|
| 679 |
+
"rewards/margins": 1.2430122738704086,
|
| 680 |
+
"logps/chosen": -75.25101128816604,
|
| 681 |
+
"logps/rejected": -73.58783400058746,
|
| 682 |
+
"epoch": 0.576,
|
| 683 |
+
"step": 360
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"loss": 0.4977682113647461,
|
| 687 |
+
"grad_norm": 0.1781209260225296,
|
| 688 |
+
"learning_rate": 7.048e-06,
|
| 689 |
+
"entropy": 0.8224106237757951,
|
| 690 |
+
"num_tokens": 4968874.0,
|
| 691 |
+
"logits/chosen": -1.3960146295474423,
|
| 692 |
+
"logits/rejected": -1.1073445296913083,
|
| 693 |
+
"mean_token_accuracy": 0.708382755331695,
|
| 694 |
+
"rewards/chosen": -1.115643077727873,
|
| 695 |
+
"rewards/rejected": -2.070754229882732,
|
| 696 |
+
"rewards/accuracies": 0.74375,
|
| 697 |
+
"rewards/margins": 0.955111154448241,
|
| 698 |
+
"logps/chosen": -61.408040976524354,
|
| 699 |
+
"logps/rejected": -78.4707190990448,
|
| 700 |
+
"epoch": 0.592,
|
| 701 |
+
"step": 370
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"loss": 0.4448493480682373,
|
| 705 |
+
"grad_norm": 0.1303374469280243,
|
| 706 |
+
"learning_rate": 6.968e-06,
|
| 707 |
+
"entropy": 0.7095136601157719,
|
| 708 |
+
"num_tokens": 5102712.0,
|
| 709 |
+
"logits/chosen": -1.550997072455129,
|
| 710 |
+
"logits/rejected": -1.2904333606869343,
|
| 711 |
+
"mean_token_accuracy": 0.69968516016379,
|
| 712 |
+
"rewards/chosen": -0.9111583859747043,
|
| 713 |
+
"rewards/rejected": -2.052812862768769,
|
| 714 |
+
"rewards/accuracies": 0.80625,
|
| 715 |
+
"rewards/margins": 1.141654483322054,
|
| 716 |
+
"logps/chosen": -59.56483067274094,
|
| 717 |
+
"logps/rejected": -70.73611218929291,
|
| 718 |
+
"epoch": 0.608,
|
| 719 |
+
"step": 380
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"loss": 0.618243932723999,
|
| 723 |
+
"grad_norm": 0.23218761384487152,
|
| 724 |
+
"learning_rate": 6.888e-06,
|
| 725 |
+
"entropy": 0.7683811950089876,
|
| 726 |
+
"num_tokens": 5236746.0,
|
| 727 |
+
"logits/chosen": -1.4908055362391708,
|
| 728 |
+
"logits/rejected": -1.2139118427508149,
|
| 729 |
+
"mean_token_accuracy": 0.7114272927865386,
|
| 730 |
+
"rewards/chosen": -1.168174527026713,
|
| 731 |
+
"rewards/rejected": -1.9601395897567273,
|
| 732 |
+
"rewards/accuracies": 0.71875,
|
| 733 |
+
"rewards/margins": 0.7919650661759079,
|
| 734 |
+
"logps/chosen": -63.611954271793365,
|
| 735 |
+
"logps/rejected": -72.28059177398681,
|
| 736 |
+
"epoch": 0.624,
|
| 737 |
+
"step": 390
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"loss": 0.5224360942840576,
|
| 741 |
+
"grad_norm": 0.16070543229579926,
|
| 742 |
+
"learning_rate": 6.808e-06,
|
| 743 |
+
"entropy": 0.8030928350694012,
|
| 744 |
+
"num_tokens": 5371265.0,
|
| 745 |
+
"logits/chosen": -1.3533366529032598,
|
| 746 |
+
"logits/rejected": -1.0578829997757886,
|
| 747 |
+
"mean_token_accuracy": 0.7191924162209034,
|
| 748 |
+
"rewards/chosen": -0.8509922233264661,
|
| 749 |
+
"rewards/rejected": -1.6920959531096742,
|
| 750 |
+
"rewards/accuracies": 0.7625,
|
| 751 |
+
"rewards/margins": 0.8411037294892594,
|
| 752 |
+
"logps/chosen": -61.14945147037506,
|
| 753 |
+
"logps/rejected": -68.67226406335831,
|
| 754 |
+
"epoch": 0.64,
|
| 755 |
+
"step": 400
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"loss": 0.41513900756835936,
|
| 759 |
+
"grad_norm": 0.19860734045505524,
|
| 760 |
+
"learning_rate": 6.728e-06,
|
| 761 |
+
"entropy": 0.7868119461927563,
|
| 762 |
+
"num_tokens": 5505311.0,
|
| 763 |
+
"logits/chosen": -1.2997119180538206,
|
| 764 |
+
"logits/rejected": -1.0947122079972385,
|
| 765 |
+
"mean_token_accuracy": 0.7141418108716607,
|
| 766 |
+
"rewards/chosen": -0.5277955924670096,
|
| 767 |
+
"rewards/rejected": -1.6158742993546185,
|
| 768 |
+
"rewards/accuracies": 0.85625,
|
| 769 |
+
"rewards/margins": 1.0880787078291179,
|
| 770 |
+
"logps/chosen": -58.874490237236024,
|
| 771 |
+
"logps/rejected": -69.12621879577637,
|
| 772 |
+
"epoch": 0.656,
|
| 773 |
+
"step": 410
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"loss": 0.4617919921875,
|
| 777 |
+
"grad_norm": 0.1614799201488495,
|
| 778 |
+
"learning_rate": 6.648e-06,
|
| 779 |
+
"entropy": 0.8407431220111903,
|
| 780 |
+
"num_tokens": 5639727.0,
|
| 781 |
+
"logits/chosen": -1.2453079399296676,
|
| 782 |
+
"logits/rejected": -0.9694453760606313,
|
| 783 |
+
"mean_token_accuracy": 0.6953371344134212,
|
| 784 |
+
"rewards/chosen": -0.6480406609189231,
|
| 785 |
+
"rewards/rejected": -1.6581713807128835,
|
| 786 |
+
"rewards/accuracies": 0.80625,
|
| 787 |
+
"rewards/margins": 1.0101307133561932,
|
| 788 |
+
"logps/chosen": -58.330216348171234,
|
| 789 |
+
"logps/rejected": -75.3308912873268,
|
| 790 |
+
"epoch": 0.672,
|
| 791 |
+
"step": 420
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"loss": 0.44162707328796386,
|
| 795 |
+
"grad_norm": 0.15507620573043823,
|
| 796 |
+
"learning_rate": 6.568000000000001e-06,
|
| 797 |
+
"entropy": 0.7857385093288031,
|
| 798 |
+
"num_tokens": 5775347.0,
|
| 799 |
+
"logits/chosen": -1.4506047377437405,
|
| 800 |
+
"logits/rejected": -1.1385212073298612,
|
| 801 |
+
"mean_token_accuracy": 0.7148427126929164,
|
| 802 |
+
"rewards/chosen": -0.9052540952921845,
|
| 803 |
+
"rewards/rejected": -1.9813320814282633,
|
| 804 |
+
"rewards/accuracies": 0.8,
|
| 805 |
+
"rewards/margins": 1.0760779848322273,
|
| 806 |
+
"logps/chosen": -70.78476246595383,
|
| 807 |
+
"logps/rejected": -75.07093167304993,
|
| 808 |
+
"epoch": 0.688,
|
| 809 |
+
"step": 430
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"loss": 0.45162124633789064,
|
| 813 |
+
"grad_norm": 0.23057712614536285,
|
| 814 |
+
"learning_rate": 6.488000000000001e-06,
|
| 815 |
+
"entropy": 0.7095841458125506,
|
| 816 |
+
"num_tokens": 5908570.0,
|
| 817 |
+
"logits/chosen": -1.522052635312034,
|
| 818 |
+
"logits/rejected": -1.1914354851620277,
|
| 819 |
+
"mean_token_accuracy": 0.7254255541600287,
|
| 820 |
+
"rewards/chosen": -1.0367633419460618,
|
| 821 |
+
"rewards/rejected": -2.0772957720793785,
|
| 822 |
+
"rewards/accuracies": 0.8,
|
| 823 |
+
"rewards/margins": 1.0405324320308864,
|
| 824 |
+
"logps/chosen": -56.933732664585115,
|
| 825 |
+
"logps/rejected": -71.38762845993043,
|
| 826 |
+
"epoch": 0.704,
|
| 827 |
+
"step": 440
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"loss": 0.5143694877624512,
|
| 831 |
+
"grad_norm": 0.1373089849948883,
|
| 832 |
+
"learning_rate": 6.408000000000001e-06,
|
| 833 |
+
"entropy": 0.7877030725649092,
|
| 834 |
+
"num_tokens": 6042989.0,
|
| 835 |
+
"logits/chosen": -1.3941457694145336,
|
| 836 |
+
"logits/rejected": -1.1642616028929327,
|
| 837 |
+
"mean_token_accuracy": 0.7271975075826049,
|
| 838 |
+
"rewards/chosen": -1.1276259648264386,
|
| 839 |
+
"rewards/rejected": -2.2803055624710398,
|
| 840 |
+
"rewards/accuracies": 0.775,
|
| 841 |
+
"rewards/margins": 1.1526795887388288,
|
| 842 |
+
"logps/chosen": -62.81108283996582,
|
| 843 |
+
"logps/rejected": -78.13692320585251,
|
| 844 |
+
"epoch": 0.72,
|
| 845 |
+
"step": 450
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"loss": 0.4237183094024658,
|
| 849 |
+
"grad_norm": 0.13749825954437256,
|
| 850 |
+
"learning_rate": 6.328000000000001e-06,
|
| 851 |
+
"entropy": 0.7828766799415462,
|
| 852 |
+
"num_tokens": 6177897.0,
|
| 853 |
+
"logits/chosen": -1.3374319642690895,
|
| 854 |
+
"logits/rejected": -1.1246487231005464,
|
| 855 |
+
"mean_token_accuracy": 0.7126569332554936,
|
| 856 |
+
"rewards/chosen": -0.9298112412194314,
|
| 857 |
+
"rewards/rejected": -2.1673050606746984,
|
| 858 |
+
"rewards/accuracies": 0.81875,
|
| 859 |
+
"rewards/margins": 1.2374938178574666,
|
| 860 |
+
"logps/chosen": -63.089643478393555,
|
| 861 |
+
"logps/rejected": -75.59077807664872,
|
| 862 |
+
"epoch": 0.736,
|
| 863 |
+
"step": 460
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"loss": 0.4884624004364014,
|
| 867 |
+
"grad_norm": 0.1415245234966278,
|
| 868 |
+
"learning_rate": 6.248000000000001e-06,
|
| 869 |
+
"entropy": 0.8562492666591425,
|
| 870 |
+
"num_tokens": 6312714.0,
|
| 871 |
+
"logits/chosen": -1.126100791332234,
|
| 872 |
+
"logits/rejected": -0.8647624546845234,
|
| 873 |
+
"mean_token_accuracy": 0.6987247484736144,
|
| 874 |
+
"rewards/chosen": -0.808612387260655,
|
| 875 |
+
"rewards/rejected": -1.6612209561048075,
|
| 876 |
+
"rewards/accuracies": 0.80625,
|
| 877 |
+
"rewards/margins": 0.8526085724122823,
|
| 878 |
+
"logps/chosen": -67.91057009696961,
|
| 879 |
+
"logps/rejected": -70.94221692085266,
|
| 880 |
+
"epoch": 0.752,
|
| 881 |
+
"step": 470
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"loss": 0.44064011573791506,
|
| 885 |
+
"grad_norm": 0.20672255754470825,
|
| 886 |
+
"learning_rate": 6.168000000000001e-06,
|
| 887 |
+
"entropy": 0.7785831509012496,
|
| 888 |
+
"num_tokens": 6446811.0,
|
| 889 |
+
"logits/chosen": -1.3451005530773972,
|
| 890 |
+
"logits/rejected": -1.0701528376820553,
|
| 891 |
+
"mean_token_accuracy": 0.7139140725135803,
|
| 892 |
+
"rewards/chosen": -0.8083824556204491,
|
| 893 |
+
"rewards/rejected": -1.8517554196994752,
|
| 894 |
+
"rewards/accuracies": 0.81875,
|
| 895 |
+
"rewards/margins": 1.0433729680255055,
|
| 896 |
+
"logps/chosen": -58.11244525909424,
|
| 897 |
+
"logps/rejected": -70.4776518702507,
|
| 898 |
+
"epoch": 0.768,
|
| 899 |
+
"step": 480
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"loss": 0.4577657222747803,
|
| 903 |
+
"grad_norm": 0.20929287374019623,
|
| 904 |
+
"learning_rate": 6.088000000000001e-06,
|
| 905 |
+
"entropy": 0.8531491419766098,
|
| 906 |
+
"num_tokens": 6581587.0,
|
| 907 |
+
"logits/chosen": -1.2832854161235994,
|
| 908 |
+
"logits/rejected": -1.1052666904723358,
|
| 909 |
+
"mean_token_accuracy": 0.7067524623125792,
|
| 910 |
+
"rewards/chosen": -0.9194631451624445,
|
| 911 |
+
"rewards/rejected": -1.912933972803876,
|
| 912 |
+
"rewards/accuracies": 0.7875,
|
| 913 |
+
"rewards/margins": 0.9934708263725043,
|
| 914 |
+
"logps/chosen": -62.308107471466066,
|
| 915 |
+
"logps/rejected": -74.89246033430099,
|
| 916 |
+
"epoch": 0.784,
|
| 917 |
+
"step": 490
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"loss": 0.4375007629394531,
|
| 921 |
+
"grad_norm": 0.20060019195079803,
|
| 922 |
+
"learning_rate": 6.008000000000001e-06,
|
| 923 |
+
"entropy": 0.7929941387919826,
|
| 924 |
+
"num_tokens": 6716289.0,
|
| 925 |
+
"logits/chosen": -1.3354546779997054,
|
| 926 |
+
"logits/rejected": -1.0329545787740195,
|
| 927 |
+
"mean_token_accuracy": 0.7229124492034316,
|
| 928 |
+
"rewards/chosen": -1.0229379917611368,
|
| 929 |
+
"rewards/rejected": -2.1298453632509338,
|
| 930 |
+
"rewards/accuracies": 0.8,
|
| 931 |
+
"rewards/margins": 1.1069073686376214,
|
| 932 |
+
"logps/chosen": -63.658102416992186,
|
| 933 |
+
"logps/rejected": -75.60895298719406,
|
| 934 |
+
"epoch": 0.8,
|
| 935 |
+
"step": 500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"loss": 0.44527058601379393,
|
| 939 |
+
"grad_norm": 0.16677559912204742,
|
| 940 |
+
"learning_rate": 5.928000000000001e-06,
|
| 941 |
+
"entropy": 0.8354968667088543,
|
| 942 |
+
"num_tokens": 6850716.0,
|
| 943 |
+
"logits/chosen": -1.4591034667577056,
|
| 944 |
+
"logits/rejected": -1.0641008106016938,
|
| 945 |
+
"mean_token_accuracy": 0.705034868977964,
|
| 946 |
+
"rewards/chosen": -0.8949406793020899,
|
| 947 |
+
"rewards/rejected": -2.185808292112779,
|
| 948 |
+
"rewards/accuracies": 0.81875,
|
| 949 |
+
"rewards/margins": 1.2908676087856292,
|
| 950 |
+
"logps/chosen": -58.768228101730344,
|
| 951 |
+
"logps/rejected": -78.82726471424102,
|
| 952 |
+
"epoch": 0.816,
|
| 953 |
+
"step": 510
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"loss": 0.3973827600479126,
|
| 957 |
+
"grad_norm": 0.15284304320812225,
|
| 958 |
+
"learning_rate": 5.848000000000001e-06,
|
| 959 |
+
"entropy": 0.8099856940709287,
|
| 960 |
+
"num_tokens": 6984942.0,
|
| 961 |
+
"logits/chosen": -1.3746344108751551,
|
| 962 |
+
"logits/rejected": -0.9748087908165729,
|
| 963 |
+
"mean_token_accuracy": 0.7168531034141779,
|
| 964 |
+
"rewards/chosen": -0.8542922898486722,
|
| 965 |
+
"rewards/rejected": -2.102445878717117,
|
| 966 |
+
"rewards/accuracies": 0.85625,
|
| 967 |
+
"rewards/margins": 1.2481535905040801,
|
| 968 |
+
"logps/chosen": -63.956273436546326,
|
| 969 |
+
"logps/rejected": -74.7093752503395,
|
| 970 |
+
"epoch": 0.832,
|
| 971 |
+
"step": 520
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"loss": 0.39521021842956544,
|
| 975 |
+
"grad_norm": 0.19803296029567719,
|
| 976 |
+
"learning_rate": 5.7680000000000005e-06,
|
| 977 |
+
"entropy": 0.819463662133785,
|
| 978 |
+
"num_tokens": 7118390.0,
|
| 979 |
+
"logits/chosen": -1.4278405835062675,
|
| 980 |
+
"logits/rejected": -1.1354102461541689,
|
| 981 |
+
"mean_token_accuracy": 0.7017366614192724,
|
| 982 |
+
"rewards/chosen": -0.8539736664190286,
|
| 983 |
+
"rewards/rejected": -2.189559509139508,
|
| 984 |
+
"rewards/accuracies": 0.81875,
|
| 985 |
+
"rewards/margins": 1.3355858475901186,
|
| 986 |
+
"logps/chosen": -59.81350688934326,
|
| 987 |
+
"logps/rejected": -72.4623057961464,
|
| 988 |
+
"epoch": 0.848,
|
| 989 |
+
"step": 530
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"loss": 0.4910752773284912,
|
| 993 |
+
"grad_norm": 0.17725913226604462,
|
| 994 |
+
"learning_rate": 5.6880000000000004e-06,
|
| 995 |
+
"entropy": 0.8000208166384255,
|
| 996 |
+
"num_tokens": 7252607.0,
|
| 997 |
+
"logits/chosen": -1.4821630663306287,
|
| 998 |
+
"logits/rejected": -1.1388486199317076,
|
| 999 |
+
"mean_token_accuracy": 0.7186640851199627,
|
| 1000 |
+
"rewards/chosen": -1.0058101782691664,
|
| 1001 |
+
"rewards/rejected": -2.2602031591522973,
|
| 1002 |
+
"rewards/accuracies": 0.78125,
|
| 1003 |
+
"rewards/margins": 1.2543929865583778,
|
| 1004 |
+
"logps/chosen": -59.514958798885345,
|
| 1005 |
+
"logps/rejected": -78.18541886806489,
|
| 1006 |
+
"epoch": 0.864,
|
| 1007 |
+
"step": 540
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"loss": 0.4494024276733398,
|
| 1011 |
+
"grad_norm": 0.18910986185073853,
|
| 1012 |
+
"learning_rate": 5.608e-06,
|
| 1013 |
+
"entropy": 0.7143417345014313,
|
| 1014 |
+
"num_tokens": 7386208.0,
|
| 1015 |
+
"logits/chosen": -1.621721678971833,
|
| 1016 |
+
"logits/rejected": -1.4079677960636496,
|
| 1017 |
+
"mean_token_accuracy": 0.6997427928261459,
|
| 1018 |
+
"rewards/chosen": -1.0504559374399833,
|
| 1019 |
+
"rewards/rejected": -2.317046788427979,
|
| 1020 |
+
"rewards/accuracies": 0.7875,
|
| 1021 |
+
"rewards/margins": 1.2665908511728048,
|
| 1022 |
+
"logps/chosen": -59.609275901317595,
|
| 1023 |
+
"logps/rejected": -73.99389593601227,
|
| 1024 |
+
"epoch": 0.88,
|
| 1025 |
+
"step": 550
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"loss": 0.3651487588882446,
|
| 1029 |
+
"grad_norm": 0.14056329429149628,
|
| 1030 |
+
"learning_rate": 5.528e-06,
|
| 1031 |
+
"entropy": 0.814640334653086,
|
| 1032 |
+
"num_tokens": 7521519.0,
|
| 1033 |
+
"logits/chosen": -1.6321557376252653,
|
| 1034 |
+
"logits/rejected": -1.287688344637761,
|
| 1035 |
+
"mean_token_accuracy": 0.7133785426616669,
|
| 1036 |
+
"rewards/chosen": -0.9613554299203543,
|
| 1037 |
+
"rewards/rejected": -2.597287955011416,
|
| 1038 |
+
"rewards/accuracies": 0.83125,
|
| 1039 |
+
"rewards/margins": 1.6359325245954097,
|
| 1040 |
+
"logps/chosen": -64.85036475658417,
|
| 1041 |
+
"logps/rejected": -87.75885683298111,
|
| 1042 |
+
"epoch": 0.896,
|
| 1043 |
+
"step": 560
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"loss": 0.47167210578918456,
|
| 1047 |
+
"grad_norm": 0.15875674784183502,
|
| 1048 |
+
"learning_rate": 5.448e-06,
|
| 1049 |
+
"entropy": 0.8689434929285198,
|
| 1050 |
+
"num_tokens": 7656864.0,
|
| 1051 |
+
"logits/chosen": -1.484541877747767,
|
| 1052 |
+
"logits/rejected": -1.210293212753148,
|
| 1053 |
+
"mean_token_accuracy": 0.6996884623542428,
|
| 1054 |
+
"rewards/chosen": -1.097544879911584,
|
| 1055 |
+
"rewards/rejected": -2.3577645811019465,
|
| 1056 |
+
"rewards/accuracies": 0.78125,
|
| 1057 |
+
"rewards/margins": 1.260219706967473,
|
| 1058 |
+
"logps/chosen": -71.07291498184205,
|
| 1059 |
+
"logps/rejected": -80.84261965751648,
|
| 1060 |
+
"epoch": 0.912,
|
| 1061 |
+
"step": 570
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"loss": 0.4431413173675537,
|
| 1065 |
+
"grad_norm": 0.08832120895385742,
|
| 1066 |
+
"learning_rate": 5.368000000000001e-06,
|
| 1067 |
+
"entropy": 0.8448504954460077,
|
| 1068 |
+
"num_tokens": 7792899.0,
|
| 1069 |
+
"logits/chosen": -1.3894132053730714,
|
| 1070 |
+
"logits/rejected": -1.1916056589936446,
|
| 1071 |
+
"mean_token_accuracy": 0.6817449929192663,
|
| 1072 |
+
"rewards/chosen": -1.1424140176386572,
|
| 1073 |
+
"rewards/rejected": -2.407927218545228,
|
| 1074 |
+
"rewards/accuracies": 0.8375,
|
| 1075 |
+
"rewards/margins": 1.2655131912790238,
|
| 1076 |
+
"logps/chosen": -75.6745502948761,
|
| 1077 |
+
"logps/rejected": -80.21998720169067,
|
| 1078 |
+
"epoch": 0.928,
|
| 1079 |
+
"step": 580
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"loss": 0.4135700225830078,
|
| 1083 |
+
"grad_norm": 0.2704945206642151,
|
| 1084 |
+
"learning_rate": 5.288000000000001e-06,
|
| 1085 |
+
"entropy": 0.7824862065215712,
|
| 1086 |
+
"num_tokens": 7926779.0,
|
| 1087 |
+
"logits/chosen": -1.6528558851348272,
|
| 1088 |
+
"logits/rejected": -1.292456634430272,
|
| 1089 |
+
"mean_token_accuracy": 0.705788871459663,
|
| 1090 |
+
"rewards/chosen": -0.8465097412496106,
|
| 1091 |
+
"rewards/rejected": -2.368203252152307,
|
| 1092 |
+
"rewards/accuracies": 0.825,
|
| 1093 |
+
"rewards/margins": 1.5216935152653606,
|
| 1094 |
+
"logps/chosen": -60.099838173389436,
|
| 1095 |
+
"logps/rejected": -76.38954356908798,
|
| 1096 |
+
"epoch": 0.944,
|
| 1097 |
+
"step": 590
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"loss": 0.44308009147644045,
|
| 1101 |
+
"grad_norm": 0.24440231919288635,
|
| 1102 |
+
"learning_rate": 5.208000000000001e-06,
|
| 1103 |
+
"entropy": 0.6905676309019327,
|
| 1104 |
+
"num_tokens": 8060079.0,
|
| 1105 |
+
"logits/chosen": -1.726859618494725,
|
| 1106 |
+
"logits/rejected": -1.3902765776081965,
|
| 1107 |
+
"mean_token_accuracy": 0.7324171539396047,
|
| 1108 |
+
"rewards/chosen": -0.8706245078108623,
|
| 1109 |
+
"rewards/rejected": -2.1242041391320527,
|
| 1110 |
+
"rewards/accuracies": 0.84375,
|
| 1111 |
+
"rewards/margins": 1.253579631447792,
|
| 1112 |
+
"logps/chosen": -54.19212522506714,
|
| 1113 |
+
"logps/rejected": -68.28248654603958,
|
| 1114 |
+
"epoch": 0.96,
|
| 1115 |
+
"step": 600
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"loss": 0.4900795936584473,
|
| 1119 |
+
"grad_norm": 0.1992669254541397,
|
| 1120 |
+
"learning_rate": 5.128000000000001e-06,
|
| 1121 |
+
"entropy": 0.7946285988775343,
|
| 1122 |
+
"num_tokens": 8194850.0,
|
| 1123 |
+
"logits/chosen": -1.432810063268619,
|
| 1124 |
+
"logits/rejected": -1.1435234217692902,
|
| 1125 |
+
"mean_token_accuracy": 0.7105734800919891,
|
| 1126 |
+
"rewards/chosen": -0.8311952710559126,
|
| 1127 |
+
"rewards/rejected": -1.92433615202317,
|
| 1128 |
+
"rewards/accuracies": 0.80625,
|
| 1129 |
+
"rewards/margins": 1.0931408811360597,
|
| 1130 |
+
"logps/chosen": -59.344405829906464,
|
| 1131 |
+
"logps/rejected": -72.60044294595718,
|
| 1132 |
+
"epoch": 0.976,
|
| 1133 |
+
"step": 610
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"loss": 0.4234670639038086,
|
| 1137 |
+
"grad_norm": 0.26938584446907043,
|
| 1138 |
+
"learning_rate": 5.048000000000001e-06,
|
| 1139 |
+
"entropy": 0.846720263955649,
|
| 1140 |
+
"num_tokens": 8329816.0,
|
| 1141 |
+
"logits/chosen": -1.482459889192245,
|
| 1142 |
+
"logits/rejected": -1.2096144698576996,
|
| 1143 |
+
"mean_token_accuracy": 0.7066002277657389,
|
| 1144 |
+
"rewards/chosen": -0.7947340043901931,
|
| 1145 |
+
"rewards/rejected": -2.0874571030028166,
|
| 1146 |
+
"rewards/accuracies": 0.81875,
|
| 1147 |
+
"rewards/margins": 1.292723097000271,
|
| 1148 |
+
"logps/chosen": -61.47416917085648,
|
| 1149 |
+
"logps/rejected": -79.54253326654434,
|
| 1150 |
+
"epoch": 0.992,
|
| 1151 |
+
"step": 620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"loss": 0.42821288108825684,
|
| 1155 |
+
"grad_norm": 0.20475246012210846,
|
| 1156 |
+
"learning_rate": 4.9680000000000005e-06,
|
| 1157 |
+
"entropy": 0.8765991456399206,
|
| 1158 |
+
"num_tokens": 8464146.0,
|
| 1159 |
+
"logits/chosen": -1.3660770117134349,
|
| 1160 |
+
"logits/rejected": -1.1283678471483534,
|
| 1161 |
+
"mean_token_accuracy": 0.6944115529768169,
|
| 1162 |
+
"rewards/chosen": -0.758866243439843,
|
| 1163 |
+
"rewards/rejected": -1.9606408149003982,
|
| 1164 |
+
"rewards/accuracies": 0.8,
|
| 1165 |
+
"rewards/margins": 1.2017745693214237,
|
| 1166 |
+
"logps/chosen": -59.6711128950119,
|
| 1167 |
+
"logps/rejected": -75.17260210514068,
|
| 1168 |
+
"epoch": 1.008,
|
| 1169 |
+
"step": 630
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"loss": 0.4121402740478516,
|
| 1173 |
+
"grad_norm": 0.19181285798549652,
|
| 1174 |
+
"learning_rate": 4.8880000000000005e-06,
|
| 1175 |
+
"entropy": 0.8112090851645917,
|
| 1176 |
+
"num_tokens": 8598165.0,
|
| 1177 |
+
"logits/chosen": -1.6396478070003788,
|
| 1178 |
+
"logits/rejected": -1.201064509639671,
|
| 1179 |
+
"mean_token_accuracy": 0.7071607926860451,
|
| 1180 |
+
"rewards/chosen": -0.8190987646230496,
|
| 1181 |
+
"rewards/rejected": -2.1065064918162535,
|
| 1182 |
+
"rewards/accuracies": 0.85,
|
| 1183 |
+
"rewards/margins": 1.2874077258165926,
|
| 1184 |
+
"logps/chosen": -58.56870514154434,
|
| 1185 |
+
"logps/rejected": -75.47631640434265,
|
| 1186 |
+
"epoch": 1.024,
|
| 1187 |
+
"step": 640
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"loss": 0.3832432746887207,
|
| 1191 |
+
"grad_norm": 0.16693222522735596,
|
| 1192 |
+
"learning_rate": 4.808e-06,
|
| 1193 |
+
"entropy": 0.8202884538564831,
|
| 1194 |
+
"num_tokens": 8733182.0,
|
| 1195 |
+
"logits/chosen": -1.4093097295310522,
|
| 1196 |
+
"logits/rejected": -1.0764184005015662,
|
| 1197 |
+
"mean_token_accuracy": 0.7213250549510122,
|
| 1198 |
+
"rewards/chosen": -0.8947586458227306,
|
| 1199 |
+
"rewards/rejected": -2.217220963910222,
|
| 1200 |
+
"rewards/accuracies": 0.85,
|
| 1201 |
+
"rewards/margins": 1.3224623166024685,
|
| 1202 |
+
"logps/chosen": -63.64674232006073,
|
| 1203 |
+
"logps/rejected": -79.54431369304658,
|
| 1204 |
+
"epoch": 1.04,
|
| 1205 |
+
"step": 650
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"loss": 0.467090368270874,
|
| 1209 |
+
"grad_norm": 0.2820304334163666,
|
| 1210 |
+
"learning_rate": 4.728e-06,
|
| 1211 |
+
"entropy": 0.8175160668091849,
|
| 1212 |
+
"num_tokens": 8866532.0,
|
| 1213 |
+
"logits/chosen": -1.4113409272988666,
|
| 1214 |
+
"logits/rejected": -1.0253050648283575,
|
| 1215 |
+
"mean_token_accuracy": 0.7056375283747911,
|
| 1216 |
+
"rewards/chosen": -1.1019376051408472,
|
| 1217 |
+
"rewards/rejected": -2.4279107422335073,
|
| 1218 |
+
"rewards/accuracies": 0.83125,
|
| 1219 |
+
"rewards/margins": 1.3259731331840157,
|
| 1220 |
+
"logps/chosen": -62.17448818683624,
|
| 1221 |
+
"logps/rejected": -77.10730903148651,
|
| 1222 |
+
"epoch": 1.056,
|
| 1223 |
+
"step": 660
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"loss": 0.4856276512145996,
|
| 1227 |
+
"grad_norm": 0.1508263647556305,
|
| 1228 |
+
"learning_rate": 4.648e-06,
|
| 1229 |
+
"entropy": 0.8777147593849804,
|
| 1230 |
+
"num_tokens": 9002058.0,
|
| 1231 |
+
"logits/chosen": -1.237333422062607,
|
| 1232 |
+
"logits/rejected": -0.9867713449001874,
|
| 1233 |
+
"mean_token_accuracy": 0.6980989784002304,
|
| 1234 |
+
"rewards/chosen": -1.102817886834964,
|
| 1235 |
+
"rewards/rejected": -2.2164312085602433,
|
| 1236 |
+
"rewards/accuracies": 0.7875,
|
| 1237 |
+
"rewards/margins": 1.113613315252587,
|
| 1238 |
+
"logps/chosen": -72.5415987610817,
|
| 1239 |
+
"logps/rejected": -79.59271297454833,
|
| 1240 |
+
"epoch": 1.072,
|
| 1241 |
+
"step": 670
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"loss": 0.40852723121643064,
|
| 1245 |
+
"grad_norm": 0.1356440633535385,
|
| 1246 |
+
"learning_rate": 4.568e-06,
|
| 1247 |
+
"entropy": 0.8335771631682292,
|
| 1248 |
+
"num_tokens": 9136559.0,
|
| 1249 |
+
"logits/chosen": -1.4357955653994279,
|
| 1250 |
+
"logits/rejected": -1.0632368045892524,
|
| 1251 |
+
"mean_token_accuracy": 0.7067163791507483,
|
| 1252 |
+
"rewards/chosen": -0.8871819378109649,
|
| 1253 |
+
"rewards/rejected": -2.2391563730299824,
|
| 1254 |
+
"rewards/accuracies": 0.8,
|
| 1255 |
+
"rewards/margins": 1.3519744293764233,
|
| 1256 |
+
"logps/chosen": -65.51536420583724,
|
| 1257 |
+
"logps/rejected": -74.92358832359314,
|
| 1258 |
+
"epoch": 1.088,
|
| 1259 |
+
"step": 680
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"loss": 0.42632465362548827,
|
| 1263 |
+
"grad_norm": 0.2446223795413971,
|
| 1264 |
+
"learning_rate": 4.488e-06,
|
| 1265 |
+
"entropy": 0.7106906755536329,
|
| 1266 |
+
"num_tokens": 9269437.0,
|
| 1267 |
+
"logits/chosen": -1.7027034766334481,
|
| 1268 |
+
"logits/rejected": -1.3692697426790565,
|
| 1269 |
+
"mean_token_accuracy": 0.7298609726130962,
|
| 1270 |
+
"rewards/chosen": -0.9662762339459732,
|
| 1271 |
+
"rewards/rejected": -2.2450467050541194,
|
| 1272 |
+
"rewards/accuracies": 0.825,
|
| 1273 |
+
"rewards/margins": 1.2787704736925662,
|
| 1274 |
+
"logps/chosen": -57.37728985548019,
|
| 1275 |
+
"logps/rejected": -68.59222289323807,
|
| 1276 |
+
"epoch": 1.104,
|
| 1277 |
+
"step": 690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"loss": 0.35138611793518065,
|
| 1281 |
+
"grad_norm": 0.12836965918540955,
|
| 1282 |
+
"learning_rate": 4.408000000000001e-06,
|
| 1283 |
+
"entropy": 0.8666601853678003,
|
| 1284 |
+
"num_tokens": 9404559.0,
|
| 1285 |
+
"logits/chosen": -1.3400690205304637,
|
| 1286 |
+
"logits/rejected": -1.0684046347991027,
|
| 1287 |
+
"mean_token_accuracy": 0.6958472795784474,
|
| 1288 |
+
"rewards/chosen": -0.7967793260220788,
|
| 1289 |
+
"rewards/rejected": -2.3613458889187315,
|
| 1290 |
+
"rewards/accuracies": 0.84375,
|
| 1291 |
+
"rewards/margins": 1.5645665610209107,
|
| 1292 |
+
"logps/chosen": -66.0351133108139,
|
| 1293 |
+
"logps/rejected": -81.51442708969117,
|
| 1294 |
+
"epoch": 1.12,
|
| 1295 |
+
"step": 700
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"loss": 0.4332468032836914,
|
| 1299 |
+
"grad_norm": 0.20934900641441345,
|
| 1300 |
+
"learning_rate": 4.328000000000001e-06,
|
| 1301 |
+
"entropy": 0.854771285172319,
|
| 1302 |
+
"num_tokens": 9539800.0,
|
| 1303 |
+
"logits/chosen": -1.3352951022130157,
|
| 1304 |
+
"logits/rejected": -0.9464862926160796,
|
| 1305 |
+
"mean_token_accuracy": 0.6998173886910081,
|
| 1306 |
+
"rewards/chosen": -0.8727404756122269,
|
| 1307 |
+
"rewards/rejected": -2.1966246593976395,
|
| 1308 |
+
"rewards/accuracies": 0.85,
|
| 1309 |
+
"rewards/margins": 1.3238841853104533,
|
| 1310 |
+
"logps/chosen": -61.889275419712064,
|
| 1311 |
+
"logps/rejected": -78.38478618860245,
|
| 1312 |
+
"epoch": 1.1360000000000001,
|
| 1313 |
+
"step": 710
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"loss": 0.4375255584716797,
|
| 1317 |
+
"grad_norm": 0.14867250621318817,
|
| 1318 |
+
"learning_rate": 4.248000000000001e-06,
|
| 1319 |
+
"entropy": 0.7838106972893002,
|
| 1320 |
+
"num_tokens": 9673582.0,
|
| 1321 |
+
"logits/chosen": -1.6430084008837227,
|
| 1322 |
+
"logits/rejected": -1.178061142015777,
|
| 1323 |
+
"mean_token_accuracy": 0.7228047780692577,
|
| 1324 |
+
"rewards/chosen": -0.8893449913855875,
|
| 1325 |
+
"rewards/rejected": -2.1319361824076624,
|
| 1326 |
+
"rewards/accuracies": 0.78125,
|
| 1327 |
+
"rewards/margins": 1.2425911880563945,
|
| 1328 |
+
"logps/chosen": -53.7250174164772,
|
| 1329 |
+
"logps/rejected": -75.55973731279373,
|
| 1330 |
+
"epoch": 1.152,
|
| 1331 |
+
"step": 720
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"loss": 0.4083255767822266,
|
| 1335 |
+
"grad_norm": 0.15854252874851227,
|
| 1336 |
+
"learning_rate": 4.168000000000001e-06,
|
| 1337 |
+
"entropy": 0.8028846303815953,
|
| 1338 |
+
"num_tokens": 9807723.0,
|
| 1339 |
+
"logits/chosen": -1.5796788729741613,
|
| 1340 |
+
"logits/rejected": -1.2279314483195203,
|
| 1341 |
+
"mean_token_accuracy": 0.7037102231755853,
|
| 1342 |
+
"rewards/chosen": -0.8733749952283688,
|
| 1343 |
+
"rewards/rejected": -2.346686244173907,
|
| 1344 |
+
"rewards/accuracies": 0.79375,
|
| 1345 |
+
"rewards/margins": 1.47331124804914,
|
| 1346 |
+
"logps/chosen": -57.2558536529541,
|
| 1347 |
+
"logps/rejected": -78.73293989896774,
|
| 1348 |
+
"epoch": 1.168,
|
| 1349 |
+
"step": 730
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"loss": 0.40847029685974123,
|
| 1353 |
+
"grad_norm": 0.2271832525730133,
|
| 1354 |
+
"learning_rate": 4.0880000000000005e-06,
|
| 1355 |
+
"entropy": 0.7124190992151853,
|
| 1356 |
+
"num_tokens": 9940521.0,
|
| 1357 |
+
"logits/chosen": -1.7263337940860637,
|
| 1358 |
+
"logits/rejected": -1.2993630722529041,
|
| 1359 |
+
"mean_token_accuracy": 0.7212265940383077,
|
| 1360 |
+
"rewards/chosen": -0.749589913454838,
|
| 1361 |
+
"rewards/rejected": -2.0937022533995333,
|
| 1362 |
+
"rewards/accuracies": 0.81875,
|
| 1363 |
+
"rewards/margins": 1.3441123379394413,
|
| 1364 |
+
"logps/chosen": -54.285783529281616,
|
| 1365 |
+
"logps/rejected": -67.44227703809739,
|
| 1366 |
+
"epoch": 1.184,
|
| 1367 |
+
"step": 740
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"loss": 0.4263965129852295,
|
| 1371 |
+
"grad_norm": 0.16508913040161133,
|
| 1372 |
+
"learning_rate": 4.008e-06,
|
| 1373 |
+
"entropy": 0.8771127343410626,
|
| 1374 |
+
"num_tokens": 10075349.0,
|
| 1375 |
+
"logits/chosen": -1.3213851460192345,
|
| 1376 |
+
"logits/rejected": -1.057589043568979,
|
| 1377 |
+
"mean_token_accuracy": 0.6872733032330871,
|
| 1378 |
+
"rewards/chosen": -0.9163320093342918,
|
| 1379 |
+
"rewards/rejected": -2.138802810528432,
|
| 1380 |
+
"rewards/accuracies": 0.80625,
|
| 1381 |
+
"rewards/margins": 1.2224708050955087,
|
| 1382 |
+
"logps/chosen": -65.35894116163254,
|
| 1383 |
+
"logps/rejected": -74.31111829280853,
|
| 1384 |
+
"epoch": 1.2,
|
| 1385 |
+
"step": 750
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"loss": 0.3466779708862305,
|
| 1389 |
+
"grad_norm": 0.14594385027885437,
|
| 1390 |
+
"learning_rate": 3.928e-06,
|
| 1391 |
+
"entropy": 0.8973389053368009,
|
| 1392 |
+
"num_tokens": 10210309.0,
|
| 1393 |
+
"logits/chosen": -1.429195333570735,
|
| 1394 |
+
"logits/rejected": -1.0293746532892192,
|
| 1395 |
+
"mean_token_accuracy": 0.6843982387334109,
|
| 1396 |
+
"rewards/chosen": -0.790108532004524,
|
| 1397 |
+
"rewards/rejected": -2.4523484482313505,
|
| 1398 |
+
"rewards/accuracies": 0.88125,
|
| 1399 |
+
"rewards/margins": 1.6622399202547968,
|
| 1400 |
+
"logps/chosen": -64.03692282438278,
|
| 1401 |
+
"logps/rejected": -83.50145937204361,
|
| 1402 |
+
"epoch": 1.216,
|
| 1403 |
+
"step": 760
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"loss": 0.3902191400527954,
|
| 1407 |
+
"grad_norm": 0.21555303037166595,
|
| 1408 |
+
"learning_rate": 3.848e-06,
|
| 1409 |
+
"entropy": 0.8176256978302263,
|
| 1410 |
+
"num_tokens": 10344571.0,
|
| 1411 |
+
"logits/chosen": -1.571296518421704,
|
| 1412 |
+
"logits/rejected": -1.1463247268095662,
|
| 1413 |
+
"mean_token_accuracy": 0.7237087743356824,
|
| 1414 |
+
"rewards/chosen": -0.9857207721681334,
|
| 1415 |
+
"rewards/rejected": -2.467484907130711,
|
| 1416 |
+
"rewards/accuracies": 0.85,
|
| 1417 |
+
"rewards/margins": 1.4817641287110745,
|
| 1418 |
+
"logps/chosen": -55.41268019676208,
|
| 1419 |
+
"logps/rejected": -83.52744359970093,
|
| 1420 |
+
"epoch": 1.232,
|
| 1421 |
+
"step": 770
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"loss": 0.4406851291656494,
|
| 1425 |
+
"grad_norm": 0.21594573557376862,
|
| 1426 |
+
"learning_rate": 3.7680000000000006e-06,
|
| 1427 |
+
"entropy": 0.7303940998070175,
|
| 1428 |
+
"num_tokens": 10478286.0,
|
| 1429 |
+
"logits/chosen": -1.8663801540109504,
|
| 1430 |
+
"logits/rejected": -1.4280047600468566,
|
| 1431 |
+
"mean_token_accuracy": 0.7173186991363764,
|
| 1432 |
+
"rewards/chosen": -1.1608861902008356,
|
| 1433 |
+
"rewards/rejected": -2.671035580892203,
|
| 1434 |
+
"rewards/accuracies": 0.79375,
|
| 1435 |
+
"rewards/margins": 1.51014938436565,
|
| 1436 |
+
"logps/chosen": -60.46502422094345,
|
| 1437 |
+
"logps/rejected": -77.58741216659546,
|
| 1438 |
+
"epoch": 1.248,
|
| 1439 |
+
"step": 780
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"loss": 0.4125234127044678,
|
| 1443 |
+
"grad_norm": 0.2962269186973572,
|
| 1444 |
+
"learning_rate": 3.6880000000000005e-06,
|
| 1445 |
+
"entropy": 0.8331925821723416,
|
| 1446 |
+
"num_tokens": 10612286.0,
|
| 1447 |
+
"logits/chosen": -1.5659800443469807,
|
| 1448 |
+
"logits/rejected": -1.2150587345047403,
|
| 1449 |
+
"mean_token_accuracy": 0.7073991242796183,
|
| 1450 |
+
"rewards/chosen": -1.1026862843311391,
|
| 1451 |
+
"rewards/rejected": -2.5605769436224364,
|
| 1452 |
+
"rewards/accuracies": 0.78125,
|
| 1453 |
+
"rewards/margins": 1.4578906406648457,
|
| 1454 |
+
"logps/chosen": -62.54402623176575,
|
| 1455 |
+
"logps/rejected": -79.81341508626937,
|
| 1456 |
+
"epoch": 1.264,
|
| 1457 |
+
"step": 790
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"loss": 0.49402365684509275,
|
| 1461 |
+
"grad_norm": 0.155300110578537,
|
| 1462 |
+
"learning_rate": 3.6080000000000004e-06,
|
| 1463 |
+
"entropy": 0.7988151058845687,
|
| 1464 |
+
"num_tokens": 10747462.0,
|
| 1465 |
+
"logits/chosen": -1.605360317746176,
|
| 1466 |
+
"logits/rejected": -1.2881572925592417,
|
| 1467 |
+
"mean_token_accuracy": 0.7002569252625108,
|
| 1468 |
+
"rewards/chosen": -1.0135203968995483,
|
| 1469 |
+
"rewards/rejected": -2.2969876725808716,
|
| 1470 |
+
"rewards/accuracies": 0.79375,
|
| 1471 |
+
"rewards/margins": 1.2834672711789608,
|
| 1472 |
+
"logps/chosen": -67.37470293045044,
|
| 1473 |
+
"logps/rejected": -79.85493696928025,
|
| 1474 |
+
"epoch": 1.28,
|
| 1475 |
+
"step": 800
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"loss": 0.4200831413269043,
|
| 1479 |
+
"grad_norm": 0.14519786834716797,
|
| 1480 |
+
"learning_rate": 3.5280000000000004e-06,
|
| 1481 |
+
"entropy": 0.8120490956622234,
|
| 1482 |
+
"num_tokens": 10882594.0,
|
| 1483 |
+
"logits/chosen": -1.5574423493502165,
|
| 1484 |
+
"logits/rejected": -1.3764162162150948,
|
| 1485 |
+
"mean_token_accuracy": 0.702505898848176,
|
| 1486 |
+
"rewards/chosen": -0.979388279729028,
|
| 1487 |
+
"rewards/rejected": -2.461002457328141,
|
| 1488 |
+
"rewards/accuracies": 0.8,
|
| 1489 |
+
"rewards/margins": 1.4816141836112364,
|
| 1490 |
+
"logps/chosen": -66.68434045314788,
|
| 1491 |
+
"logps/rejected": -83.09649764299392,
|
| 1492 |
+
"epoch": 1.296,
|
| 1493 |
+
"step": 810
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"loss": 0.4704907894134521,
|
| 1497 |
+
"grad_norm": 0.22803907096385956,
|
| 1498 |
+
"learning_rate": 3.4480000000000003e-06,
|
| 1499 |
+
"entropy": 0.8039654018590227,
|
| 1500 |
+
"num_tokens": 11016931.0,
|
| 1501 |
+
"logits/chosen": -1.5012306281869234,
|
| 1502 |
+
"logits/rejected": -1.2876392053232184,
|
| 1503 |
+
"mean_token_accuracy": 0.7034664511680603,
|
| 1504 |
+
"rewards/chosen": -1.1309354332945076,
|
| 1505 |
+
"rewards/rejected": -2.3758349375304535,
|
| 1506 |
+
"rewards/accuracies": 0.76875,
|
| 1507 |
+
"rewards/margins": 1.2448995084967465,
|
| 1508 |
+
"logps/chosen": -67.47772282361984,
|
| 1509 |
+
"logps/rejected": -78.3834860920906,
|
| 1510 |
+
"epoch": 1.312,
|
| 1511 |
+
"step": 820
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"loss": 0.46920132637023926,
|
| 1515 |
+
"grad_norm": 0.2387949824333191,
|
| 1516 |
+
"learning_rate": 3.368e-06,
|
| 1517 |
+
"entropy": 0.732459101951099,
|
| 1518 |
+
"num_tokens": 11150609.0,
|
| 1519 |
+
"logits/chosen": -1.7111103661903264,
|
| 1520 |
+
"logits/rejected": -1.3267267526306643,
|
| 1521 |
+
"mean_token_accuracy": 0.7098248641937971,
|
| 1522 |
+
"rewards/chosen": -0.9866411694383714,
|
| 1523 |
+
"rewards/rejected": -2.3024006365798413,
|
| 1524 |
+
"rewards/accuracies": 0.8,
|
| 1525 |
+
"rewards/margins": 1.3157594701158815,
|
| 1526 |
+
"logps/chosen": -60.04266767501831,
|
| 1527 |
+
"logps/rejected": -73.96700737476348,
|
| 1528 |
+
"epoch": 1.328,
|
| 1529 |
+
"step": 830
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"loss": 0.3219926357269287,
|
| 1533 |
+
"grad_norm": 0.13847558200359344,
|
| 1534 |
+
"learning_rate": 3.288e-06,
|
| 1535 |
+
"entropy": 0.7705854054656811,
|
| 1536 |
+
"num_tokens": 11284845.0,
|
| 1537 |
+
"logits/chosen": -1.7253091998162051,
|
| 1538 |
+
"logits/rejected": -1.305342065365432,
|
| 1539 |
+
"mean_token_accuracy": 0.7240345129743219,
|
| 1540 |
+
"rewards/chosen": -0.9102539952356892,
|
| 1541 |
+
"rewards/rejected": -2.657541433814913,
|
| 1542 |
+
"rewards/accuracies": 0.875,
|
| 1543 |
+
"rewards/margins": 1.7472874362953008,
|
| 1544 |
+
"logps/chosen": -56.32816154956818,
|
| 1545 |
+
"logps/rejected": -82.49419674873351,
|
| 1546 |
+
"epoch": 1.3439999999999999,
|
| 1547 |
+
"step": 840
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"loss": 0.39888598918914797,
|
| 1551 |
+
"grad_norm": 0.22760330140590668,
|
| 1552 |
+
"learning_rate": 3.208e-06,
|
| 1553 |
+
"entropy": 0.6883504351309966,
|
| 1554 |
+
"num_tokens": 11418114.0,
|
| 1555 |
+
"logits/chosen": -1.8767877806339819,
|
| 1556 |
+
"logits/rejected": -1.502016394931313,
|
| 1557 |
+
"mean_token_accuracy": 0.7253359684720635,
|
| 1558 |
+
"rewards/chosen": -0.836689239833504,
|
| 1559 |
+
"rewards/rejected": -2.3422608138527723,
|
| 1560 |
+
"rewards/accuracies": 0.80625,
|
| 1561 |
+
"rewards/margins": 1.5055715713184328,
|
| 1562 |
+
"logps/chosen": -53.072984755039215,
|
| 1563 |
+
"logps/rejected": -72.80904539823533,
|
| 1564 |
+
"epoch": 1.3599999999999999,
|
| 1565 |
+
"step": 850
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"loss": 0.40503392219543455,
|
| 1569 |
+
"grad_norm": 0.18328320980072021,
|
| 1570 |
+
"learning_rate": 3.1280000000000004e-06,
|
| 1571 |
+
"entropy": 0.7598929285246413,
|
| 1572 |
+
"num_tokens": 11552146.0,
|
| 1573 |
+
"logits/chosen": -1.644595842548687,
|
| 1574 |
+
"logits/rejected": -1.378459610872036,
|
| 1575 |
+
"mean_token_accuracy": 0.7259372064843774,
|
| 1576 |
+
"rewards/chosen": -0.9007043580990285,
|
| 1577 |
+
"rewards/rejected": -2.372628803132102,
|
| 1578 |
+
"rewards/accuracies": 0.8,
|
| 1579 |
+
"rewards/margins": 1.4719244507141411,
|
| 1580 |
+
"logps/chosen": -55.648898458480836,
|
| 1581 |
+
"logps/rejected": -78.23955315351486,
|
| 1582 |
+
"epoch": 1.376,
|
| 1583 |
+
"step": 860
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"loss": 0.3458548545837402,
|
| 1587 |
+
"grad_norm": 0.25410881638526917,
|
| 1588 |
+
"learning_rate": 3.0480000000000003e-06,
|
| 1589 |
+
"entropy": 0.7665736552706222,
|
| 1590 |
+
"num_tokens": 11686051.0,
|
| 1591 |
+
"logits/chosen": -1.578677616557072,
|
| 1592 |
+
"logits/rejected": -1.3612439345571947,
|
| 1593 |
+
"mean_token_accuracy": 0.713687221519649,
|
| 1594 |
+
"rewards/chosen": -0.9132042807443213,
|
| 1595 |
+
"rewards/rejected": -2.5326640743296593,
|
| 1596 |
+
"rewards/accuracies": 0.875,
|
| 1597 |
+
"rewards/margins": 1.6194597949273883,
|
| 1598 |
+
"logps/chosen": -62.23388094902039,
|
| 1599 |
+
"logps/rejected": -74.41176266670227,
|
| 1600 |
+
"epoch": 1.392,
|
| 1601 |
+
"step": 870
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"loss": 0.45064554214477537,
|
| 1605 |
+
"grad_norm": 0.31622838973999023,
|
| 1606 |
+
"learning_rate": 2.9680000000000002e-06,
|
| 1607 |
+
"entropy": 0.8155602383427322,
|
| 1608 |
+
"num_tokens": 11820419.0,
|
| 1609 |
+
"logits/chosen": -1.5256738113106172,
|
| 1610 |
+
"logits/rejected": -1.2444007046615775,
|
| 1611 |
+
"mean_token_accuracy": 0.7064365288242698,
|
| 1612 |
+
"rewards/chosen": -0.8790285978000612,
|
| 1613 |
+
"rewards/rejected": -2.279954434429237,
|
| 1614 |
+
"rewards/accuracies": 0.81875,
|
| 1615 |
+
"rewards/margins": 1.400925842532888,
|
| 1616 |
+
"logps/chosen": -61.98214646577835,
|
| 1617 |
+
"logps/rejected": -77.71078071594238,
|
| 1618 |
+
"epoch": 1.408,
|
| 1619 |
+
"step": 880
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"loss": 0.3549015045166016,
|
| 1623 |
+
"grad_norm": 0.11679533123970032,
|
| 1624 |
+
"learning_rate": 2.888e-06,
|
| 1625 |
+
"entropy": 0.8460370196611621,
|
| 1626 |
+
"num_tokens": 11954898.0,
|
| 1627 |
+
"logits/chosen": -1.501683007594535,
|
| 1628 |
+
"logits/rejected": -1.153882464001247,
|
| 1629 |
+
"mean_token_accuracy": 0.7017565036192537,
|
| 1630 |
+
"rewards/chosen": -0.9793404275318608,
|
| 1631 |
+
"rewards/rejected": -2.5882866755331633,
|
| 1632 |
+
"rewards/accuracies": 0.89375,
|
| 1633 |
+
"rewards/margins": 1.6089462437666953,
|
| 1634 |
+
"logps/chosen": -62.671991884708405,
|
| 1635 |
+
"logps/rejected": -85.92449575662613,
|
| 1636 |
+
"epoch": 1.424,
|
| 1637 |
+
"step": 890
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"loss": 0.32079691886901857,
|
| 1641 |
+
"grad_norm": 0.17164677381515503,
|
| 1642 |
+
"learning_rate": 2.808e-06,
|
| 1643 |
+
"entropy": 0.7836568029015325,
|
| 1644 |
+
"num_tokens": 12089403.0,
|
| 1645 |
+
"logits/chosen": -1.5902203304330556,
|
| 1646 |
+
"logits/rejected": -1.348343018569439,
|
| 1647 |
+
"mean_token_accuracy": 0.7099103134125471,
|
| 1648 |
+
"rewards/chosen": -0.906367156367196,
|
| 1649 |
+
"rewards/rejected": -2.5413650372705887,
|
| 1650 |
+
"rewards/accuracies": 0.88125,
|
| 1651 |
+
"rewards/margins": 1.634997878689319,
|
| 1652 |
+
"logps/chosen": -65.62818305492401,
|
| 1653 |
+
"logps/rejected": -79.59799456596375,
|
| 1654 |
+
"epoch": 1.44,
|
| 1655 |
+
"step": 900
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"loss": 0.3879056215286255,
|
| 1659 |
+
"grad_norm": 0.15152597427368164,
|
| 1660 |
+
"learning_rate": 2.728e-06,
|
| 1661 |
+
"entropy": 0.7635506895370782,
|
| 1662 |
+
"num_tokens": 12223715.0,
|
| 1663 |
+
"logits/chosen": -1.808664480653039,
|
| 1664 |
+
"logits/rejected": -1.4436211706139415,
|
| 1665 |
+
"mean_token_accuracy": 0.716391022503376,
|
| 1666 |
+
"rewards/chosen": -1.0626204300948303,
|
| 1667 |
+
"rewards/rejected": -2.617245429044124,
|
| 1668 |
+
"rewards/accuracies": 0.8125,
|
| 1669 |
+
"rewards/margins": 1.5546250022249297,
|
| 1670 |
+
"logps/chosen": -59.04966660737991,
|
| 1671 |
+
"logps/rejected": -79.71140024662017,
|
| 1672 |
+
"epoch": 1.456,
|
| 1673 |
+
"step": 910
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"loss": 0.5003737926483154,
|
| 1677 |
+
"grad_norm": 0.23257283866405487,
|
| 1678 |
+
"learning_rate": 2.648e-06,
|
| 1679 |
+
"entropy": 0.7805526316573378,
|
| 1680 |
+
"num_tokens": 12357865.0,
|
| 1681 |
+
"logits/chosen": -1.7092802822631128,
|
| 1682 |
+
"logits/rejected": -1.3571210243590142,
|
| 1683 |
+
"mean_token_accuracy": 0.7101537603884935,
|
| 1684 |
+
"rewards/chosen": -1.1259239912003978,
|
| 1685 |
+
"rewards/rejected": -2.455858718138188,
|
| 1686 |
+
"rewards/accuracies": 0.7875,
|
| 1687 |
+
"rewards/margins": 1.3299347181804477,
|
| 1688 |
+
"logps/chosen": -62.12466698884964,
|
| 1689 |
+
"logps/rejected": -79.51555901765823,
|
| 1690 |
+
"epoch": 1.472,
|
| 1691 |
+
"step": 920
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"loss": 0.4326589107513428,
|
| 1695 |
+
"grad_norm": 0.22872693836688995,
|
| 1696 |
+
"learning_rate": 2.568e-06,
|
| 1697 |
+
"entropy": 0.7554121573979501,
|
| 1698 |
+
"num_tokens": 12492128.0,
|
| 1699 |
+
"logits/chosen": -1.6719579694003723,
|
| 1700 |
+
"logits/rejected": -1.3663122032286132,
|
| 1701 |
+
"mean_token_accuracy": 0.6942645735107362,
|
| 1702 |
+
"rewards/chosen": -1.0867409846643568,
|
| 1703 |
+
"rewards/rejected": -2.5454429023200644,
|
| 1704 |
+
"rewards/accuracies": 0.79375,
|
| 1705 |
+
"rewards/margins": 1.4587019056081771,
|
| 1706 |
+
"logps/chosen": -63.67592179775238,
|
| 1707 |
+
"logps/rejected": -80.17804374694825,
|
| 1708 |
+
"epoch": 1.488,
|
| 1709 |
+
"step": 930
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"loss": 0.4602513790130615,
|
| 1713 |
+
"grad_norm": 0.23694008588790894,
|
| 1714 |
+
"learning_rate": 2.488e-06,
|
| 1715 |
+
"entropy": 0.7785145582747646,
|
| 1716 |
+
"num_tokens": 12626077.0,
|
| 1717 |
+
"logits/chosen": -1.5935042309292489,
|
| 1718 |
+
"logits/rejected": -1.3624577984140387,
|
| 1719 |
+
"mean_token_accuracy": 0.6916681522503495,
|
| 1720 |
+
"rewards/chosen": -1.1778735827072524,
|
| 1721 |
+
"rewards/rejected": -2.502696030540392,
|
| 1722 |
+
"rewards/accuracies": 0.8125,
|
| 1723 |
+
"rewards/margins": 1.3248224433511495,
|
| 1724 |
+
"logps/chosen": -63.580693411827085,
|
| 1725 |
+
"logps/rejected": -74.72070835828781,
|
| 1726 |
+
"epoch": 1.504,
|
| 1727 |
+
"step": 940
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"loss": 0.4067554473876953,
|
| 1731 |
+
"grad_norm": 0.14756956696510315,
|
| 1732 |
+
"learning_rate": 2.408e-06,
|
| 1733 |
+
"entropy": 0.8026386831275886,
|
| 1734 |
+
"num_tokens": 12760098.0,
|
| 1735 |
+
"logits/chosen": -1.5458875154452285,
|
| 1736 |
+
"logits/rejected": -1.2564218268845326,
|
| 1737 |
+
"mean_token_accuracy": 0.7152364812791348,
|
| 1738 |
+
"rewards/chosen": -1.2055871986289275,
|
| 1739 |
+
"rewards/rejected": -2.6404173804068707,
|
| 1740 |
+
"rewards/accuracies": 0.85,
|
| 1741 |
+
"rewards/margins": 1.4348301694728434,
|
| 1742 |
+
"logps/chosen": -66.13761007785797,
|
| 1743 |
+
"logps/rejected": -80.10225908756256,
|
| 1744 |
+
"epoch": 1.52,
|
| 1745 |
+
"step": 950
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"loss": 0.42372560501098633,
|
| 1749 |
+
"grad_norm": 0.14788998663425446,
|
| 1750 |
+
"learning_rate": 2.3280000000000004e-06,
|
| 1751 |
+
"entropy": 0.8012942865549121,
|
| 1752 |
+
"num_tokens": 12894652.0,
|
| 1753 |
+
"logits/chosen": -1.635073784869764,
|
| 1754 |
+
"logits/rejected": -1.3146647893445076,
|
| 1755 |
+
"mean_token_accuracy": 0.7114136775955557,
|
| 1756 |
+
"rewards/chosen": -1.1332746736516128,
|
| 1757 |
+
"rewards/rejected": -2.6066967224120163,
|
| 1758 |
+
"rewards/accuracies": 0.775,
|
| 1759 |
+
"rewards/margins": 1.4734220502898097,
|
| 1760 |
+
"logps/chosen": -63.05048147439957,
|
| 1761 |
+
"logps/rejected": -82.07224049568177,
|
| 1762 |
+
"epoch": 1.536,
|
| 1763 |
+
"step": 960
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"loss": 0.412662935256958,
|
| 1767 |
+
"grad_norm": 0.14732256531715393,
|
| 1768 |
+
"learning_rate": 2.2480000000000003e-06,
|
| 1769 |
+
"entropy": 0.8334445571847027,
|
| 1770 |
+
"num_tokens": 13030162.0,
|
| 1771 |
+
"logits/chosen": -1.6011668878902328,
|
| 1772 |
+
"logits/rejected": -1.2587357097834815,
|
| 1773 |
+
"mean_token_accuracy": 0.702854485437274,
|
| 1774 |
+
"rewards/chosen": -1.0372710642812308,
|
| 1775 |
+
"rewards/rejected": -2.6535557047463953,
|
| 1776 |
+
"rewards/accuracies": 0.8375,
|
| 1777 |
+
"rewards/margins": 1.6162846368737518,
|
| 1778 |
+
"logps/chosen": -65.74512588381768,
|
| 1779 |
+
"logps/rejected": -88.88693373203277,
|
| 1780 |
+
"epoch": 1.552,
|
| 1781 |
+
"step": 970
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"loss": 0.3407352209091187,
|
| 1785 |
+
"grad_norm": 0.2886792719364166,
|
| 1786 |
+
"learning_rate": 2.1680000000000002e-06,
|
| 1787 |
+
"entropy": 0.8762669585994445,
|
| 1788 |
+
"num_tokens": 13165409.0,
|
| 1789 |
+
"logits/chosen": -1.383209431558544,
|
| 1790 |
+
"logits/rejected": -1.0744050889575703,
|
| 1791 |
+
"mean_token_accuracy": 0.6733901240862906,
|
| 1792 |
+
"rewards/chosen": -1.008211946907977,
|
| 1793 |
+
"rewards/rejected": -2.624329062405741,
|
| 1794 |
+
"rewards/accuracies": 0.86875,
|
| 1795 |
+
"rewards/margins": 1.6161171085201205,
|
| 1796 |
+
"logps/chosen": -69.2023845911026,
|
| 1797 |
+
"logps/rejected": -83.00953713655471,
|
| 1798 |
+
"epoch": 1.568,
|
| 1799 |
+
"step": 980
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"loss": 0.4386160850524902,
|
| 1803 |
+
"grad_norm": 0.2656986713409424,
|
| 1804 |
+
"learning_rate": 2.088e-06,
|
| 1805 |
+
"entropy": 0.8209956398721261,
|
| 1806 |
+
"num_tokens": 13299607.0,
|
| 1807 |
+
"logits/chosen": -1.4395911330145508,
|
| 1808 |
+
"logits/rejected": -1.2677628081064927,
|
| 1809 |
+
"mean_token_accuracy": 0.6904863499104976,
|
| 1810 |
+
"rewards/chosen": -1.0870793154346756,
|
| 1811 |
+
"rewards/rejected": -2.4575889983447268,
|
| 1812 |
+
"rewards/accuracies": 0.78125,
|
| 1813 |
+
"rewards/margins": 1.3705096821766347,
|
| 1814 |
+
"logps/chosen": -65.42525545358657,
|
| 1815 |
+
"logps/rejected": -74.03723225593566,
|
| 1816 |
+
"epoch": 1.584,
|
| 1817 |
+
"step": 990
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"loss": 0.35819923877716064,
|
| 1821 |
+
"grad_norm": 0.35382354259490967,
|
| 1822 |
+
"learning_rate": 2.008e-06,
|
| 1823 |
+
"entropy": 0.7399830836133333,
|
| 1824 |
+
"num_tokens": 13434502.0,
|
| 1825 |
+
"logits/chosen": -1.6976982361606354,
|
| 1826 |
+
"logits/rejected": -1.3072192337860213,
|
| 1827 |
+
"mean_token_accuracy": 0.7421066265553236,
|
| 1828 |
+
"rewards/chosen": -1.0984582830686123,
|
| 1829 |
+
"rewards/rejected": -2.7966011623386295,
|
| 1830 |
+
"rewards/accuracies": 0.875,
|
| 1831 |
+
"rewards/margins": 1.6981428703293204,
|
| 1832 |
+
"logps/chosen": -62.93297493457794,
|
| 1833 |
+
"logps/rejected": -82.4026222705841,
|
| 1834 |
+
"epoch": 1.6,
|
| 1835 |
+
"step": 1000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"loss": 0.4160293102264404,
|
| 1839 |
+
"grad_norm": 0.11767679452896118,
|
| 1840 |
+
"learning_rate": 1.928e-06,
|
| 1841 |
+
"entropy": 0.7990678637885139,
|
| 1842 |
+
"num_tokens": 13568600.0,
|
| 1843 |
+
"logits/chosen": -1.5975146089430179,
|
| 1844 |
+
"logits/rejected": -1.2924786059115378,
|
| 1845 |
+
"mean_token_accuracy": 0.7110372580587864,
|
| 1846 |
+
"rewards/chosen": -0.9529279105423484,
|
| 1847 |
+
"rewards/rejected": -2.457513489993289,
|
| 1848 |
+
"rewards/accuracies": 0.83125,
|
| 1849 |
+
"rewards/margins": 1.504585579968989,
|
| 1850 |
+
"logps/chosen": -63.974093568325046,
|
| 1851 |
+
"logps/rejected": -74.71904839277268,
|
| 1852 |
+
"epoch": 1.616,
|
| 1853 |
+
"step": 1010
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"loss": 0.3250428199768066,
|
| 1857 |
+
"grad_norm": 0.12981919944286346,
|
| 1858 |
+
"learning_rate": 1.8480000000000001e-06,
|
| 1859 |
+
"entropy": 0.7811902783811092,
|
| 1860 |
+
"num_tokens": 13702787.0,
|
| 1861 |
+
"logits/chosen": -1.587494817410056,
|
| 1862 |
+
"logits/rejected": -1.1941186838362041,
|
| 1863 |
+
"mean_token_accuracy": 0.7234507920220494,
|
| 1864 |
+
"rewards/chosen": -0.8217556012241403,
|
| 1865 |
+
"rewards/rejected": -2.597050206689164,
|
| 1866 |
+
"rewards/accuracies": 0.86875,
|
| 1867 |
+
"rewards/margins": 1.7752946069464088,
|
| 1868 |
+
"logps/chosen": -61.340254080295566,
|
| 1869 |
+
"logps/rejected": -79.60141725540161,
|
| 1870 |
+
"epoch": 1.6320000000000001,
|
| 1871 |
+
"step": 1020
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"loss": 0.4002103805541992,
|
| 1875 |
+
"grad_norm": 0.2534341812133789,
|
| 1876 |
+
"learning_rate": 1.7680000000000003e-06,
|
| 1877 |
+
"entropy": 0.8721957165267668,
|
| 1878 |
+
"num_tokens": 13837799.0,
|
| 1879 |
+
"logits/chosen": -1.4787466162956828,
|
| 1880 |
+
"logits/rejected": -1.1347675946257396,
|
| 1881 |
+
"mean_token_accuracy": 0.7160949306562543,
|
| 1882 |
+
"rewards/chosen": -0.9889563272474333,
|
| 1883 |
+
"rewards/rejected": -2.5054163753055034,
|
| 1884 |
+
"rewards/accuracies": 0.81875,
|
| 1885 |
+
"rewards/margins": 1.5164600439369678,
|
| 1886 |
+
"logps/chosen": -69.03193366527557,
|
| 1887 |
+
"logps/rejected": -80.24655103683472,
|
| 1888 |
+
"epoch": 1.6480000000000001,
|
| 1889 |
+
"step": 1030
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"loss": 0.4034124374389648,
|
| 1893 |
+
"grad_norm": 0.23815952241420746,
|
| 1894 |
+
"learning_rate": 1.6880000000000002e-06,
|
| 1895 |
+
"entropy": 0.7898992500049644,
|
| 1896 |
+
"num_tokens": 13972529.0,
|
| 1897 |
+
"logits/chosen": -1.6163878058701155,
|
| 1898 |
+
"logits/rejected": -1.2316415786069164,
|
| 1899 |
+
"mean_token_accuracy": 0.7335911913774907,
|
| 1900 |
+
"rewards/chosen": -0.9068507085920828,
|
| 1901 |
+
"rewards/rejected": -2.4930222455412148,
|
| 1902 |
+
"rewards/accuracies": 0.84375,
|
| 1903 |
+
"rewards/margins": 1.5861715397797524,
|
| 1904 |
+
"logps/chosen": -62.6472428560257,
|
| 1905 |
+
"logps/rejected": -78.13663581609725,
|
| 1906 |
+
"epoch": 1.6640000000000001,
|
| 1907 |
+
"step": 1040
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"loss": 0.4241136074066162,
|
| 1911 |
+
"grad_norm": 0.1419256180524826,
|
| 1912 |
+
"learning_rate": 1.608e-06,
|
| 1913 |
+
"entropy": 0.9256169050706375,
|
| 1914 |
+
"num_tokens": 14107336.0,
|
| 1915 |
+
"logits/chosen": -1.2789081893064282,
|
| 1916 |
+
"logits/rejected": -1.098782251423331,
|
| 1917 |
+
"mean_token_accuracy": 0.696486575063318,
|
| 1918 |
+
"rewards/chosen": -0.8607778700679773,
|
| 1919 |
+
"rewards/rejected": -2.315709656581748,
|
| 1920 |
+
"rewards/accuracies": 0.8125,
|
| 1921 |
+
"rewards/margins": 1.454931782837957,
|
| 1922 |
+
"logps/chosen": -74.47855192422867,
|
| 1923 |
+
"logps/rejected": -76.75005420446396,
|
| 1924 |
+
"epoch": 1.6800000000000002,
|
| 1925 |
+
"step": 1050
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"loss": 0.48003559112548827,
|
| 1929 |
+
"grad_norm": 0.14790725708007812,
|
| 1930 |
+
"learning_rate": 1.528e-06,
|
| 1931 |
+
"entropy": 0.7936980344704352,
|
| 1932 |
+
"num_tokens": 14241540.0,
|
| 1933 |
+
"logits/chosen": -1.5867842467297604,
|
| 1934 |
+
"logits/rejected": -1.1837827649291766,
|
| 1935 |
+
"mean_token_accuracy": 0.7259030997753143,
|
| 1936 |
+
"rewards/chosen": -0.8804566722927121,
|
| 1937 |
+
"rewards/rejected": -2.2063653921242805,
|
| 1938 |
+
"rewards/accuracies": 0.79375,
|
| 1939 |
+
"rewards/margins": 1.3259087240789085,
|
| 1940 |
+
"logps/chosen": -58.475341248512265,
|
| 1941 |
+
"logps/rejected": -74.59980441331864,
|
| 1942 |
+
"epoch": 1.696,
|
| 1943 |
+
"step": 1060
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"loss": 0.3922820329666138,
|
| 1947 |
+
"grad_norm": 0.1496279537677765,
|
| 1948 |
+
"learning_rate": 1.4480000000000002e-06,
|
| 1949 |
+
"entropy": 0.8726246880483813,
|
| 1950 |
+
"num_tokens": 14376121.0,
|
| 1951 |
+
"logits/chosen": -1.4914376459842063,
|
| 1952 |
+
"logits/rejected": -1.1848033257275015,
|
| 1953 |
+
"mean_token_accuracy": 0.7143584050238132,
|
| 1954 |
+
"rewards/chosen": -0.9215551663655788,
|
| 1955 |
+
"rewards/rejected": -2.562272682785988,
|
| 1956 |
+
"rewards/accuracies": 0.84375,
|
| 1957 |
+
"rewards/margins": 1.6407175094820559,
|
| 1958 |
+
"logps/chosen": -67.5427442073822,
|
| 1959 |
+
"logps/rejected": -79.30532623529434,
|
| 1960 |
+
"epoch": 1.712,
|
| 1961 |
+
"step": 1070
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"loss": 0.44224209785461427,
|
| 1965 |
+
"grad_norm": 0.2878839075565338,
|
| 1966 |
+
"learning_rate": 1.368e-06,
|
| 1967 |
+
"entropy": 0.7888671966618859,
|
| 1968 |
+
"num_tokens": 14510553.0,
|
| 1969 |
+
"logits/chosen": -1.6546901899116329,
|
| 1970 |
+
"logits/rejected": -1.3682432627615229,
|
| 1971 |
+
"mean_token_accuracy": 0.6996359150856734,
|
| 1972 |
+
"rewards/chosen": -0.8950480898056412,
|
| 1973 |
+
"rewards/rejected": -2.3655205052462405,
|
| 1974 |
+
"rewards/accuracies": 0.80625,
|
| 1975 |
+
"rewards/margins": 1.4704724150244146,
|
| 1976 |
+
"logps/chosen": -63.91799589395523,
|
| 1977 |
+
"logps/rejected": -74.64101424217225,
|
| 1978 |
+
"epoch": 1.728,
|
| 1979 |
+
"step": 1080
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"loss": 0.3927299499511719,
|
| 1983 |
+
"grad_norm": 0.1370965987443924,
|
| 1984 |
+
"learning_rate": 1.288e-06,
|
| 1985 |
+
"entropy": 0.8091648026223993,
|
| 1986 |
+
"num_tokens": 14644507.0,
|
| 1987 |
+
"logits/chosen": -1.687283619747059,
|
| 1988 |
+
"logits/rejected": -1.3548523346162953,
|
| 1989 |
+
"mean_token_accuracy": 0.7228656150400639,
|
| 1990 |
+
"rewards/chosen": -0.7743405980079843,
|
| 1991 |
+
"rewards/rejected": -2.3230774260126053,
|
| 1992 |
+
"rewards/accuracies": 0.7875,
|
| 1993 |
+
"rewards/margins": 1.5487368311733007,
|
| 1994 |
+
"logps/chosen": -56.80163584947586,
|
| 1995 |
+
"logps/rejected": -80.1871037721634,
|
| 1996 |
+
"epoch": 1.744,
|
| 1997 |
+
"step": 1090
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"loss": 0.36075007915496826,
|
| 2001 |
+
"grad_norm": 0.12493802607059479,
|
| 2002 |
+
"learning_rate": 1.2080000000000001e-06,
|
| 2003 |
+
"entropy": 0.8516323209594703,
|
| 2004 |
+
"num_tokens": 14779558.0,
|
| 2005 |
+
"logits/chosen": -1.5479666328178479,
|
| 2006 |
+
"logits/rejected": -1.1764522167627494,
|
| 2007 |
+
"mean_token_accuracy": 0.7081021483056247,
|
| 2008 |
+
"rewards/chosen": -0.8559462582834385,
|
| 2009 |
+
"rewards/rejected": -2.430596137570683,
|
| 2010 |
+
"rewards/accuracies": 0.84375,
|
| 2011 |
+
"rewards/margins": 1.5746498768217863,
|
| 2012 |
+
"logps/chosen": -63.79940243959427,
|
| 2013 |
+
"logps/rejected": -79.56198363304138,
|
| 2014 |
+
"epoch": 1.76,
|
| 2015 |
+
"step": 1100
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"loss": 0.49372134208679197,
|
| 2019 |
+
"grad_norm": 0.24344250559806824,
|
| 2020 |
+
"learning_rate": 1.128e-06,
|
| 2021 |
+
"entropy": 0.8455155969480984,
|
| 2022 |
+
"num_tokens": 14914286.0,
|
| 2023 |
+
"logits/chosen": -1.498396463441885,
|
| 2024 |
+
"logits/rejected": -1.1425197974463714,
|
| 2025 |
+
"mean_token_accuracy": 0.7100071370601654,
|
| 2026 |
+
"rewards/chosen": -0.8169066427948565,
|
| 2027 |
+
"rewards/rejected": -2.166107503604144,
|
| 2028 |
+
"rewards/accuracies": 0.76875,
|
| 2029 |
+
"rewards/margins": 1.3492008646018803,
|
| 2030 |
+
"logps/chosen": -61.403113543987274,
|
| 2031 |
+
"logps/rejected": -75.34919095039368,
|
| 2032 |
+
"epoch": 1.776,
|
| 2033 |
+
"step": 1110
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"loss": 0.4005413055419922,
|
| 2037 |
+
"grad_norm": 0.20426344871520996,
|
| 2038 |
+
"learning_rate": 1.0480000000000002e-06,
|
| 2039 |
+
"entropy": 0.8211736791708972,
|
| 2040 |
+
"num_tokens": 15048840.0,
|
| 2041 |
+
"logits/chosen": -1.582013316231499,
|
| 2042 |
+
"logits/rejected": -1.3164341995727113,
|
| 2043 |
+
"mean_token_accuracy": 0.7223803894594312,
|
| 2044 |
+
"rewards/chosen": -0.6959973029966932,
|
| 2045 |
+
"rewards/rejected": -2.1918047105107687,
|
| 2046 |
+
"rewards/accuracies": 0.84375,
|
| 2047 |
+
"rewards/margins": 1.4958074030466377,
|
| 2048 |
+
"logps/chosen": -59.653766822814944,
|
| 2049 |
+
"logps/rejected": -75.93207306861878,
|
| 2050 |
+
"epoch": 1.792,
|
| 2051 |
+
"step": 1120
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"loss": 0.5007998466491699,
|
| 2055 |
+
"grad_norm": 0.1154891774058342,
|
| 2056 |
+
"learning_rate": 9.68e-07,
|
| 2057 |
+
"entropy": 0.8647114810533821,
|
| 2058 |
+
"num_tokens": 15184276.0,
|
| 2059 |
+
"logits/chosen": -1.356775653423806,
|
| 2060 |
+
"logits/rejected": -1.1500322818720279,
|
| 2061 |
+
"mean_token_accuracy": 0.7039438035339117,
|
| 2062 |
+
"rewards/chosen": -0.8606041681859097,
|
| 2063 |
+
"rewards/rejected": -2.0742746030678973,
|
| 2064 |
+
"rewards/accuracies": 0.7375,
|
| 2065 |
+
"rewards/margins": 1.2136704298667609,
|
| 2066 |
+
"logps/chosen": -64.9673285484314,
|
| 2067 |
+
"logps/rejected": -80.83175637722016,
|
| 2068 |
+
"epoch": 1.808,
|
| 2069 |
+
"step": 1130
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"loss": 0.4986130714416504,
|
| 2073 |
+
"grad_norm": 0.6216063499450684,
|
| 2074 |
+
"learning_rate": 8.880000000000001e-07,
|
| 2075 |
+
"entropy": 0.774267910355411,
|
| 2076 |
+
"num_tokens": 15317721.0,
|
| 2077 |
+
"logits/chosen": -1.4605975439966772,
|
| 2078 |
+
"logits/rejected": -1.3212052490062418,
|
| 2079 |
+
"mean_token_accuracy": 0.7165707518346608,
|
| 2080 |
+
"rewards/chosen": -0.8740379733324517,
|
| 2081 |
+
"rewards/rejected": -1.9363951487233861,
|
| 2082 |
+
"rewards/accuracies": 0.76875,
|
| 2083 |
+
"rewards/margins": 1.0623571707867085,
|
| 2084 |
+
"logps/chosen": -58.174967527389526,
|
| 2085 |
+
"logps/rejected": -67.10608189105987,
|
| 2086 |
+
"epoch": 1.8239999999999998,
|
| 2087 |
+
"step": 1140
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"loss": 0.39597718715667723,
|
| 2091 |
+
"grad_norm": 0.1611914187669754,
|
| 2092 |
+
"learning_rate": 8.08e-07,
|
| 2093 |
+
"entropy": 0.8271302699897205,
|
| 2094 |
+
"num_tokens": 15452553.0,
|
| 2095 |
+
"logits/chosen": -1.50679093223041,
|
| 2096 |
+
"logits/rejected": -1.147970850967657,
|
| 2097 |
+
"mean_token_accuracy": 0.7078391901217401,
|
| 2098 |
+
"rewards/chosen": -0.8506218325230293,
|
| 2099 |
+
"rewards/rejected": -2.2930383594852173,
|
| 2100 |
+
"rewards/accuracies": 0.8375,
|
| 2101 |
+
"rewards/margins": 1.4424165301956235,
|
| 2102 |
+
"logps/chosen": -60.48249053955078,
|
| 2103 |
+
"logps/rejected": -79.1380724787712,
|
| 2104 |
+
"epoch": 1.8399999999999999,
|
| 2105 |
+
"step": 1150
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"loss": 0.3954659700393677,
|
| 2109 |
+
"grad_norm": 0.19070428609848022,
|
| 2110 |
+
"learning_rate": 7.280000000000001e-07,
|
| 2111 |
+
"entropy": 0.7992117294867057,
|
| 2112 |
+
"num_tokens": 15586655.0,
|
| 2113 |
+
"logits/chosen": -1.6674989181953035,
|
| 2114 |
+
"logits/rejected": -1.2958623246252912,
|
| 2115 |
+
"mean_token_accuracy": 0.7274953337386251,
|
| 2116 |
+
"rewards/chosen": -0.897261195579631,
|
| 2117 |
+
"rewards/rejected": -2.37544046624098,
|
| 2118 |
+
"rewards/accuracies": 0.8,
|
| 2119 |
+
"rewards/margins": 1.4781792684458197,
|
| 2120 |
+
"logps/chosen": -55.330790144205096,
|
| 2121 |
+
"logps/rejected": -79.54153176546097,
|
| 2122 |
+
"epoch": 1.8559999999999999,
|
| 2123 |
+
"step": 1160
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"loss": 0.5150039196014404,
|
| 2127 |
+
"grad_norm": 0.2727234959602356,
|
| 2128 |
+
"learning_rate": 6.48e-07,
|
| 2129 |
+
"entropy": 0.7774595006601885,
|
| 2130 |
+
"num_tokens": 15720172.0,
|
| 2131 |
+
"logits/chosen": -1.673701828468026,
|
| 2132 |
+
"logits/rejected": -1.2861514487494707,
|
| 2133 |
+
"mean_token_accuracy": 0.7126229584217072,
|
| 2134 |
+
"rewards/chosen": -0.9092265904924716,
|
| 2135 |
+
"rewards/rejected": -2.033220373163931,
|
| 2136 |
+
"rewards/accuracies": 0.71875,
|
| 2137 |
+
"rewards/margins": 1.1239937825594097,
|
| 2138 |
+
"logps/chosen": -54.55541696548462,
|
| 2139 |
+
"logps/rejected": -76.10742880105973,
|
| 2140 |
+
"epoch": 1.8719999999999999,
|
| 2141 |
+
"step": 1170
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"loss": 0.4881202220916748,
|
| 2145 |
+
"grad_norm": 0.21918685734272003,
|
| 2146 |
+
"learning_rate": 5.680000000000001e-07,
|
| 2147 |
+
"entropy": 0.7866302890935912,
|
| 2148 |
+
"num_tokens": 15853755.0,
|
| 2149 |
+
"logits/chosen": -1.4762931082547661,
|
| 2150 |
+
"logits/rejected": -1.2942548537923684,
|
| 2151 |
+
"mean_token_accuracy": 0.7225816631689668,
|
| 2152 |
+
"rewards/chosen": -0.9264017570319993,
|
| 2153 |
+
"rewards/rejected": -1.9995961244334466,
|
| 2154 |
+
"rewards/accuracies": 0.7875,
|
| 2155 |
+
"rewards/margins": 1.0731943636201322,
|
| 2156 |
+
"logps/chosen": -60.705457758903506,
|
| 2157 |
+
"logps/rejected": -66.6190501689911,
|
| 2158 |
+
"epoch": 1.888,
|
| 2159 |
+
"step": 1180
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"loss": 0.45177521705627444,
|
| 2163 |
+
"grad_norm": 0.23347221314907074,
|
| 2164 |
+
"learning_rate": 4.88e-07,
|
| 2165 |
+
"entropy": 0.8309914332116023,
|
| 2166 |
+
"num_tokens": 15988701.0,
|
| 2167 |
+
"logits/chosen": -1.4661994627051458,
|
| 2168 |
+
"logits/rejected": -1.2287666153016743,
|
| 2169 |
+
"mean_token_accuracy": 0.7092148935422301,
|
| 2170 |
+
"rewards/chosen": -0.8807039022503886,
|
| 2171 |
+
"rewards/rejected": -2.185251401356072,
|
| 2172 |
+
"rewards/accuracies": 0.80625,
|
| 2173 |
+
"rewards/margins": 1.304547501122579,
|
| 2174 |
+
"logps/chosen": -62.431247127056125,
|
| 2175 |
+
"logps/rejected": -74.59568692445755,
|
| 2176 |
+
"epoch": 1.904,
|
| 2177 |
+
"step": 1190
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"loss": 0.44033398628234866,
|
| 2181 |
+
"grad_norm": 0.13558809459209442,
|
| 2182 |
+
"learning_rate": 4.0800000000000005e-07,
|
| 2183 |
+
"entropy": 0.8481319433776662,
|
| 2184 |
+
"num_tokens": 16123227.0,
|
| 2185 |
+
"logits/chosen": -1.4890639792428868,
|
| 2186 |
+
"logits/rejected": -1.1712801983612529,
|
| 2187 |
+
"mean_token_accuracy": 0.7045220224186778,
|
| 2188 |
+
"rewards/chosen": -0.8116515700123272,
|
| 2189 |
+
"rewards/rejected": -2.0684703564504163,
|
| 2190 |
+
"rewards/accuracies": 0.83125,
|
| 2191 |
+
"rewards/margins": 1.2568187874741852,
|
| 2192 |
+
"logps/chosen": -63.18868860006332,
|
| 2193 |
+
"logps/rejected": -74.17321823835373,
|
| 2194 |
+
"epoch": 1.92,
|
| 2195 |
+
"step": 1200
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"loss": 0.4489591598510742,
|
| 2199 |
+
"grad_norm": 0.2724771499633789,
|
| 2200 |
+
"learning_rate": 3.280000000000001e-07,
|
| 2201 |
+
"entropy": 0.8390395241905935,
|
| 2202 |
+
"num_tokens": 16257799.0,
|
| 2203 |
+
"logits/chosen": -1.3757150913324974,
|
| 2204 |
+
"logits/rejected": -1.180797201944608,
|
| 2205 |
+
"mean_token_accuracy": 0.6906995047815144,
|
| 2206 |
+
"rewards/chosen": -0.8883007764234208,
|
| 2207 |
+
"rewards/rejected": -2.3104404117213564,
|
| 2208 |
+
"rewards/accuracies": 0.8125,
|
| 2209 |
+
"rewards/margins": 1.422139625577256,
|
| 2210 |
+
"logps/chosen": -64.57728606462479,
|
| 2211 |
+
"logps/rejected": -76.76991089582444,
|
| 2212 |
+
"epoch": 1.936,
|
| 2213 |
+
"step": 1210
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"loss": 0.3736592769622803,
|
| 2217 |
+
"grad_norm": 0.23324978351593018,
|
| 2218 |
+
"learning_rate": 2.48e-07,
|
| 2219 |
+
"entropy": 0.7958670913358219,
|
| 2220 |
+
"num_tokens": 16392755.0,
|
| 2221 |
+
"logits/chosen": -1.5214667053964295,
|
| 2222 |
+
"logits/rejected": -1.1308945168158175,
|
| 2223 |
+
"mean_token_accuracy": 0.7339793419465422,
|
| 2224 |
+
"rewards/chosen": -0.8573693673475645,
|
| 2225 |
+
"rewards/rejected": -2.422400009748526,
|
| 2226 |
+
"rewards/accuracies": 0.84375,
|
| 2227 |
+
"rewards/margins": 1.565030633006245,
|
| 2228 |
+
"logps/chosen": -62.808938360214235,
|
| 2229 |
+
"logps/rejected": -81.37407802343368,
|
| 2230 |
+
"epoch": 1.952,
|
| 2231 |
+
"step": 1220
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"loss": 0.33856496810913084,
|
| 2235 |
+
"grad_norm": 0.2365734577178955,
|
| 2236 |
+
"learning_rate": 1.68e-07,
|
| 2237 |
+
"entropy": 0.7709657683037221,
|
| 2238 |
+
"num_tokens": 16526347.0,
|
| 2239 |
+
"logits/chosen": -1.5951650245485083,
|
| 2240 |
+
"logits/rejected": -1.2316822978483537,
|
| 2241 |
+
"mean_token_accuracy": 0.7203758641146123,
|
| 2242 |
+
"rewards/chosen": -0.7306884591787821,
|
| 2243 |
+
"rewards/rejected": -2.310912328850827,
|
| 2244 |
+
"rewards/accuracies": 0.8625,
|
| 2245 |
+
"rewards/margins": 1.5802238748408854,
|
| 2246 |
+
"logps/chosen": -53.69367561340332,
|
| 2247 |
+
"logps/rejected": -75.45296376943588,
|
| 2248 |
+
"epoch": 1.968,
|
| 2249 |
+
"step": 1230
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"loss": 0.47231736183166506,
|
| 2253 |
+
"grad_norm": 0.1407570242881775,
|
| 2254 |
+
"learning_rate": 8.800000000000001e-08,
|
| 2255 |
+
"entropy": 0.9006856581720057,
|
| 2256 |
+
"num_tokens": 16660816.0,
|
| 2257 |
+
"logits/chosen": -1.4061471100194554,
|
| 2258 |
+
"logits/rejected": -1.0629769642593176,
|
| 2259 |
+
"mean_token_accuracy": 0.697626062296331,
|
| 2260 |
+
"rewards/chosen": -0.7163842913665576,
|
| 2261 |
+
"rewards/rejected": -1.9068784552206126,
|
| 2262 |
+
"rewards/accuracies": 0.8,
|
| 2263 |
+
"rewards/margins": 1.1904941664426587,
|
| 2264 |
+
"logps/chosen": -60.04284707307816,
|
| 2265 |
+
"logps/rejected": -76.30784704685212,
|
| 2266 |
+
"epoch": 1.984,
|
| 2267 |
+
"step": 1240
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"loss": 0.4329345226287842,
|
| 2271 |
+
"grad_norm": 0.20022091269493103,
|
| 2272 |
+
"learning_rate": 8e-09,
|
| 2273 |
+
"entropy": 0.7137106273818062,
|
| 2274 |
+
"num_tokens": 16794078.0,
|
| 2275 |
+
"logits/chosen": -1.7487986939590061,
|
| 2276 |
+
"logits/rejected": -1.4640005781460887,
|
| 2277 |
+
"mean_token_accuracy": 0.7085168479010463,
|
| 2278 |
+
"rewards/chosen": -0.7161212222184986,
|
| 2279 |
+
"rewards/rejected": -2.086236947233556,
|
| 2280 |
+
"rewards/accuracies": 0.80625,
|
| 2281 |
+
"rewards/margins": 1.3701157211326063,
|
| 2282 |
+
"logps/chosen": -54.929140663146974,
|
| 2283 |
+
"logps/rejected": -69.96422597169877,
|
| 2284 |
+
"epoch": 2.0,
|
| 2285 |
+
"step": 1250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"train_runtime": 5977.8981,
|
| 2289 |
+
"train_samples_per_second": 3.346,
|
| 2290 |
+
"train_steps_per_second": 0.209,
|
| 2291 |
+
"total_flos": 1.3612284112597402e+17,
|
| 2292 |
+
"train_loss": 0.45482119541168214,
|
| 2293 |
+
"epoch": 2.0,
|
| 2294 |
+
"step": 1250
|
| 2295 |
+
}
|
| 2296 |
+
],
|
| 2297 |
+
"validation_monitor_size": 0,
|
| 2298 |
+
"validation_monitor_selection": "fixed_seed_random_without_replacement",
|
| 2299 |
+
"validation_monitor_seed": 22,
|
| 2300 |
+
"validation_monitor_indices": [],
|
| 2301 |
+
"early_stopping_patience": null
|
| 2302 |
+
}
|