Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- D_full_corrected_seed23/adapter_config.json +43 -0
- D_full_corrected_seed23/adapter_model.safetensors +3 -0
- D_full_corrected_seed23/chat_template.jinja +54 -0
- D_full_corrected_seed23/tokenizer.json +3 -0
- D_full_corrected_seed23/tokenizer_config.json +29 -0
- D_full_corrected_seed23/training_summary.json +2302 -0
.gitattributes
CHANGED
|
@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
D_full_emf1_seed23/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
D_full_corrected_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
D_full_emf1_seed23/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
D_full_corrected_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
D_full_corrected_seed23/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
D_full_corrected_seed23/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_corrected_seed23/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:012aaaf7c27a174fdfe9ddcc8b50f75dfd2ff216a889e8d85c85544911da20df
|
| 3 |
+
size 17462432
|
D_full_corrected_seed23/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_corrected_seed23/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_corrected_seed23/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_corrected_seed23/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": 23,
|
| 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": 5990.5564,
|
| 31 |
+
"train_samples_per_second": 3.339,
|
| 32 |
+
"train_steps_per_second": 0.209,
|
| 33 |
+
"total_flos": 1.3612284112597402e+17,
|
| 34 |
+
"train_loss": 0.4133561391830444
|
| 35 |
+
},
|
| 36 |
+
"log_history": [
|
| 37 |
+
{
|
| 38 |
+
"loss": 0.6908764839172363,
|
| 39 |
+
"grad_norm": 0.1349368393421173,
|
| 40 |
+
"learning_rate": 9.928e-06,
|
| 41 |
+
"entropy": 0.8426685474347323,
|
| 42 |
+
"num_tokens": 133829.0,
|
| 43 |
+
"logits/chosen": -1.0278949735298577,
|
| 44 |
+
"logits/rejected": -1.0196253362684828,
|
| 45 |
+
"mean_token_accuracy": 0.714401988312602,
|
| 46 |
+
"rewards/chosen": 0.0029785050165173745,
|
| 47 |
+
"rewards/rejected": -0.0019342578076248173,
|
| 48 |
+
"rewards/accuracies": 0.50625,
|
| 49 |
+
"rewards/margins": 0.004912762764615763,
|
| 50 |
+
"logps/chosen": -57.39018324613571,
|
| 51 |
+
"logps/rejected": -48.962822711467744,
|
| 52 |
+
"epoch": 0.016,
|
| 53 |
+
"step": 10
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"loss": 0.6819325447082519,
|
| 57 |
+
"grad_norm": 0.09928478300571442,
|
| 58 |
+
"learning_rate": 9.848000000000001e-06,
|
| 59 |
+
"entropy": 0.7880316659575328,
|
| 60 |
+
"num_tokens": 268096.0,
|
| 61 |
+
"logits/chosen": -1.1536350940643518,
|
| 62 |
+
"logits/rejected": -1.0090869685947004,
|
| 63 |
+
"mean_token_accuracy": 0.740577093232423,
|
| 64 |
+
"rewards/chosen": -0.012110545788709714,
|
| 65 |
+
"rewards/rejected": -0.03527573887986364,
|
| 66 |
+
"rewards/accuracies": 0.6625,
|
| 67 |
+
"rewards/margins": 0.023165192987653425,
|
| 68 |
+
"logps/chosen": -48.57938648462296,
|
| 69 |
+
"logps/rejected": -55.74204361438751,
|
| 70 |
+
"epoch": 0.032,
|
| 71 |
+
"step": 20
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"loss": 0.651966142654419,
|
| 75 |
+
"grad_norm": 0.12356256693601608,
|
| 76 |
+
"learning_rate": 9.768e-06,
|
| 77 |
+
"entropy": 0.9091600199579262,
|
| 78 |
+
"num_tokens": 403100.0,
|
| 79 |
+
"logits/chosen": -0.8498058642694668,
|
| 80 |
+
"logits/rejected": -0.8488367408309598,
|
| 81 |
+
"mean_token_accuracy": 0.722683965601027,
|
| 82 |
+
"rewards/chosen": -0.02561774755413353,
|
| 83 |
+
"rewards/rejected": -0.11284091770812665,
|
| 84 |
+
"rewards/accuracies": 0.81875,
|
| 85 |
+
"rewards/margins": 0.08722316985949874,
|
| 86 |
+
"logps/chosen": -63.05102525949478,
|
| 87 |
+
"logps/rejected": -59.17153347730637,
|
| 88 |
+
"epoch": 0.048,
|
| 89 |
+
"step": 30
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"loss": 0.638601016998291,
|
| 93 |
+
"grad_norm": 0.1074950322508812,
|
| 94 |
+
"learning_rate": 9.688000000000001e-06,
|
| 95 |
+
"entropy": 0.8092274450114928,
|
| 96 |
+
"num_tokens": 536957.0,
|
| 97 |
+
"logits/chosen": -0.8933127736032072,
|
| 98 |
+
"logits/rejected": -0.8002460568149278,
|
| 99 |
+
"mean_token_accuracy": 0.7233862653374672,
|
| 100 |
+
"rewards/chosen": -0.08373595155862859,
|
| 101 |
+
"rewards/rejected": -0.20470629430783446,
|
| 102 |
+
"rewards/accuracies": 0.8125,
|
| 103 |
+
"rewards/margins": 0.12097034289326984,
|
| 104 |
+
"logps/chosen": -56.76135518550873,
|
| 105 |
+
"logps/rejected": -51.07406049966812,
|
| 106 |
+
"epoch": 0.064,
|
| 107 |
+
"step": 40
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"loss": 0.6213541984558105,
|
| 111 |
+
"grad_norm": 0.11525370925664902,
|
| 112 |
+
"learning_rate": 9.608e-06,
|
| 113 |
+
"entropy": 0.8385081959189847,
|
| 114 |
+
"num_tokens": 672162.0,
|
| 115 |
+
"logits/chosen": -0.8699766650620504,
|
| 116 |
+
"logits/rejected": -0.8224479540489131,
|
| 117 |
+
"mean_token_accuracy": 0.7261195071041584,
|
| 118 |
+
"rewards/chosen": -0.1754664451329518,
|
| 119 |
+
"rewards/rejected": -0.34559402727754784,
|
| 120 |
+
"rewards/accuracies": 0.71875,
|
| 121 |
+
"rewards/margins": 0.17012758103664966,
|
| 122 |
+
"logps/chosen": -58.15660923719406,
|
| 123 |
+
"logps/rejected": -58.908082497119906,
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"step": 50
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"loss": 0.5792030334472656,
|
| 129 |
+
"grad_norm": 0.10603554546833038,
|
| 130 |
+
"learning_rate": 9.528000000000001e-06,
|
| 131 |
+
"entropy": 0.7932545269140974,
|
| 132 |
+
"num_tokens": 806028.0,
|
| 133 |
+
"logits/chosen": -1.040485449115422,
|
| 134 |
+
"logits/rejected": -0.8688068371256747,
|
| 135 |
+
"mean_token_accuracy": 0.7223709706217051,
|
| 136 |
+
"rewards/chosen": -0.22816310470516327,
|
| 137 |
+
"rewards/rejected": -0.5157650222856318,
|
| 138 |
+
"rewards/accuracies": 0.75625,
|
| 139 |
+
"rewards/margins": 0.28760191788896916,
|
| 140 |
+
"logps/chosen": -57.834456419944765,
|
| 141 |
+
"logps/rejected": -54.969275665283206,
|
| 142 |
+
"epoch": 0.096,
|
| 143 |
+
"step": 60
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"loss": 0.5693038463592529,
|
| 147 |
+
"grad_norm": 0.13141675293445587,
|
| 148 |
+
"learning_rate": 9.448e-06,
|
| 149 |
+
"entropy": 0.7932507971301674,
|
| 150 |
+
"num_tokens": 940207.0,
|
| 151 |
+
"logits/chosen": -0.9470030204140493,
|
| 152 |
+
"logits/rejected": -0.8312474098211338,
|
| 153 |
+
"mean_token_accuracy": 0.7141493072733283,
|
| 154 |
+
"rewards/chosen": -0.38701098853271104,
|
| 155 |
+
"rewards/rejected": -0.7214406291190244,
|
| 156 |
+
"rewards/accuracies": 0.75,
|
| 157 |
+
"rewards/margins": 0.33442963876295834,
|
| 158 |
+
"logps/chosen": -57.73893855810165,
|
| 159 |
+
"logps/rejected": -59.55150935649872,
|
| 160 |
+
"epoch": 0.112,
|
| 161 |
+
"step": 70
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"loss": 0.5150018215179444,
|
| 165 |
+
"grad_norm": 0.12327121198177338,
|
| 166 |
+
"learning_rate": 9.368e-06,
|
| 167 |
+
"entropy": 0.8867666751611978,
|
| 168 |
+
"num_tokens": 1075324.0,
|
| 169 |
+
"logits/chosen": -0.8451518329835009,
|
| 170 |
+
"logits/rejected": -0.7344587441011935,
|
| 171 |
+
"mean_token_accuracy": 0.697543959133327,
|
| 172 |
+
"rewards/chosen": -0.4674373689107597,
|
| 173 |
+
"rewards/rejected": -0.9770577104296535,
|
| 174 |
+
"rewards/accuracies": 0.8125,
|
| 175 |
+
"rewards/margins": 0.5096203419845551,
|
| 176 |
+
"logps/chosen": -62.60571091175079,
|
| 177 |
+
"logps/rejected": -64.44814534187317,
|
| 178 |
+
"epoch": 0.128,
|
| 179 |
+
"step": 80
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"loss": 0.5565267086029053,
|
| 183 |
+
"grad_norm": 0.15871000289916992,
|
| 184 |
+
"learning_rate": 9.288e-06,
|
| 185 |
+
"entropy": 0.7912994117708877,
|
| 186 |
+
"num_tokens": 1210083.0,
|
| 187 |
+
"logits/chosen": -1.141903348240454,
|
| 188 |
+
"logits/rejected": -0.9025658047991666,
|
| 189 |
+
"mean_token_accuracy": 0.696558403223753,
|
| 190 |
+
"rewards/chosen": -0.6701661852537655,
|
| 191 |
+
"rewards/rejected": -1.1687032264657318,
|
| 192 |
+
"rewards/accuracies": 0.71875,
|
| 193 |
+
"rewards/margins": 0.49853704115375874,
|
| 194 |
+
"logps/chosen": -61.645342516899106,
|
| 195 |
+
"logps/rejected": -64.29151567220688,
|
| 196 |
+
"epoch": 0.144,
|
| 197 |
+
"step": 90
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"loss": 0.4938487529754639,
|
| 201 |
+
"grad_norm": 0.14912934601306915,
|
| 202 |
+
"learning_rate": 9.208e-06,
|
| 203 |
+
"entropy": 0.730810348119121,
|
| 204 |
+
"num_tokens": 1343049.0,
|
| 205 |
+
"logits/chosen": -1.320068194428384,
|
| 206 |
+
"logits/rejected": -0.9305093509728117,
|
| 207 |
+
"mean_token_accuracy": 0.7150107763707638,
|
| 208 |
+
"rewards/chosen": -0.7319991766475141,
|
| 209 |
+
"rewards/rejected": -1.3618548762926366,
|
| 210 |
+
"rewards/accuracies": 0.775,
|
| 211 |
+
"rewards/margins": 0.6298556990455836,
|
| 212 |
+
"logps/chosen": -47.61162393093109,
|
| 213 |
+
"logps/rejected": -63.62128868103027,
|
| 214 |
+
"epoch": 0.16,
|
| 215 |
+
"step": 100
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"loss": 0.5095012187957764,
|
| 219 |
+
"grad_norm": 0.19329985976219177,
|
| 220 |
+
"learning_rate": 9.128e-06,
|
| 221 |
+
"entropy": 0.767873613483971,
|
| 222 |
+
"num_tokens": 1477024.0,
|
| 223 |
+
"logits/chosen": -1.272959069252792,
|
| 224 |
+
"logits/rejected": -1.089739362694075,
|
| 225 |
+
"mean_token_accuracy": 0.7073799153789878,
|
| 226 |
+
"rewards/chosen": -0.9055291980272159,
|
| 227 |
+
"rewards/rejected": -1.6229321473278104,
|
| 228 |
+
"rewards/accuracies": 0.76875,
|
| 229 |
+
"rewards/margins": 0.7174029435962439,
|
| 230 |
+
"logps/chosen": -60.94918806552887,
|
| 231 |
+
"logps/rejected": -69.470434319973,
|
| 232 |
+
"epoch": 0.176,
|
| 233 |
+
"step": 110
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"loss": 0.4905715465545654,
|
| 237 |
+
"grad_norm": 0.1474754810333252,
|
| 238 |
+
"learning_rate": 9.048e-06,
|
| 239 |
+
"entropy": 0.7697478765854612,
|
| 240 |
+
"num_tokens": 1610835.0,
|
| 241 |
+
"logits/chosen": -1.2505374480949096,
|
| 242 |
+
"logits/rejected": -0.9930596218749196,
|
| 243 |
+
"mean_token_accuracy": 0.6980024211108684,
|
| 244 |
+
"rewards/chosen": -0.9649846871383488,
|
| 245 |
+
"rewards/rejected": -1.6785709237214177,
|
| 246 |
+
"rewards/accuracies": 0.8125,
|
| 247 |
+
"rewards/margins": 0.7135862361639738,
|
| 248 |
+
"logps/chosen": -60.80166560411453,
|
| 249 |
+
"logps/rejected": -67.57921611070633,
|
| 250 |
+
"epoch": 0.192,
|
| 251 |
+
"step": 120
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"loss": 0.5163051605224609,
|
| 255 |
+
"grad_norm": 0.1452934294939041,
|
| 256 |
+
"learning_rate": 8.968000000000001e-06,
|
| 257 |
+
"entropy": 0.8311835750704631,
|
| 258 |
+
"num_tokens": 1745826.0,
|
| 259 |
+
"logits/chosen": -1.1844953880498643,
|
| 260 |
+
"logits/rejected": -0.9830533050504325,
|
| 261 |
+
"mean_token_accuracy": 0.6839988188818097,
|
| 262 |
+
"rewards/chosen": -1.0836696863756514,
|
| 263 |
+
"rewards/rejected": -1.8168972163461148,
|
| 264 |
+
"rewards/accuracies": 0.775,
|
| 265 |
+
"rewards/margins": 0.7332275293767452,
|
| 266 |
+
"logps/chosen": -66.47507886886596,
|
| 267 |
+
"logps/rejected": -72.6287369966507,
|
| 268 |
+
"epoch": 0.208,
|
| 269 |
+
"step": 130
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"loss": 0.4344949722290039,
|
| 273 |
+
"grad_norm": 0.14412587881088257,
|
| 274 |
+
"learning_rate": 8.888e-06,
|
| 275 |
+
"entropy": 0.8209783418795269,
|
| 276 |
+
"num_tokens": 1880204.0,
|
| 277 |
+
"logits/chosen": -1.3630924177733654,
|
| 278 |
+
"logits/rejected": -1.0051007545666888,
|
| 279 |
+
"mean_token_accuracy": 0.6901104470714927,
|
| 280 |
+
"rewards/chosen": -1.0612068250833544,
|
| 281 |
+
"rewards/rejected": -2.040816684626043,
|
| 282 |
+
"rewards/accuracies": 0.81875,
|
| 283 |
+
"rewards/margins": 0.9796098598279059,
|
| 284 |
+
"logps/chosen": -64.26051410436631,
|
| 285 |
+
"logps/rejected": -75.64944772720337,
|
| 286 |
+
"epoch": 0.224,
|
| 287 |
+
"step": 140
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"loss": 0.43866467475891113,
|
| 291 |
+
"grad_norm": 0.11717677116394043,
|
| 292 |
+
"learning_rate": 8.808000000000001e-06,
|
| 293 |
+
"entropy": 0.7623928170069121,
|
| 294 |
+
"num_tokens": 2014712.0,
|
| 295 |
+
"logits/chosen": -1.2756038667122866,
|
| 296 |
+
"logits/rejected": -0.9577477036871918,
|
| 297 |
+
"mean_token_accuracy": 0.6972886839881539,
|
| 298 |
+
"rewards/chosen": -1.0979995009372943,
|
| 299 |
+
"rewards/rejected": -2.088811414106749,
|
| 300 |
+
"rewards/accuracies": 0.80625,
|
| 301 |
+
"rewards/margins": 0.9908119091764093,
|
| 302 |
+
"logps/chosen": -59.780848908424375,
|
| 303 |
+
"logps/rejected": -77.88960106372834,
|
| 304 |
+
"epoch": 0.24,
|
| 305 |
+
"step": 150
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"loss": 0.4745968818664551,
|
| 309 |
+
"grad_norm": 0.15543906390666962,
|
| 310 |
+
"learning_rate": 8.728e-06,
|
| 311 |
+
"entropy": 0.7725743347415118,
|
| 312 |
+
"num_tokens": 2148822.0,
|
| 313 |
+
"logits/chosen": -1.484901146092101,
|
| 314 |
+
"logits/rejected": -1.1511150508100232,
|
| 315 |
+
"mean_token_accuracy": 0.702755116764456,
|
| 316 |
+
"rewards/chosen": -1.2833676048379856,
|
| 317 |
+
"rewards/rejected": -2.2331671537132936,
|
| 318 |
+
"rewards/accuracies": 0.775,
|
| 319 |
+
"rewards/margins": 0.9497995547950268,
|
| 320 |
+
"logps/chosen": -66.71142189502716,
|
| 321 |
+
"logps/rejected": -73.32775690555573,
|
| 322 |
+
"epoch": 0.256,
|
| 323 |
+
"step": 160
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"loss": 0.5096800327301025,
|
| 327 |
+
"grad_norm": 0.17330285906791687,
|
| 328 |
+
"learning_rate": 8.648000000000001e-06,
|
| 329 |
+
"entropy": 0.7661999334814027,
|
| 330 |
+
"num_tokens": 2282595.0,
|
| 331 |
+
"logits/chosen": -1.3000051864690811,
|
| 332 |
+
"logits/rejected": -1.1044259002930392,
|
| 333 |
+
"mean_token_accuracy": 0.7065533282235265,
|
| 334 |
+
"rewards/chosen": -1.2363940134644509,
|
| 335 |
+
"rewards/rejected": -2.0990007982123644,
|
| 336 |
+
"rewards/accuracies": 0.725,
|
| 337 |
+
"rewards/margins": 0.86260677985847,
|
| 338 |
+
"logps/chosen": -66.75396176576615,
|
| 339 |
+
"logps/rejected": -70.63486285209656,
|
| 340 |
+
"epoch": 0.272,
|
| 341 |
+
"step": 170
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"loss": 0.44024109840393066,
|
| 345 |
+
"grad_norm": 0.11513521522283554,
|
| 346 |
+
"learning_rate": 8.568e-06,
|
| 347 |
+
"entropy": 0.791322482819669,
|
| 348 |
+
"num_tokens": 2418209.0,
|
| 349 |
+
"logits/chosen": -1.335577737989283,
|
| 350 |
+
"logits/rejected": -1.1066254954392574,
|
| 351 |
+
"mean_token_accuracy": 0.690339633077383,
|
| 352 |
+
"rewards/chosen": -1.3229667855077423,
|
| 353 |
+
"rewards/rejected": -2.4642039353144356,
|
| 354 |
+
"rewards/accuracies": 0.80625,
|
| 355 |
+
"rewards/margins": 1.1412371447309853,
|
| 356 |
+
"logps/chosen": -72.83452732563019,
|
| 357 |
+
"logps/rejected": -85.12001547813415,
|
| 358 |
+
"epoch": 0.288,
|
| 359 |
+
"step": 180
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"loss": 0.504111909866333,
|
| 363 |
+
"grad_norm": 0.156762033700943,
|
| 364 |
+
"learning_rate": 8.488e-06,
|
| 365 |
+
"entropy": 0.8087703323311871,
|
| 366 |
+
"num_tokens": 2552678.0,
|
| 367 |
+
"logits/chosen": -1.2479021492111673,
|
| 368 |
+
"logits/rejected": -1.1869599471548637,
|
| 369 |
+
"mean_token_accuracy": 0.673186760675162,
|
| 370 |
+
"rewards/chosen": -1.2775754334405065,
|
| 371 |
+
"rewards/rejected": -2.152550506545231,
|
| 372 |
+
"rewards/accuracies": 0.7625,
|
| 373 |
+
"rewards/margins": 0.8749750746414066,
|
| 374 |
+
"logps/chosen": -72.5224678516388,
|
| 375 |
+
"logps/rejected": -72.16854902505875,
|
| 376 |
+
"epoch": 0.304,
|
| 377 |
+
"step": 190
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"loss": 0.5158682346343995,
|
| 381 |
+
"grad_norm": 0.12845639884471893,
|
| 382 |
+
"learning_rate": 8.408e-06,
|
| 383 |
+
"entropy": 0.7313262579322327,
|
| 384 |
+
"num_tokens": 2686349.0,
|
| 385 |
+
"logits/chosen": -1.3158707246374635,
|
| 386 |
+
"logits/rejected": -1.1155622394775986,
|
| 387 |
+
"mean_token_accuracy": 0.7055080824531614,
|
| 388 |
+
"rewards/chosen": -1.1535688072181074,
|
| 389 |
+
"rewards/rejected": -1.949167560931528,
|
| 390 |
+
"rewards/accuracies": 0.75,
|
| 391 |
+
"rewards/margins": 0.7955987584777177,
|
| 392 |
+
"logps/chosen": -66.22938264608383,
|
| 393 |
+
"logps/rejected": -67.70417716503144,
|
| 394 |
+
"epoch": 0.32,
|
| 395 |
+
"step": 200
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"loss": 0.3683095693588257,
|
| 399 |
+
"grad_norm": 0.15806466341018677,
|
| 400 |
+
"learning_rate": 8.328e-06,
|
| 401 |
+
"entropy": 0.7535996982711367,
|
| 402 |
+
"num_tokens": 2821059.0,
|
| 403 |
+
"logits/chosen": -1.5549069948264784,
|
| 404 |
+
"logits/rejected": -1.1385768780607908,
|
| 405 |
+
"mean_token_accuracy": 0.7171918565407396,
|
| 406 |
+
"rewards/chosen": -0.9996276974328794,
|
| 407 |
+
"rewards/rejected": -2.2987426282372327,
|
| 408 |
+
"rewards/accuracies": 0.85,
|
| 409 |
+
"rewards/margins": 1.299114930210635,
|
| 410 |
+
"logps/chosen": -58.6371146440506,
|
| 411 |
+
"logps/rejected": -80.23583896160126,
|
| 412 |
+
"epoch": 0.336,
|
| 413 |
+
"step": 210
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"loss": 0.3892038106918335,
|
| 417 |
+
"grad_norm": 0.19750693440437317,
|
| 418 |
+
"learning_rate": 8.248e-06,
|
| 419 |
+
"entropy": 0.6967509233363671,
|
| 420 |
+
"num_tokens": 2954598.0,
|
| 421 |
+
"logits/chosen": -1.601990247352082,
|
| 422 |
+
"logits/rejected": -1.2623563139239278,
|
| 423 |
+
"mean_token_accuracy": 0.708772998303175,
|
| 424 |
+
"rewards/chosen": -1.0824068963585887,
|
| 425 |
+
"rewards/rejected": -2.34019776461646,
|
| 426 |
+
"rewards/accuracies": 0.84375,
|
| 427 |
+
"rewards/margins": 1.2577908653765917,
|
| 428 |
+
"logps/chosen": -59.04450926780701,
|
| 429 |
+
"logps/rejected": -72.98424959182739,
|
| 430 |
+
"epoch": 0.352,
|
| 431 |
+
"step": 220
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"loss": 0.4733689785003662,
|
| 435 |
+
"grad_norm": 0.17226403951644897,
|
| 436 |
+
"learning_rate": 8.168e-06,
|
| 437 |
+
"entropy": 0.657842610750231,
|
| 438 |
+
"num_tokens": 3087946.0,
|
| 439 |
+
"logits/chosen": -1.6329175719438567,
|
| 440 |
+
"logits/rejected": -1.3952135735634263,
|
| 441 |
+
"mean_token_accuracy": 0.714717129431665,
|
| 442 |
+
"rewards/chosen": -1.212909367471002,
|
| 443 |
+
"rewards/rejected": -2.3671575023792686,
|
| 444 |
+
"rewards/accuracies": 0.775,
|
| 445 |
+
"rewards/margins": 1.1542481290176512,
|
| 446 |
+
"logps/chosen": -58.4173419713974,
|
| 447 |
+
"logps/rejected": -71.25410054922104,
|
| 448 |
+
"epoch": 0.368,
|
| 449 |
+
"step": 230
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"loss": 0.35283257961273196,
|
| 453 |
+
"grad_norm": 0.12192272394895554,
|
| 454 |
+
"learning_rate": 8.088e-06,
|
| 455 |
+
"entropy": 0.7537508669774979,
|
| 456 |
+
"num_tokens": 3221706.0,
|
| 457 |
+
"logits/chosen": -1.4900533239242644,
|
| 458 |
+
"logits/rejected": -1.2093497528084223,
|
| 459 |
+
"mean_token_accuracy": 0.7051469705067575,
|
| 460 |
+
"rewards/chosen": -1.1929488512105308,
|
| 461 |
+
"rewards/rejected": -2.568997914995998,
|
| 462 |
+
"rewards/accuracies": 0.9,
|
| 463 |
+
"rewards/margins": 1.376049063075334,
|
| 464 |
+
"logps/chosen": -59.28422422409058,
|
| 465 |
+
"logps/rejected": -77.90960043668747,
|
| 466 |
+
"epoch": 0.384,
|
| 467 |
+
"step": 240
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"loss": 0.4659436702728271,
|
| 471 |
+
"grad_norm": 0.22347807884216309,
|
| 472 |
+
"learning_rate": 8.008e-06,
|
| 473 |
+
"entropy": 0.8218240110814804,
|
| 474 |
+
"num_tokens": 3356185.0,
|
| 475 |
+
"logits/chosen": -1.3824732789417833,
|
| 476 |
+
"logits/rejected": -1.1605517896529085,
|
| 477 |
+
"mean_token_accuracy": 0.6925541674718261,
|
| 478 |
+
"rewards/chosen": -1.3342948321445873,
|
| 479 |
+
"rewards/rejected": -2.4782243948429823,
|
| 480 |
+
"rewards/accuracies": 0.8,
|
| 481 |
+
"rewards/margins": 1.1439295556396245,
|
| 482 |
+
"logps/chosen": -73.3686304807663,
|
| 483 |
+
"logps/rejected": -74.8317013144493,
|
| 484 |
+
"epoch": 0.4,
|
| 485 |
+
"step": 250
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"loss": 0.4091994285583496,
|
| 489 |
+
"grad_norm": 0.20474763214588165,
|
| 490 |
+
"learning_rate": 7.928e-06,
|
| 491 |
+
"entropy": 0.8001895788867841,
|
| 492 |
+
"num_tokens": 3490468.0,
|
| 493 |
+
"logits/chosen": -1.3735195504245614,
|
| 494 |
+
"logits/rejected": -1.16435433493973,
|
| 495 |
+
"mean_token_accuracy": 0.6946983626112342,
|
| 496 |
+
"rewards/chosen": -1.2219555595249403,
|
| 497 |
+
"rewards/rejected": -2.5509443364106117,
|
| 498 |
+
"rewards/accuracies": 0.84375,
|
| 499 |
+
"rewards/margins": 1.328988773515448,
|
| 500 |
+
"logps/chosen": -68.65264060497285,
|
| 501 |
+
"logps/rejected": -80.02137950658798,
|
| 502 |
+
"epoch": 0.416,
|
| 503 |
+
"step": 260
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"loss": 0.3889183521270752,
|
| 507 |
+
"grad_norm": 0.19864033162593842,
|
| 508 |
+
"learning_rate": 7.848000000000002e-06,
|
| 509 |
+
"entropy": 0.7910746831679717,
|
| 510 |
+
"num_tokens": 3625248.0,
|
| 511 |
+
"logits/chosen": -1.457904188026588,
|
| 512 |
+
"logits/rejected": -1.158327298867405,
|
| 513 |
+
"mean_token_accuracy": 0.6931218519806862,
|
| 514 |
+
"rewards/chosen": -1.2410434904973955,
|
| 515 |
+
"rewards/rejected": -2.4440282738534735,
|
| 516 |
+
"rewards/accuracies": 0.8375,
|
| 517 |
+
"rewards/margins": 1.2029847822152078,
|
| 518 |
+
"logps/chosen": -70.0272049665451,
|
| 519 |
+
"logps/rejected": -79.16600930690765,
|
| 520 |
+
"epoch": 0.432,
|
| 521 |
+
"step": 270
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"loss": 0.3985002040863037,
|
| 525 |
+
"grad_norm": 0.19505871832370758,
|
| 526 |
+
"learning_rate": 7.768e-06,
|
| 527 |
+
"entropy": 0.7678435662121046,
|
| 528 |
+
"num_tokens": 3759703.0,
|
| 529 |
+
"logits/chosen": -1.6367592534525586,
|
| 530 |
+
"logits/rejected": -1.353649134707537,
|
| 531 |
+
"mean_token_accuracy": 0.7048457863740623,
|
| 532 |
+
"rewards/chosen": -1.180449191952357,
|
| 533 |
+
"rewards/rejected": -2.5172444540075958,
|
| 534 |
+
"rewards/accuracies": 0.83125,
|
| 535 |
+
"rewards/margins": 1.3367952583008447,
|
| 536 |
+
"logps/chosen": -65.0837183713913,
|
| 537 |
+
"logps/rejected": -79.13848531246185,
|
| 538 |
+
"epoch": 0.448,
|
| 539 |
+
"step": 280
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"loss": 0.4443049430847168,
|
| 543 |
+
"grad_norm": 0.17939473688602448,
|
| 544 |
+
"learning_rate": 7.688000000000002e-06,
|
| 545 |
+
"entropy": 0.726059359870851,
|
| 546 |
+
"num_tokens": 3893200.0,
|
| 547 |
+
"logits/chosen": -1.6981963699206202,
|
| 548 |
+
"logits/rejected": -1.4103382131432336,
|
| 549 |
+
"mean_token_accuracy": 0.7085827089846134,
|
| 550 |
+
"rewards/chosen": -1.1276641700067558,
|
| 551 |
+
"rewards/rejected": -2.3673765807703604,
|
| 552 |
+
"rewards/accuracies": 0.78125,
|
| 553 |
+
"rewards/margins": 1.2397124132607131,
|
| 554 |
+
"logps/chosen": -60.80485708713532,
|
| 555 |
+
"logps/rejected": -71.46178147792816,
|
| 556 |
+
"epoch": 0.464,
|
| 557 |
+
"step": 290
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"loss": 0.4806370258331299,
|
| 561 |
+
"grad_norm": 0.19454504549503326,
|
| 562 |
+
"learning_rate": 7.608000000000001e-06,
|
| 563 |
+
"entropy": 0.7791819364763797,
|
| 564 |
+
"num_tokens": 4027530.0,
|
| 565 |
+
"logits/chosen": -1.6851057889767016,
|
| 566 |
+
"logits/rejected": -1.2973655587521755,
|
| 567 |
+
"mean_token_accuracy": 0.7038618138059973,
|
| 568 |
+
"rewards/chosen": -1.2468871754863358,
|
| 569 |
+
"rewards/rejected": -2.4119768414646385,
|
| 570 |
+
"rewards/accuracies": 0.8,
|
| 571 |
+
"rewards/margins": 1.165089663118124,
|
| 572 |
+
"logps/chosen": -66.0106352686882,
|
| 573 |
+
"logps/rejected": -78.41197410821914,
|
| 574 |
+
"epoch": 0.48,
|
| 575 |
+
"step": 300
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"loss": 0.37413773536682127,
|
| 579 |
+
"grad_norm": 0.0995570570230484,
|
| 580 |
+
"learning_rate": 7.528000000000001e-06,
|
| 581 |
+
"entropy": 0.7844068950158544,
|
| 582 |
+
"num_tokens": 4162061.0,
|
| 583 |
+
"logits/chosen": -1.4213479846202168,
|
| 584 |
+
"logits/rejected": -1.1986241342964326,
|
| 585 |
+
"mean_token_accuracy": 0.6761405906639993,
|
| 586 |
+
"rewards/chosen": -1.0949653455951194,
|
| 587 |
+
"rewards/rejected": -2.420564866659697,
|
| 588 |
+
"rewards/accuracies": 0.85,
|
| 589 |
+
"rewards/margins": 1.3255995167419314,
|
| 590 |
+
"logps/chosen": -68.37061663866044,
|
| 591 |
+
"logps/rejected": -75.29490267038345,
|
| 592 |
+
"epoch": 0.496,
|
| 593 |
+
"step": 310
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"loss": 0.4030296802520752,
|
| 597 |
+
"grad_norm": 0.28332963585853577,
|
| 598 |
+
"learning_rate": 7.4480000000000005e-06,
|
| 599 |
+
"entropy": 0.7986679952475242,
|
| 600 |
+
"num_tokens": 4295767.0,
|
| 601 |
+
"logits/chosen": -1.5688449783205605,
|
| 602 |
+
"logits/rejected": -1.3472575925643246,
|
| 603 |
+
"mean_token_accuracy": 0.6999106446281076,
|
| 604 |
+
"rewards/chosen": -1.0604232914178282,
|
| 605 |
+
"rewards/rejected": -2.3024212238378823,
|
| 606 |
+
"rewards/accuracies": 0.85,
|
| 607 |
+
"rewards/margins": 1.2419979333877564,
|
| 608 |
+
"logps/chosen": -64.41045069694519,
|
| 609 |
+
"logps/rejected": -71.96095795631409,
|
| 610 |
+
"epoch": 0.512,
|
| 611 |
+
"step": 320
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"loss": 0.4251255512237549,
|
| 615 |
+
"grad_norm": 0.21712468564510345,
|
| 616 |
+
"learning_rate": 7.3680000000000004e-06,
|
| 617 |
+
"entropy": 0.771777928317897,
|
| 618 |
+
"num_tokens": 4430319.0,
|
| 619 |
+
"logits/chosen": -1.554949818555905,
|
| 620 |
+
"logits/rejected": -1.2374624330985304,
|
| 621 |
+
"mean_token_accuracy": 0.7228127410635352,
|
| 622 |
+
"rewards/chosen": -1.2476483421865852,
|
| 623 |
+
"rewards/rejected": -2.5015724537312054,
|
| 624 |
+
"rewards/accuracies": 0.79375,
|
| 625 |
+
"rewards/margins": 1.2539241114631294,
|
| 626 |
+
"logps/chosen": -66.54614742994309,
|
| 627 |
+
"logps/rejected": -82.05639437437057,
|
| 628 |
+
"epoch": 0.528,
|
| 629 |
+
"step": 330
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"loss": 0.43709683418273926,
|
| 633 |
+
"grad_norm": 0.17526720464229584,
|
| 634 |
+
"learning_rate": 7.288e-06,
|
| 635 |
+
"entropy": 0.788643982866779,
|
| 636 |
+
"num_tokens": 4565294.0,
|
| 637 |
+
"logits/chosen": -1.6649492613392822,
|
| 638 |
+
"logits/rejected": -1.3661411110846526,
|
| 639 |
+
"mean_token_accuracy": 0.683904533740133,
|
| 640 |
+
"rewards/chosen": -1.1670136523724068,
|
| 641 |
+
"rewards/rejected": -2.5708392354194074,
|
| 642 |
+
"rewards/accuracies": 0.825,
|
| 643 |
+
"rewards/margins": 1.403825580328703,
|
| 644 |
+
"logps/chosen": -65.53375207185745,
|
| 645 |
+
"logps/rejected": -82.1581961631775,
|
| 646 |
+
"epoch": 0.544,
|
| 647 |
+
"step": 340
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"loss": 0.35862901210784914,
|
| 651 |
+
"grad_norm": 0.1822371631860733,
|
| 652 |
+
"learning_rate": 7.208e-06,
|
| 653 |
+
"entropy": 0.8551028195535764,
|
| 654 |
+
"num_tokens": 4700673.0,
|
| 655 |
+
"logits/chosen": -1.456482774999826,
|
| 656 |
+
"logits/rejected": -1.1154920837633087,
|
| 657 |
+
"mean_token_accuracy": 0.6865572449751198,
|
| 658 |
+
"rewards/chosen": -1.0634017121992656,
|
| 659 |
+
"rewards/rejected": -2.5696225532796233,
|
| 660 |
+
"rewards/accuracies": 0.875,
|
| 661 |
+
"rewards/margins": 1.5062208444345742,
|
| 662 |
+
"logps/chosen": -68.0031861782074,
|
| 663 |
+
"logps/rejected": -86.1665113568306,
|
| 664 |
+
"epoch": 0.56,
|
| 665 |
+
"step": 350
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"loss": 0.3786123752593994,
|
| 669 |
+
"grad_norm": 0.13778463006019592,
|
| 670 |
+
"learning_rate": 7.128e-06,
|
| 671 |
+
"entropy": 0.7678775020700414,
|
| 672 |
+
"num_tokens": 4834914.0,
|
| 673 |
+
"logits/chosen": -1.6450069016663857,
|
| 674 |
+
"logits/rejected": -1.3522691083362712,
|
| 675 |
+
"mean_token_accuracy": 0.691734860651195,
|
| 676 |
+
"rewards/chosen": -1.1923491157358512,
|
| 677 |
+
"rewards/rejected": -2.65358635247685,
|
| 678 |
+
"rewards/accuracies": 0.8375,
|
| 679 |
+
"rewards/margins": 1.4612372373230755,
|
| 680 |
+
"logps/chosen": -63.21224284172058,
|
| 681 |
+
"logps/rejected": -77.80383843183517,
|
| 682 |
+
"epoch": 0.576,
|
| 683 |
+
"step": 360
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"loss": 0.40740342140197755,
|
| 687 |
+
"grad_norm": 0.11445042490959167,
|
| 688 |
+
"learning_rate": 7.048e-06,
|
| 689 |
+
"entropy": 0.7982966381299775,
|
| 690 |
+
"num_tokens": 4969481.0,
|
| 691 |
+
"logits/chosen": -1.5110376075863712,
|
| 692 |
+
"logits/rejected": -1.3656208793349927,
|
| 693 |
+
"mean_token_accuracy": 0.6955559890717268,
|
| 694 |
+
"rewards/chosen": -1.324278107311693,
|
| 695 |
+
"rewards/rejected": -2.7373441014206037,
|
| 696 |
+
"rewards/accuracies": 0.83125,
|
| 697 |
+
"rewards/margins": 1.4130659931339324,
|
| 698 |
+
"logps/chosen": -72.74515337944031,
|
| 699 |
+
"logps/rejected": -75.52241103649139,
|
| 700 |
+
"epoch": 0.592,
|
| 701 |
+
"step": 370
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"loss": 0.4049683570861816,
|
| 705 |
+
"grad_norm": 0.23570746183395386,
|
| 706 |
+
"learning_rate": 6.968e-06,
|
| 707 |
+
"entropy": 0.7056236127449665,
|
| 708 |
+
"num_tokens": 5104134.0,
|
| 709 |
+
"logits/chosen": -1.798725774451439,
|
| 710 |
+
"logits/rejected": -1.365983592775875,
|
| 711 |
+
"mean_token_accuracy": 0.7301392247900367,
|
| 712 |
+
"rewards/chosen": -1.2776403030846268,
|
| 713 |
+
"rewards/rejected": -2.910024182451889,
|
| 714 |
+
"rewards/accuracies": 0.83125,
|
| 715 |
+
"rewards/margins": 1.6323838772717862,
|
| 716 |
+
"logps/chosen": -64.17630796432495,
|
| 717 |
+
"logps/rejected": -82.18561379909515,
|
| 718 |
+
"epoch": 0.608,
|
| 719 |
+
"step": 380
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"loss": 0.3106002569198608,
|
| 723 |
+
"grad_norm": 0.1947896033525467,
|
| 724 |
+
"learning_rate": 6.888e-06,
|
| 725 |
+
"entropy": 0.8436481792770791,
|
| 726 |
+
"num_tokens": 5238990.0,
|
| 727 |
+
"logits/chosen": -1.5257286904400196,
|
| 728 |
+
"logits/rejected": -1.211996810178077,
|
| 729 |
+
"mean_token_accuracy": 0.7017000323161483,
|
| 730 |
+
"rewards/chosen": -1.093766995461192,
|
| 731 |
+
"rewards/rejected": -2.8707397868623956,
|
| 732 |
+
"rewards/accuracies": 0.86875,
|
| 733 |
+
"rewards/margins": 1.7769727816805243,
|
| 734 |
+
"logps/chosen": -74.20939913988113,
|
| 735 |
+
"logps/rejected": -81.55687050819397,
|
| 736 |
+
"epoch": 0.624,
|
| 737 |
+
"step": 390
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"loss": 0.388614296913147,
|
| 741 |
+
"grad_norm": 0.1308419406414032,
|
| 742 |
+
"learning_rate": 6.808e-06,
|
| 743 |
+
"entropy": 0.8186127942055463,
|
| 744 |
+
"num_tokens": 5373007.0,
|
| 745 |
+
"logits/chosen": -1.6899056237717702,
|
| 746 |
+
"logits/rejected": -1.262867669279399,
|
| 747 |
+
"mean_token_accuracy": 0.7177292765118182,
|
| 748 |
+
"rewards/chosen": -1.1862425664090552,
|
| 749 |
+
"rewards/rejected": -2.6863625437486918,
|
| 750 |
+
"rewards/accuracies": 0.8375,
|
| 751 |
+
"rewards/margins": 1.5001199764199555,
|
| 752 |
+
"logps/chosen": -65.6905270934105,
|
| 753 |
+
"logps/rejected": -76.37971225976943,
|
| 754 |
+
"epoch": 0.64,
|
| 755 |
+
"step": 400
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"loss": 0.3734513998031616,
|
| 759 |
+
"grad_norm": 0.14931273460388184,
|
| 760 |
+
"learning_rate": 6.728e-06,
|
| 761 |
+
"entropy": 0.8112484578086878,
|
| 762 |
+
"num_tokens": 5507998.0,
|
| 763 |
+
"logits/chosen": -1.4861246998222568,
|
| 764 |
+
"logits/rejected": -1.2152247862453383,
|
| 765 |
+
"mean_token_accuracy": 0.7157961593009532,
|
| 766 |
+
"rewards/chosen": -1.2374363817856646,
|
| 767 |
+
"rewards/rejected": -2.7955031143967064,
|
| 768 |
+
"rewards/accuracies": 0.85,
|
| 769 |
+
"rewards/margins": 1.5580667286179959,
|
| 770 |
+
"logps/chosen": -69.05660092830658,
|
| 771 |
+
"logps/rejected": -83.02244206666947,
|
| 772 |
+
"epoch": 0.656,
|
| 773 |
+
"step": 410
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"loss": 0.40374464988708497,
|
| 777 |
+
"grad_norm": 0.1771688610315323,
|
| 778 |
+
"learning_rate": 6.648e-06,
|
| 779 |
+
"entropy": 0.8456664968209224,
|
| 780 |
+
"num_tokens": 5643003.0,
|
| 781 |
+
"logits/chosen": -1.4466347163058084,
|
| 782 |
+
"logits/rejected": -1.1498763927681597,
|
| 783 |
+
"mean_token_accuracy": 0.706757742818445,
|
| 784 |
+
"rewards/chosen": -1.130590606055921,
|
| 785 |
+
"rewards/rejected": -2.6461291053332388,
|
| 786 |
+
"rewards/accuracies": 0.83125,
|
| 787 |
+
"rewards/margins": 1.5155384978279471,
|
| 788 |
+
"logps/chosen": -73.05048446655273,
|
| 789 |
+
"logps/rejected": -78.86939449310303,
|
| 790 |
+
"epoch": 0.672,
|
| 791 |
+
"step": 420
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"loss": 0.41454453468322755,
|
| 795 |
+
"grad_norm": 0.2073439657688141,
|
| 796 |
+
"learning_rate": 6.568000000000001e-06,
|
| 797 |
+
"entropy": 0.916799544049718,
|
| 798 |
+
"num_tokens": 5778181.0,
|
| 799 |
+
"logits/chosen": -1.4833107915614325,
|
| 800 |
+
"logits/rejected": -1.112710293552378,
|
| 801 |
+
"mean_token_accuracy": 0.7104185981675982,
|
| 802 |
+
"rewards/chosen": -0.813328304514289,
|
| 803 |
+
"rewards/rejected": -2.2752998224226757,
|
| 804 |
+
"rewards/accuracies": 0.83125,
|
| 805 |
+
"rewards/margins": 1.4619715091772378,
|
| 806 |
+
"logps/chosen": -68.41654760837555,
|
| 807 |
+
"logps/rejected": -78.04829220771789,
|
| 808 |
+
"epoch": 0.688,
|
| 809 |
+
"step": 430
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"loss": 0.42249112129211425,
|
| 813 |
+
"grad_norm": 0.13397125899791718,
|
| 814 |
+
"learning_rate": 6.488000000000001e-06,
|
| 815 |
+
"entropy": 0.7896173141780309,
|
| 816 |
+
"num_tokens": 5912330.0,
|
| 817 |
+
"logits/chosen": -1.681157098237107,
|
| 818 |
+
"logits/rejected": -1.427799018524135,
|
| 819 |
+
"mean_token_accuracy": 0.7210862101987005,
|
| 820 |
+
"rewards/chosen": -0.9002718864590861,
|
| 821 |
+
"rewards/rejected": -2.3048052706988527,
|
| 822 |
+
"rewards/accuracies": 0.81875,
|
| 823 |
+
"rewards/margins": 1.4045333787798882,
|
| 824 |
+
"logps/chosen": -64.21181601285934,
|
| 825 |
+
"logps/rejected": -71.62155838012696,
|
| 826 |
+
"epoch": 0.704,
|
| 827 |
+
"step": 440
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"loss": 0.38965840339660646,
|
| 831 |
+
"grad_norm": 0.26196277141571045,
|
| 832 |
+
"learning_rate": 6.408000000000001e-06,
|
| 833 |
+
"entropy": 0.8554482666426338,
|
| 834 |
+
"num_tokens": 6046833.0,
|
| 835 |
+
"logits/chosen": -1.6261219666437803,
|
| 836 |
+
"logits/rejected": -1.3941636311101144,
|
| 837 |
+
"mean_token_accuracy": 0.6925984876230359,
|
| 838 |
+
"rewards/chosen": -0.7953715628435021,
|
| 839 |
+
"rewards/rejected": -2.3454553585266695,
|
| 840 |
+
"rewards/accuracies": 0.825,
|
| 841 |
+
"rewards/margins": 1.5500838007777928,
|
| 842 |
+
"logps/chosen": -69.06866861581803,
|
| 843 |
+
"logps/rejected": -75.28483995199204,
|
| 844 |
+
"epoch": 0.72,
|
| 845 |
+
"step": 450
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"loss": 0.4335937976837158,
|
| 849 |
+
"grad_norm": 0.14940954744815826,
|
| 850 |
+
"learning_rate": 6.328000000000001e-06,
|
| 851 |
+
"entropy": 0.7710782749025384,
|
| 852 |
+
"num_tokens": 6180736.0,
|
| 853 |
+
"logits/chosen": -1.7704391814301037,
|
| 854 |
+
"logits/rejected": -1.471645080915709,
|
| 855 |
+
"mean_token_accuracy": 0.7167778188362718,
|
| 856 |
+
"rewards/chosen": -0.789050264639991,
|
| 857 |
+
"rewards/rejected": -2.1151811987860127,
|
| 858 |
+
"rewards/accuracies": 0.79375,
|
| 859 |
+
"rewards/margins": 1.3261309264693408,
|
| 860 |
+
"logps/chosen": -59.20472636222839,
|
| 861 |
+
"logps/rejected": -72.96591898202897,
|
| 862 |
+
"epoch": 0.736,
|
| 863 |
+
"step": 460
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"loss": 0.3879385948181152,
|
| 867 |
+
"grad_norm": 0.19128943979740143,
|
| 868 |
+
"learning_rate": 6.248000000000001e-06,
|
| 869 |
+
"entropy": 0.8313533828346408,
|
| 870 |
+
"num_tokens": 6314772.0,
|
| 871 |
+
"logits/chosen": -1.7308567879554204,
|
| 872 |
+
"logits/rejected": -1.394286283449892,
|
| 873 |
+
"mean_token_accuracy": 0.7202207330614329,
|
| 874 |
+
"rewards/chosen": -0.7537071728525916,
|
| 875 |
+
"rewards/rejected": -2.1371329107510975,
|
| 876 |
+
"rewards/accuracies": 0.84375,
|
| 877 |
+
"rewards/margins": 1.3834257372189314,
|
| 878 |
+
"logps/chosen": -59.07860852479935,
|
| 879 |
+
"logps/rejected": -74.81651721000671,
|
| 880 |
+
"epoch": 0.752,
|
| 881 |
+
"step": 470
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"loss": 0.3739017009735107,
|
| 885 |
+
"grad_norm": 0.21433700621128082,
|
| 886 |
+
"learning_rate": 6.168000000000001e-06,
|
| 887 |
+
"entropy": 0.7569797153497347,
|
| 888 |
+
"num_tokens": 6449350.0,
|
| 889 |
+
"logits/chosen": -1.8720855768489677,
|
| 890 |
+
"logits/rejected": -1.3812638967944575,
|
| 891 |
+
"mean_token_accuracy": 0.718151222448796,
|
| 892 |
+
"rewards/chosen": -0.9107469652291911,
|
| 893 |
+
"rewards/rejected": -2.5364978435100056,
|
| 894 |
+
"rewards/accuracies": 0.8375,
|
| 895 |
+
"rewards/margins": 1.6257508777081966,
|
| 896 |
+
"logps/chosen": -57.991915786266325,
|
| 897 |
+
"logps/rejected": -79.13743947744369,
|
| 898 |
+
"epoch": 0.768,
|
| 899 |
+
"step": 480
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"loss": 0.42026190757751464,
|
| 903 |
+
"grad_norm": 0.19288769364356995,
|
| 904 |
+
"learning_rate": 6.088000000000001e-06,
|
| 905 |
+
"entropy": 0.8451594576821663,
|
| 906 |
+
"num_tokens": 6584446.0,
|
| 907 |
+
"logits/chosen": -1.6996204679132532,
|
| 908 |
+
"logits/rejected": -1.4186780551259308,
|
| 909 |
+
"mean_token_accuracy": 0.7145285487174988,
|
| 910 |
+
"rewards/chosen": -0.972397522853862,
|
| 911 |
+
"rewards/rejected": -2.618794380797772,
|
| 912 |
+
"rewards/accuracies": 0.79375,
|
| 913 |
+
"rewards/margins": 1.6463968583382667,
|
| 914 |
+
"logps/chosen": -65.44049211740494,
|
| 915 |
+
"logps/rejected": -82.9235647559166,
|
| 916 |
+
"epoch": 0.784,
|
| 917 |
+
"step": 490
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"loss": 0.49308347702026367,
|
| 921 |
+
"grad_norm": 0.2881377637386322,
|
| 922 |
+
"learning_rate": 6.008000000000001e-06,
|
| 923 |
+
"entropy": 0.8130768362927483,
|
| 924 |
+
"num_tokens": 6719638.0,
|
| 925 |
+
"logits/chosen": -1.5676580468108416,
|
| 926 |
+
"logits/rejected": -1.419498501372546,
|
| 927 |
+
"mean_token_accuracy": 0.6852236403152346,
|
| 928 |
+
"rewards/chosen": -1.1576830506091937,
|
| 929 |
+
"rewards/rejected": -2.454260028176941,
|
| 930 |
+
"rewards/accuracies": 0.7875,
|
| 931 |
+
"rewards/margins": 1.2965769773349165,
|
| 932 |
+
"logps/chosen": -73.50958200693131,
|
| 933 |
+
"logps/rejected": -75.68365123271943,
|
| 934 |
+
"epoch": 0.8,
|
| 935 |
+
"step": 500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"loss": 0.4143951416015625,
|
| 939 |
+
"grad_norm": 0.19303563237190247,
|
| 940 |
+
"learning_rate": 5.928000000000001e-06,
|
| 941 |
+
"entropy": 0.8353224259713897,
|
| 942 |
+
"num_tokens": 6853713.0,
|
| 943 |
+
"logits/chosen": -1.4901775022155148,
|
| 944 |
+
"logits/rejected": -1.2796557225262943,
|
| 945 |
+
"mean_token_accuracy": 0.6960680657066405,
|
| 946 |
+
"rewards/chosen": -1.100995674492151,
|
| 947 |
+
"rewards/rejected": -2.624603364392533,
|
| 948 |
+
"rewards/accuracies": 0.79375,
|
| 949 |
+
"rewards/margins": 1.5236076795030384,
|
| 950 |
+
"logps/chosen": -65.34812902212143,
|
| 951 |
+
"logps/rejected": -78.05482840538025,
|
| 952 |
+
"epoch": 0.816,
|
| 953 |
+
"step": 510
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"loss": 0.4617034435272217,
|
| 957 |
+
"grad_norm": 0.253850519657135,
|
| 958 |
+
"learning_rate": 5.848000000000001e-06,
|
| 959 |
+
"entropy": 0.7824866577997455,
|
| 960 |
+
"num_tokens": 6988253.0,
|
| 961 |
+
"logits/chosen": -1.6970046056735602,
|
| 962 |
+
"logits/rejected": -1.3818239300070876,
|
| 963 |
+
"mean_token_accuracy": 0.7068752739578485,
|
| 964 |
+
"rewards/chosen": -1.1414498813042884,
|
| 965 |
+
"rewards/rejected": -2.5120410541421734,
|
| 966 |
+
"rewards/accuracies": 0.8,
|
| 967 |
+
"rewards/margins": 1.3705911776050925,
|
| 968 |
+
"logps/chosen": -63.807918751239775,
|
| 969 |
+
"logps/rejected": -77.26764396429061,
|
| 970 |
+
"epoch": 0.832,
|
| 971 |
+
"step": 520
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"loss": 0.41399874687194826,
|
| 975 |
+
"grad_norm": 0.39896082878112793,
|
| 976 |
+
"learning_rate": 5.7680000000000005e-06,
|
| 977 |
+
"entropy": 0.8119747741817264,
|
| 978 |
+
"num_tokens": 7122655.0,
|
| 979 |
+
"logits/chosen": -1.5383035927580562,
|
| 980 |
+
"logits/rejected": -1.1577523367736098,
|
| 981 |
+
"mean_token_accuracy": 0.715251567773521,
|
| 982 |
+
"rewards/chosen": -1.1019727620776394,
|
| 983 |
+
"rewards/rejected": -2.5670429018791765,
|
| 984 |
+
"rewards/accuracies": 0.80625,
|
| 985 |
+
"rewards/margins": 1.4650701294653117,
|
| 986 |
+
"logps/chosen": -58.73834667801857,
|
| 987 |
+
"logps/rejected": -80.83378657102585,
|
| 988 |
+
"epoch": 0.848,
|
| 989 |
+
"step": 530
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"loss": 0.40615096092224123,
|
| 993 |
+
"grad_norm": 0.24633856117725372,
|
| 994 |
+
"learning_rate": 5.6880000000000004e-06,
|
| 995 |
+
"entropy": 0.774302898303722,
|
| 996 |
+
"num_tokens": 7256116.0,
|
| 997 |
+
"logits/chosen": -1.784429702455147,
|
| 998 |
+
"logits/rejected": -1.3702712726028023,
|
| 999 |
+
"mean_token_accuracy": 0.7210501914843916,
|
| 1000 |
+
"rewards/chosen": -1.0185908507031853,
|
| 1001 |
+
"rewards/rejected": -2.475419920049899,
|
| 1002 |
+
"rewards/accuracies": 0.825,
|
| 1003 |
+
"rewards/margins": 1.4568290673196316,
|
| 1004 |
+
"logps/chosen": -57.823643791675565,
|
| 1005 |
+
"logps/rejected": -77.6337880730629,
|
| 1006 |
+
"epoch": 0.864,
|
| 1007 |
+
"step": 540
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"loss": 0.5138762950897217,
|
| 1011 |
+
"grad_norm": 0.20225152373313904,
|
| 1012 |
+
"learning_rate": 5.608e-06,
|
| 1013 |
+
"entropy": 0.8408388607262168,
|
| 1014 |
+
"num_tokens": 7390370.0,
|
| 1015 |
+
"logits/chosen": -1.452377485576904,
|
| 1016 |
+
"logits/rejected": -1.1811360608292836,
|
| 1017 |
+
"mean_token_accuracy": 0.6988154636695981,
|
| 1018 |
+
"rewards/chosen": -0.9779604363597173,
|
| 1019 |
+
"rewards/rejected": -2.077156559471041,
|
| 1020 |
+
"rewards/accuracies": 0.775,
|
| 1021 |
+
"rewards/margins": 1.0991961166262627,
|
| 1022 |
+
"logps/chosen": -62.83508495092392,
|
| 1023 |
+
"logps/rejected": -73.65640615224838,
|
| 1024 |
+
"epoch": 0.88,
|
| 1025 |
+
"step": 550
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"loss": 0.46808481216430664,
|
| 1029 |
+
"grad_norm": 0.231434628367424,
|
| 1030 |
+
"learning_rate": 5.528e-06,
|
| 1031 |
+
"entropy": 0.7946740388579201,
|
| 1032 |
+
"num_tokens": 7524399.0,
|
| 1033 |
+
"logits/chosen": -1.5223124118443634,
|
| 1034 |
+
"logits/rejected": -1.2790454498600197,
|
| 1035 |
+
"mean_token_accuracy": 0.7107807131484151,
|
| 1036 |
+
"rewards/chosen": -0.9612048123381101,
|
| 1037 |
+
"rewards/rejected": -2.1582931477809324,
|
| 1038 |
+
"rewards/accuracies": 0.825,
|
| 1039 |
+
"rewards/margins": 1.1970883319154382,
|
| 1040 |
+
"logps/chosen": -62.14965318441391,
|
| 1041 |
+
"logps/rejected": -69.43792499303818,
|
| 1042 |
+
"epoch": 0.896,
|
| 1043 |
+
"step": 560
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"loss": 0.4392838001251221,
|
| 1047 |
+
"grad_norm": 0.25191986560821533,
|
| 1048 |
+
"learning_rate": 5.448e-06,
|
| 1049 |
+
"entropy": 0.897533849766478,
|
| 1050 |
+
"num_tokens": 7658604.0,
|
| 1051 |
+
"logits/chosen": -1.3434048518547745,
|
| 1052 |
+
"logits/rejected": -1.030690605201991,
|
| 1053 |
+
"mean_token_accuracy": 0.6994335424154997,
|
| 1054 |
+
"rewards/chosen": -0.7594973979721544,
|
| 1055 |
+
"rewards/rejected": -2.0346759966923855,
|
| 1056 |
+
"rewards/accuracies": 0.81875,
|
| 1057 |
+
"rewards/margins": 1.2751785954460502,
|
| 1058 |
+
"logps/chosen": -58.43939155340195,
|
| 1059 |
+
"logps/rejected": -73.51816222667694,
|
| 1060 |
+
"epoch": 0.912,
|
| 1061 |
+
"step": 570
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"loss": 0.41029891967773435,
|
| 1065 |
+
"grad_norm": 0.1326724737882614,
|
| 1066 |
+
"learning_rate": 5.368000000000001e-06,
|
| 1067 |
+
"entropy": 0.9109846628591185,
|
| 1068 |
+
"num_tokens": 7793743.0,
|
| 1069 |
+
"logits/chosen": -1.2809799762732126,
|
| 1070 |
+
"logits/rejected": -1.0200449868068937,
|
| 1071 |
+
"mean_token_accuracy": 0.6849511336535216,
|
| 1072 |
+
"rewards/chosen": -0.8170324935868848,
|
| 1073 |
+
"rewards/rejected": -2.2066244954708965,
|
| 1074 |
+
"rewards/accuracies": 0.83125,
|
| 1075 |
+
"rewards/margins": 1.3895920076407493,
|
| 1076 |
+
"logps/chosen": -66.22026265859604,
|
| 1077 |
+
"logps/rejected": -75.75061405897141,
|
| 1078 |
+
"epoch": 0.928,
|
| 1079 |
+
"step": 580
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"loss": 0.40343465805053713,
|
| 1083 |
+
"grad_norm": 0.16180811822414398,
|
| 1084 |
+
"learning_rate": 5.288000000000001e-06,
|
| 1085 |
+
"entropy": 0.8497528363193851,
|
| 1086 |
+
"num_tokens": 7928689.0,
|
| 1087 |
+
"logits/chosen": -1.275616841111561,
|
| 1088 |
+
"logits/rejected": -1.1260256720537545,
|
| 1089 |
+
"mean_token_accuracy": 0.7137902221642435,
|
| 1090 |
+
"rewards/chosen": -0.7834315858235641,
|
| 1091 |
+
"rewards/rejected": -2.285609513189411,
|
| 1092 |
+
"rewards/accuracies": 0.85,
|
| 1093 |
+
"rewards/margins": 1.502177918702364,
|
| 1094 |
+
"logps/chosen": -66.188141644001,
|
| 1095 |
+
"logps/rejected": -77.6662427186966,
|
| 1096 |
+
"epoch": 0.944,
|
| 1097 |
+
"step": 590
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"loss": 0.3679544687271118,
|
| 1101 |
+
"grad_norm": 0.16845545172691345,
|
| 1102 |
+
"learning_rate": 5.208000000000001e-06,
|
| 1103 |
+
"entropy": 0.8121149926242651,
|
| 1104 |
+
"num_tokens": 8062710.0,
|
| 1105 |
+
"logits/chosen": -1.5301665243925657,
|
| 1106 |
+
"logits/rejected": -1.0645871439354784,
|
| 1107 |
+
"mean_token_accuracy": 0.727741417568177,
|
| 1108 |
+
"rewards/chosen": -0.8627201122930274,
|
| 1109 |
+
"rewards/rejected": -2.3583478098065824,
|
| 1110 |
+
"rewards/accuracies": 0.8375,
|
| 1111 |
+
"rewards/margins": 1.4956277024932205,
|
| 1112 |
+
"logps/chosen": -58.9771901845932,
|
| 1113 |
+
"logps/rejected": -75.46244786977768,
|
| 1114 |
+
"epoch": 0.96,
|
| 1115 |
+
"step": 600
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"loss": 0.3810621738433838,
|
| 1119 |
+
"grad_norm": 0.11300338804721832,
|
| 1120 |
+
"learning_rate": 5.128000000000001e-06,
|
| 1121 |
+
"entropy": 0.8658083314250689,
|
| 1122 |
+
"num_tokens": 8197009.0,
|
| 1123 |
+
"logits/chosen": -1.306422252398257,
|
| 1124 |
+
"logits/rejected": -1.0578630700648384,
|
| 1125 |
+
"mean_token_accuracy": 0.6950392883270979,
|
| 1126 |
+
"rewards/chosen": -0.6266167415480595,
|
| 1127 |
+
"rewards/rejected": -2.0611097692511975,
|
| 1128 |
+
"rewards/accuracies": 0.84375,
|
| 1129 |
+
"rewards/margins": 1.4344930292107165,
|
| 1130 |
+
"logps/chosen": -61.29031510353089,
|
| 1131 |
+
"logps/rejected": -72.78142138719559,
|
| 1132 |
+
"epoch": 0.976,
|
| 1133 |
+
"step": 610
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"loss": 0.4289699077606201,
|
| 1137 |
+
"grad_norm": 0.18412326276302338,
|
| 1138 |
+
"learning_rate": 5.048000000000001e-06,
|
| 1139 |
+
"entropy": 0.8150372310570674,
|
| 1140 |
+
"num_tokens": 8330156.0,
|
| 1141 |
+
"logits/chosen": -1.5934983785184191,
|
| 1142 |
+
"logits/rejected": -1.323951400892812,
|
| 1143 |
+
"mean_token_accuracy": 0.7182830918580294,
|
| 1144 |
+
"rewards/chosen": -0.6873905764863594,
|
| 1145 |
+
"rewards/rejected": -1.9271712393616327,
|
| 1146 |
+
"rewards/accuracies": 0.825,
|
| 1147 |
+
"rewards/margins": 1.2397806531749667,
|
| 1148 |
+
"logps/chosen": -55.59258867502213,
|
| 1149 |
+
"logps/rejected": -68.15274711847306,
|
| 1150 |
+
"epoch": 0.992,
|
| 1151 |
+
"step": 620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"loss": 0.4300729751586914,
|
| 1155 |
+
"grad_norm": 0.28634947538375854,
|
| 1156 |
+
"learning_rate": 4.9680000000000005e-06,
|
| 1157 |
+
"entropy": 0.7724389405047987,
|
| 1158 |
+
"num_tokens": 8465041.0,
|
| 1159 |
+
"logits/chosen": -1.7089824220533327,
|
| 1160 |
+
"logits/rejected": -1.3800794928226323,
|
| 1161 |
+
"mean_token_accuracy": 0.6946423420682549,
|
| 1162 |
+
"rewards/chosen": -0.8758971913543064,
|
| 1163 |
+
"rewards/rejected": -2.3667486871592702,
|
| 1164 |
+
"rewards/accuracies": 0.79375,
|
| 1165 |
+
"rewards/margins": 1.4908514943905176,
|
| 1166 |
+
"logps/chosen": -60.0764141201973,
|
| 1167 |
+
"logps/rejected": -77.36668430566787,
|
| 1168 |
+
"epoch": 1.008,
|
| 1169 |
+
"step": 630
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"loss": 0.351672101020813,
|
| 1173 |
+
"grad_norm": 0.17595279216766357,
|
| 1174 |
+
"learning_rate": 4.8880000000000005e-06,
|
| 1175 |
+
"entropy": 0.8523599599866429,
|
| 1176 |
+
"num_tokens": 8599500.0,
|
| 1177 |
+
"logits/chosen": -1.5566354847185286,
|
| 1178 |
+
"logits/rejected": -1.2416854425462645,
|
| 1179 |
+
"mean_token_accuracy": 0.6994511602446437,
|
| 1180 |
+
"rewards/chosen": -0.8965433251578361,
|
| 1181 |
+
"rewards/rejected": -2.5663361197221093,
|
| 1182 |
+
"rewards/accuracies": 0.8375,
|
| 1183 |
+
"rewards/margins": 1.6697927916189657,
|
| 1184 |
+
"logps/chosen": -63.743375754356386,
|
| 1185 |
+
"logps/rejected": -78.3684019446373,
|
| 1186 |
+
"epoch": 1.024,
|
| 1187 |
+
"step": 640
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"loss": 0.39403135776519777,
|
| 1191 |
+
"grad_norm": 0.1984976977109909,
|
| 1192 |
+
"learning_rate": 4.808e-06,
|
| 1193 |
+
"entropy": 0.8592765646695625,
|
| 1194 |
+
"num_tokens": 8733676.0,
|
| 1195 |
+
"logits/chosen": -1.5561912591094866,
|
| 1196 |
+
"logits/rejected": -1.1381176918309974,
|
| 1197 |
+
"mean_token_accuracy": 0.6983322446234524,
|
| 1198 |
+
"rewards/chosen": -0.9339752380794379,
|
| 1199 |
+
"rewards/rejected": -2.5663581168220846,
|
| 1200 |
+
"rewards/accuracies": 0.8125,
|
| 1201 |
+
"rewards/margins": 1.6323828775901348,
|
| 1202 |
+
"logps/chosen": -60.757939064502715,
|
| 1203 |
+
"logps/rejected": -81.85455255508423,
|
| 1204 |
+
"epoch": 1.04,
|
| 1205 |
+
"step": 650
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"loss": 0.39429287910461425,
|
| 1209 |
+
"grad_norm": 0.1759313941001892,
|
| 1210 |
+
"learning_rate": 4.728e-06,
|
| 1211 |
+
"entropy": 0.8048280104179867,
|
| 1212 |
+
"num_tokens": 8866920.0,
|
| 1213 |
+
"logits/chosen": -1.6457790961867915,
|
| 1214 |
+
"logits/rejected": -1.3161897829243776,
|
| 1215 |
+
"mean_token_accuracy": 0.6984306069090962,
|
| 1216 |
+
"rewards/chosen": -1.0328793285123539,
|
| 1217 |
+
"rewards/rejected": -2.5047683740616775,
|
| 1218 |
+
"rewards/accuracies": 0.84375,
|
| 1219 |
+
"rewards/margins": 1.4718890445830766,
|
| 1220 |
+
"logps/chosen": -60.58936309814453,
|
| 1221 |
+
"logps/rejected": -75.65209144353867,
|
| 1222 |
+
"epoch": 1.056,
|
| 1223 |
+
"step": 660
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"loss": 0.36423916816711427,
|
| 1227 |
+
"grad_norm": 0.13459306955337524,
|
| 1228 |
+
"learning_rate": 4.648e-06,
|
| 1229 |
+
"entropy": 0.9544580190035049,
|
| 1230 |
+
"num_tokens": 9002809.0,
|
| 1231 |
+
"logits/chosen": -1.4630308122069675,
|
| 1232 |
+
"logits/rejected": -1.1763000890003652,
|
| 1233 |
+
"mean_token_accuracy": 0.66888557523489,
|
| 1234 |
+
"rewards/chosen": -1.155039997237327,
|
| 1235 |
+
"rewards/rejected": -2.982239642040804,
|
| 1236 |
+
"rewards/accuracies": 0.83125,
|
| 1237 |
+
"rewards/margins": 1.8271996420808136,
|
| 1238 |
+
"logps/chosen": -78.547492313385,
|
| 1239 |
+
"logps/rejected": -90.53760310411454,
|
| 1240 |
+
"epoch": 1.072,
|
| 1241 |
+
"step": 670
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"loss": 0.4182881355285645,
|
| 1245 |
+
"grad_norm": 0.26433658599853516,
|
| 1246 |
+
"learning_rate": 4.568e-06,
|
| 1247 |
+
"entropy": 0.8703971885872306,
|
| 1248 |
+
"num_tokens": 9138006.0,
|
| 1249 |
+
"logits/chosen": -1.6505449533579177,
|
| 1250 |
+
"logits/rejected": -1.2025132287051878,
|
| 1251 |
+
"mean_token_accuracy": 0.6970316587015987,
|
| 1252 |
+
"rewards/chosen": -1.281551358116849,
|
| 1253 |
+
"rewards/rejected": -2.89622505933512,
|
| 1254 |
+
"rewards/accuracies": 0.7625,
|
| 1255 |
+
"rewards/margins": 1.614673706702888,
|
| 1256 |
+
"logps/chosen": -66.8018369436264,
|
| 1257 |
+
"logps/rejected": -89.63993077278137,
|
| 1258 |
+
"epoch": 1.088,
|
| 1259 |
+
"step": 680
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"loss": 0.30962481498718264,
|
| 1263 |
+
"grad_norm": 0.27371224761009216,
|
| 1264 |
+
"learning_rate": 4.488e-06,
|
| 1265 |
+
"entropy": 0.7996644546336029,
|
| 1266 |
+
"num_tokens": 9272926.0,
|
| 1267 |
+
"logits/chosen": -1.6190867560520428,
|
| 1268 |
+
"logits/rejected": -1.1754203887360388,
|
| 1269 |
+
"mean_token_accuracy": 0.7229850918054581,
|
| 1270 |
+
"rewards/chosen": -1.1402118994039483,
|
| 1271 |
+
"rewards/rejected": -3.1755229078815317,
|
| 1272 |
+
"rewards/accuracies": 0.86875,
|
| 1273 |
+
"rewards/margins": 2.0353110095486047,
|
| 1274 |
+
"logps/chosen": -63.039683640003204,
|
| 1275 |
+
"logps/rejected": -86.67976603507995,
|
| 1276 |
+
"epoch": 1.104,
|
| 1277 |
+
"step": 690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"loss": 0.3763152599334717,
|
| 1281 |
+
"grad_norm": 0.25219184160232544,
|
| 1282 |
+
"learning_rate": 4.408000000000001e-06,
|
| 1283 |
+
"entropy": 0.8175220161298057,
|
| 1284 |
+
"num_tokens": 9407243.0,
|
| 1285 |
+
"logits/chosen": -1.638960367267067,
|
| 1286 |
+
"logits/rejected": -1.2401611862490487,
|
| 1287 |
+
"mean_token_accuracy": 0.7072740724310279,
|
| 1288 |
+
"rewards/chosen": -1.237404835311463,
|
| 1289 |
+
"rewards/rejected": -2.8441451793536543,
|
| 1290 |
+
"rewards/accuracies": 0.84375,
|
| 1291 |
+
"rewards/margins": 1.6067403441295027,
|
| 1292 |
+
"logps/chosen": -69.88301295042038,
|
| 1293 |
+
"logps/rejected": -76.65401461124421,
|
| 1294 |
+
"epoch": 1.12,
|
| 1295 |
+
"step": 700
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"loss": 0.32136104106903074,
|
| 1299 |
+
"grad_norm": 0.20682720839977264,
|
| 1300 |
+
"learning_rate": 4.328000000000001e-06,
|
| 1301 |
+
"entropy": 0.7797062049154192,
|
| 1302 |
+
"num_tokens": 9541881.0,
|
| 1303 |
+
"logits/chosen": -1.8241846089858633,
|
| 1304 |
+
"logits/rejected": -1.2498475755377636,
|
| 1305 |
+
"mean_token_accuracy": 0.7237331310287118,
|
| 1306 |
+
"rewards/chosen": -1.282835571677424,
|
| 1307 |
+
"rewards/rejected": -3.2540257604792715,
|
| 1308 |
+
"rewards/accuracies": 0.86875,
|
| 1309 |
+
"rewards/margins": 1.9711901864036918,
|
| 1310 |
+
"logps/chosen": -64.261137342453,
|
| 1311 |
+
"logps/rejected": -85.98668245077133,
|
| 1312 |
+
"epoch": 1.1360000000000001,
|
| 1313 |
+
"step": 710
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"loss": 0.36211495399475097,
|
| 1317 |
+
"grad_norm": 0.18496981263160706,
|
| 1318 |
+
"learning_rate": 4.248000000000001e-06,
|
| 1319 |
+
"entropy": 0.8136180234898347,
|
| 1320 |
+
"num_tokens": 9676124.0,
|
| 1321 |
+
"logits/chosen": -1.662963946550788,
|
| 1322 |
+
"logits/rejected": -1.2975660610969042,
|
| 1323 |
+
"mean_token_accuracy": 0.7034357896074652,
|
| 1324 |
+
"rewards/chosen": -1.2120525182137498,
|
| 1325 |
+
"rewards/rejected": -2.9877946829539725,
|
| 1326 |
+
"rewards/accuracies": 0.83125,
|
| 1327 |
+
"rewards/margins": 1.775742164812982,
|
| 1328 |
+
"logps/chosen": -65.20653177499771,
|
| 1329 |
+
"logps/rejected": -82.91743679046631,
|
| 1330 |
+
"epoch": 1.152,
|
| 1331 |
+
"step": 720
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"loss": 0.3944878339767456,
|
| 1335 |
+
"grad_norm": 0.14160765707492828,
|
| 1336 |
+
"learning_rate": 4.168000000000001e-06,
|
| 1337 |
+
"entropy": 0.8193648672022391,
|
| 1338 |
+
"num_tokens": 9809272.0,
|
| 1339 |
+
"logits/chosen": -1.6782603406741465,
|
| 1340 |
+
"logits/rejected": -1.2885239180924282,
|
| 1341 |
+
"mean_token_accuracy": 0.6822029747068882,
|
| 1342 |
+
"rewards/chosen": -1.041324549249839,
|
| 1343 |
+
"rewards/rejected": -2.6822686459869147,
|
| 1344 |
+
"rewards/accuracies": 0.8375,
|
| 1345 |
+
"rewards/margins": 1.6409440968185662,
|
| 1346 |
+
"logps/chosen": -64.25518968105317,
|
| 1347 |
+
"logps/rejected": -73.46110652685165,
|
| 1348 |
+
"epoch": 1.168,
|
| 1349 |
+
"step": 730
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"loss": 0.3898670434951782,
|
| 1353 |
+
"grad_norm": 0.477348268032074,
|
| 1354 |
+
"learning_rate": 4.0880000000000005e-06,
|
| 1355 |
+
"entropy": 0.904066581139341,
|
| 1356 |
+
"num_tokens": 9944472.0,
|
| 1357 |
+
"logits/chosen": -1.4180159506110437,
|
| 1358 |
+
"logits/rejected": -1.1167911378919633,
|
| 1359 |
+
"mean_token_accuracy": 0.6910396177321673,
|
| 1360 |
+
"rewards/chosen": -1.0064560518600048,
|
| 1361 |
+
"rewards/rejected": -2.755757818929851,
|
| 1362 |
+
"rewards/accuracies": 0.84375,
|
| 1363 |
+
"rewards/margins": 1.7493017674423754,
|
| 1364 |
+
"logps/chosen": -67.54992797374726,
|
| 1365 |
+
"logps/rejected": -83.48455815315246,
|
| 1366 |
+
"epoch": 1.184,
|
| 1367 |
+
"step": 740
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"loss": 0.3514470338821411,
|
| 1371 |
+
"grad_norm": 0.21692045032978058,
|
| 1372 |
+
"learning_rate": 4.008e-06,
|
| 1373 |
+
"entropy": 0.832500480231829,
|
| 1374 |
+
"num_tokens": 10078733.0,
|
| 1375 |
+
"logits/chosen": -1.4611039879162244,
|
| 1376 |
+
"logits/rejected": -1.1890601720671152,
|
| 1377 |
+
"mean_token_accuracy": 0.7059988923370838,
|
| 1378 |
+
"rewards/chosen": -0.9490231054311152,
|
| 1379 |
+
"rewards/rejected": -2.8147553551010787,
|
| 1380 |
+
"rewards/accuracies": 0.86875,
|
| 1381 |
+
"rewards/margins": 1.8657322534825653,
|
| 1382 |
+
"logps/chosen": -59.17188249826431,
|
| 1383 |
+
"logps/rejected": -79.69901636838912,
|
| 1384 |
+
"epoch": 1.2,
|
| 1385 |
+
"step": 750
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"loss": 0.39118783473968505,
|
| 1389 |
+
"grad_norm": 0.08524914085865021,
|
| 1390 |
+
"learning_rate": 3.928e-06,
|
| 1391 |
+
"entropy": 0.8148924410372274,
|
| 1392 |
+
"num_tokens": 10212826.0,
|
| 1393 |
+
"logits/chosen": -1.6316838252528025,
|
| 1394 |
+
"logits/rejected": -1.3824933858769524,
|
| 1395 |
+
"mean_token_accuracy": 0.7193772990256548,
|
| 1396 |
+
"rewards/chosen": -1.0879453276837012,
|
| 1397 |
+
"rewards/rejected": -2.708411489869468,
|
| 1398 |
+
"rewards/accuracies": 0.85,
|
| 1399 |
+
"rewards/margins": 1.6204661595635117,
|
| 1400 |
+
"logps/chosen": -61.753991997241975,
|
| 1401 |
+
"logps/rejected": -86.23918843269348,
|
| 1402 |
+
"epoch": 1.216,
|
| 1403 |
+
"step": 760
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"loss": 0.2837193489074707,
|
| 1407 |
+
"grad_norm": 0.14890901744365692,
|
| 1408 |
+
"learning_rate": 3.848e-06,
|
| 1409 |
+
"entropy": 0.8147611759137362,
|
| 1410 |
+
"num_tokens": 10346556.0,
|
| 1411 |
+
"logits/chosen": -1.6339600569053574,
|
| 1412 |
+
"logits/rejected": -1.335518427193826,
|
| 1413 |
+
"mean_token_accuracy": 0.7048663256689907,
|
| 1414 |
+
"rewards/chosen": -0.8854538638086524,
|
| 1415 |
+
"rewards/rejected": -2.8079105648466793,
|
| 1416 |
+
"rewards/accuracies": 0.9,
|
| 1417 |
+
"rewards/margins": 1.9224567006342113,
|
| 1418 |
+
"logps/chosen": -59.59974846839905,
|
| 1419 |
+
"logps/rejected": -79.61685808897019,
|
| 1420 |
+
"epoch": 1.232,
|
| 1421 |
+
"step": 770
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"loss": 0.3447991609573364,
|
| 1425 |
+
"grad_norm": 0.2853667140007019,
|
| 1426 |
+
"learning_rate": 3.7680000000000006e-06,
|
| 1427 |
+
"entropy": 0.7747858772752807,
|
| 1428 |
+
"num_tokens": 10480829.0,
|
| 1429 |
+
"logits/chosen": -1.7813727559070558,
|
| 1430 |
+
"logits/rejected": -1.3194815841216347,
|
| 1431 |
+
"mean_token_accuracy": 0.7163394480012357,
|
| 1432 |
+
"rewards/chosen": -1.2985240196736414,
|
| 1433 |
+
"rewards/rejected": -3.340971144242212,
|
| 1434 |
+
"rewards/accuracies": 0.8375,
|
| 1435 |
+
"rewards/margins": 2.0424471154808996,
|
| 1436 |
+
"logps/chosen": -65.54894176721572,
|
| 1437 |
+
"logps/rejected": -86.83000173568726,
|
| 1438 |
+
"epoch": 1.248,
|
| 1439 |
+
"step": 780
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"loss": 0.4076092720031738,
|
| 1443 |
+
"grad_norm": 0.25816309452056885,
|
| 1444 |
+
"learning_rate": 3.6880000000000005e-06,
|
| 1445 |
+
"entropy": 0.7755509500217158,
|
| 1446 |
+
"num_tokens": 10614414.0,
|
| 1447 |
+
"logits/chosen": -1.811802842191796,
|
| 1448 |
+
"logits/rejected": -1.4966397742998652,
|
| 1449 |
+
"mean_token_accuracy": 0.7023726673796773,
|
| 1450 |
+
"rewards/chosen": -1.261663666684217,
|
| 1451 |
+
"rewards/rejected": -2.9303010510047898,
|
| 1452 |
+
"rewards/accuracies": 0.83125,
|
| 1453 |
+
"rewards/margins": 1.668637388665229,
|
| 1454 |
+
"logps/chosen": -63.06534553766251,
|
| 1455 |
+
"logps/rejected": -80.0437378168106,
|
| 1456 |
+
"epoch": 1.264,
|
| 1457 |
+
"step": 790
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"loss": 0.3804662227630615,
|
| 1461 |
+
"grad_norm": 0.4363705813884735,
|
| 1462 |
+
"learning_rate": 3.6080000000000004e-06,
|
| 1463 |
+
"entropy": 0.8125547816307517,
|
| 1464 |
+
"num_tokens": 10748424.0,
|
| 1465 |
+
"logits/chosen": -1.850811308539048,
|
| 1466 |
+
"logits/rejected": -1.3608433440017866,
|
| 1467 |
+
"mean_token_accuracy": 0.6973903543315828,
|
| 1468 |
+
"rewards/chosen": -1.3105751807102934,
|
| 1469 |
+
"rewards/rejected": -3.2067231149965663,
|
| 1470 |
+
"rewards/accuracies": 0.81875,
|
| 1471 |
+
"rewards/margins": 1.896147930342704,
|
| 1472 |
+
"logps/chosen": -64.90501267910004,
|
| 1473 |
+
"logps/rejected": -85.0686597943306,
|
| 1474 |
+
"epoch": 1.28,
|
| 1475 |
+
"step": 800
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"loss": 0.3295221090316772,
|
| 1479 |
+
"grad_norm": 0.19928871095180511,
|
| 1480 |
+
"learning_rate": 3.5280000000000004e-06,
|
| 1481 |
+
"entropy": 0.8632584328879602,
|
| 1482 |
+
"num_tokens": 10884117.0,
|
| 1483 |
+
"logits/chosen": -1.585350777368448,
|
| 1484 |
+
"logits/rejected": -1.2601394946588103,
|
| 1485 |
+
"mean_token_accuracy": 0.7163022305816412,
|
| 1486 |
+
"rewards/chosen": -1.1345790353720075,
|
| 1487 |
+
"rewards/rejected": -3.0583988034166394,
|
| 1488 |
+
"rewards/accuracies": 0.8375,
|
| 1489 |
+
"rewards/margins": 1.923819756321609,
|
| 1490 |
+
"logps/chosen": -75.44103773832322,
|
| 1491 |
+
"logps/rejected": -84.95239799022674,
|
| 1492 |
+
"epoch": 1.296,
|
| 1493 |
+
"step": 810
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"loss": 0.4117788791656494,
|
| 1497 |
+
"grad_norm": 0.3535456955432892,
|
| 1498 |
+
"learning_rate": 3.4480000000000003e-06,
|
| 1499 |
+
"entropy": 0.7660285441175801,
|
| 1500 |
+
"num_tokens": 11018693.0,
|
| 1501 |
+
"logits/chosen": -1.6485695304175976,
|
| 1502 |
+
"logits/rejected": -1.3562505212611475,
|
| 1503 |
+
"mean_token_accuracy": 0.7182219873182476,
|
| 1504 |
+
"rewards/chosen": -1.1482815635274164,
|
| 1505 |
+
"rewards/rejected": -2.8863029033353085,
|
| 1506 |
+
"rewards/accuracies": 0.80625,
|
| 1507 |
+
"rewards/margins": 1.7380213363096118,
|
| 1508 |
+
"logps/chosen": -60.99362645149231,
|
| 1509 |
+
"logps/rejected": -83.36733787059784,
|
| 1510 |
+
"epoch": 1.312,
|
| 1511 |
+
"step": 820
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"loss": 0.3215837001800537,
|
| 1515 |
+
"grad_norm": 0.2147626429796219,
|
| 1516 |
+
"learning_rate": 3.368e-06,
|
| 1517 |
+
"entropy": 0.8505063455617347,
|
| 1518 |
+
"num_tokens": 11153738.0,
|
| 1519 |
+
"logits/chosen": -1.6468285197599049,
|
| 1520 |
+
"logits/rejected": -1.3726640135538268,
|
| 1521 |
+
"mean_token_accuracy": 0.6910511496476829,
|
| 1522 |
+
"rewards/chosen": -1.1186513825552538,
|
| 1523 |
+
"rewards/rejected": -2.932185032824054,
|
| 1524 |
+
"rewards/accuracies": 0.86875,
|
| 1525 |
+
"rewards/margins": 1.8135336461942644,
|
| 1526 |
+
"logps/chosen": -70.2434514760971,
|
| 1527 |
+
"logps/rejected": -84.44510225057601,
|
| 1528 |
+
"epoch": 1.328,
|
| 1529 |
+
"step": 830
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"loss": 0.370072078704834,
|
| 1533 |
+
"grad_norm": 0.2706229090690613,
|
| 1534 |
+
"learning_rate": 3.288e-06,
|
| 1535 |
+
"entropy": 0.8658261221717112,
|
| 1536 |
+
"num_tokens": 11288148.0,
|
| 1537 |
+
"logits/chosen": -1.4132377863770709,
|
| 1538 |
+
"logits/rejected": -1.1302670444944378,
|
| 1539 |
+
"mean_token_accuracy": 0.704999322257936,
|
| 1540 |
+
"rewards/chosen": -0.9550284901866689,
|
| 1541 |
+
"rewards/rejected": -2.5942960902582852,
|
| 1542 |
+
"rewards/accuracies": 0.84375,
|
| 1543 |
+
"rewards/margins": 1.6392675936222076,
|
| 1544 |
+
"logps/chosen": -62.77679759263992,
|
| 1545 |
+
"logps/rejected": -78.73835257291793,
|
| 1546 |
+
"epoch": 1.3439999999999999,
|
| 1547 |
+
"step": 840
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"loss": 0.3280667781829834,
|
| 1551 |
+
"grad_norm": 0.24525395035743713,
|
| 1552 |
+
"learning_rate": 3.208e-06,
|
| 1553 |
+
"entropy": 0.8233829756863997,
|
| 1554 |
+
"num_tokens": 11422122.0,
|
| 1555 |
+
"logits/chosen": -1.6491031347464493,
|
| 1556 |
+
"logits/rejected": -1.2391678362276777,
|
| 1557 |
+
"mean_token_accuracy": 0.7019108822569251,
|
| 1558 |
+
"rewards/chosen": -0.8939242239226587,
|
| 1559 |
+
"rewards/rejected": -2.7150645235553386,
|
| 1560 |
+
"rewards/accuracies": 0.86875,
|
| 1561 |
+
"rewards/margins": 1.821140293776989,
|
| 1562 |
+
"logps/chosen": -59.29202291965485,
|
| 1563 |
+
"logps/rejected": -79.48249032497407,
|
| 1564 |
+
"epoch": 1.3599999999999999,
|
| 1565 |
+
"step": 850
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"loss": 0.3633824110031128,
|
| 1569 |
+
"grad_norm": 0.2849610447883606,
|
| 1570 |
+
"learning_rate": 3.1280000000000004e-06,
|
| 1571 |
+
"entropy": 0.8163787097088061,
|
| 1572 |
+
"num_tokens": 11555439.0,
|
| 1573 |
+
"logits/chosen": -1.6588279042547849,
|
| 1574 |
+
"logits/rejected": -1.3349530506310818,
|
| 1575 |
+
"mean_token_accuracy": 0.6893864122219384,
|
| 1576 |
+
"rewards/chosen": -0.9723918261996005,
|
| 1577 |
+
"rewards/rejected": -2.620256224478362,
|
| 1578 |
+
"rewards/accuracies": 0.8625,
|
| 1579 |
+
"rewards/margins": 1.6478643929818646,
|
| 1580 |
+
"logps/chosen": -62.69446007013321,
|
| 1581 |
+
"logps/rejected": -73.10199151039123,
|
| 1582 |
+
"epoch": 1.376,
|
| 1583 |
+
"step": 860
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"loss": 0.36836340427398684,
|
| 1587 |
+
"grad_norm": 0.33433133363723755,
|
| 1588 |
+
"learning_rate": 3.0480000000000003e-06,
|
| 1589 |
+
"entropy": 0.7954415735090151,
|
| 1590 |
+
"num_tokens": 11689826.0,
|
| 1591 |
+
"logits/chosen": -1.7840552560337684,
|
| 1592 |
+
"logits/rejected": -1.405669701827261,
|
| 1593 |
+
"mean_token_accuracy": 0.7021645514294506,
|
| 1594 |
+
"rewards/chosen": -1.0290490832645447,
|
| 1595 |
+
"rewards/rejected": -2.7825961942784487,
|
| 1596 |
+
"rewards/accuracies": 0.7875,
|
| 1597 |
+
"rewards/margins": 1.7535471043549478,
|
| 1598 |
+
"logps/chosen": -62.958821547031405,
|
| 1599 |
+
"logps/rejected": -79.9969735980034,
|
| 1600 |
+
"epoch": 1.392,
|
| 1601 |
+
"step": 870
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"loss": 0.39579946994781495,
|
| 1605 |
+
"grad_norm": 0.20239688456058502,
|
| 1606 |
+
"learning_rate": 2.9680000000000002e-06,
|
| 1607 |
+
"entropy": 0.8648671839095187,
|
| 1608 |
+
"num_tokens": 11824262.0,
|
| 1609 |
+
"logits/chosen": -1.517392932998193,
|
| 1610 |
+
"logits/rejected": -1.3057744264956395,
|
| 1611 |
+
"mean_token_accuracy": 0.6842759451828897,
|
| 1612 |
+
"rewards/chosen": -1.0850564706523982,
|
| 1613 |
+
"rewards/rejected": -2.8189570247835944,
|
| 1614 |
+
"rewards/accuracies": 0.80625,
|
| 1615 |
+
"rewards/margins": 1.7339005566667764,
|
| 1616 |
+
"logps/chosen": -70.37592285871506,
|
| 1617 |
+
"logps/rejected": -79.68846825361251,
|
| 1618 |
+
"epoch": 1.408,
|
| 1619 |
+
"step": 880
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"loss": 0.32751243114471434,
|
| 1623 |
+
"grad_norm": 0.11899573355913162,
|
| 1624 |
+
"learning_rate": 2.888e-06,
|
| 1625 |
+
"entropy": 0.8026103116819285,
|
| 1626 |
+
"num_tokens": 11958211.0,
|
| 1627 |
+
"logits/chosen": -1.5754916948022033,
|
| 1628 |
+
"logits/rejected": -1.2621632796686542,
|
| 1629 |
+
"mean_token_accuracy": 0.709551977738738,
|
| 1630 |
+
"rewards/chosen": -1.0446015263500157,
|
| 1631 |
+
"rewards/rejected": -2.8077158498344943,
|
| 1632 |
+
"rewards/accuracies": 0.86875,
|
| 1633 |
+
"rewards/margins": 1.7631143254227937,
|
| 1634 |
+
"logps/chosen": -63.91794439554214,
|
| 1635 |
+
"logps/rejected": -76.60812841653824,
|
| 1636 |
+
"epoch": 1.424,
|
| 1637 |
+
"step": 890
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"loss": 0.311025595664978,
|
| 1641 |
+
"grad_norm": 0.18633148074150085,
|
| 1642 |
+
"learning_rate": 2.808e-06,
|
| 1643 |
+
"entropy": 0.8465069067373406,
|
| 1644 |
+
"num_tokens": 12092461.0,
|
| 1645 |
+
"logits/chosen": -1.684217637752368,
|
| 1646 |
+
"logits/rejected": -1.4311675631899698,
|
| 1647 |
+
"mean_token_accuracy": 0.7081242019310594,
|
| 1648 |
+
"rewards/chosen": -0.9291278672142653,
|
| 1649 |
+
"rewards/rejected": -2.7168329820036887,
|
| 1650 |
+
"rewards/accuracies": 0.9125,
|
| 1651 |
+
"rewards/margins": 1.787705100607127,
|
| 1652 |
+
"logps/chosen": -65.40056767463685,
|
| 1653 |
+
"logps/rejected": -77.32470668554306,
|
| 1654 |
+
"epoch": 1.44,
|
| 1655 |
+
"step": 900
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"loss": 0.41963820457458495,
|
| 1659 |
+
"grad_norm": 0.3252865970134735,
|
| 1660 |
+
"learning_rate": 2.728e-06,
|
| 1661 |
+
"entropy": 0.8206196324783377,
|
| 1662 |
+
"num_tokens": 12227051.0,
|
| 1663 |
+
"logits/chosen": -1.751943145276467,
|
| 1664 |
+
"logits/rejected": -1.4631973368450717,
|
| 1665 |
+
"mean_token_accuracy": 0.7006220754235983,
|
| 1666 |
+
"rewards/chosen": -1.120255030062981,
|
| 1667 |
+
"rewards/rejected": -2.866658524505328,
|
| 1668 |
+
"rewards/accuracies": 0.85625,
|
| 1669 |
+
"rewards/margins": 1.7464034849777819,
|
| 1670 |
+
"logps/chosen": -66.31769858598709,
|
| 1671 |
+
"logps/rejected": -78.83246475458145,
|
| 1672 |
+
"epoch": 1.456,
|
| 1673 |
+
"step": 910
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"loss": 0.3600214958190918,
|
| 1677 |
+
"grad_norm": 0.19774270057678223,
|
| 1678 |
+
"learning_rate": 2.648e-06,
|
| 1679 |
+
"entropy": 0.8105380321110716,
|
| 1680 |
+
"num_tokens": 12360999.0,
|
| 1681 |
+
"logits/chosen": -1.754174652794114,
|
| 1682 |
+
"logits/rejected": -1.3756857513271954,
|
| 1683 |
+
"mean_token_accuracy": 0.6845561027526855,
|
| 1684 |
+
"rewards/chosen": -1.1735321268170082,
|
| 1685 |
+
"rewards/rejected": -3.0875361559868906,
|
| 1686 |
+
"rewards/accuracies": 0.83125,
|
| 1687 |
+
"rewards/margins": 1.9140040219761432,
|
| 1688 |
+
"logps/chosen": -63.802481186389926,
|
| 1689 |
+
"logps/rejected": -82.82446665763855,
|
| 1690 |
+
"epoch": 1.472,
|
| 1691 |
+
"step": 920
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"loss": 0.38991284370422363,
|
| 1695 |
+
"grad_norm": 0.17794227600097656,
|
| 1696 |
+
"learning_rate": 2.568e-06,
|
| 1697 |
+
"entropy": 0.8483509451092687,
|
| 1698 |
+
"num_tokens": 12495017.0,
|
| 1699 |
+
"logits/chosen": -1.6839636590332638,
|
| 1700 |
+
"logits/rejected": -1.3878673548862333,
|
| 1701 |
+
"mean_token_accuracy": 0.6770037472248077,
|
| 1702 |
+
"rewards/chosen": -1.1462732924119337,
|
| 1703 |
+
"rewards/rejected": -2.936118917202111,
|
| 1704 |
+
"rewards/accuracies": 0.81875,
|
| 1705 |
+
"rewards/margins": 1.789845616556704,
|
| 1706 |
+
"logps/chosen": -67.04533059597016,
|
| 1707 |
+
"logps/rejected": -81.32393040657044,
|
| 1708 |
+
"epoch": 1.488,
|
| 1709 |
+
"step": 930
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"loss": 0.3819059133529663,
|
| 1713 |
+
"grad_norm": 0.19907650351524353,
|
| 1714 |
+
"learning_rate": 2.488e-06,
|
| 1715 |
+
"entropy": 0.8542897907755105,
|
| 1716 |
+
"num_tokens": 12628825.0,
|
| 1717 |
+
"logits/chosen": -1.71170331896165,
|
| 1718 |
+
"logits/rejected": -1.374216479000578,
|
| 1719 |
+
"mean_token_accuracy": 0.7099258149042725,
|
| 1720 |
+
"rewards/chosen": -1.0272940277209273,
|
| 1721 |
+
"rewards/rejected": -2.9895632124185796,
|
| 1722 |
+
"rewards/accuracies": 0.83125,
|
| 1723 |
+
"rewards/margins": 1.9622691813856363,
|
| 1724 |
+
"logps/chosen": -68.72389820814132,
|
| 1725 |
+
"logps/rejected": -80.29468423128128,
|
| 1726 |
+
"epoch": 1.504,
|
| 1727 |
+
"step": 940
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"loss": 0.36449341773986815,
|
| 1731 |
+
"grad_norm": 0.13733473420143127,
|
| 1732 |
+
"learning_rate": 2.408e-06,
|
| 1733 |
+
"entropy": 0.8203515770670492,
|
| 1734 |
+
"num_tokens": 12763474.0,
|
| 1735 |
+
"logits/chosen": -1.7479245215911259,
|
| 1736 |
+
"logits/rejected": -1.3614535160872505,
|
| 1737 |
+
"mean_token_accuracy": 0.7107045710086822,
|
| 1738 |
+
"rewards/chosen": -1.0464620275946799,
|
| 1739 |
+
"rewards/rejected": -2.9747184620238842,
|
| 1740 |
+
"rewards/accuracies": 0.85625,
|
| 1741 |
+
"rewards/margins": 1.9282564282417298,
|
| 1742 |
+
"logps/chosen": -64.21307734847069,
|
| 1743 |
+
"logps/rejected": -85.41888726949692,
|
| 1744 |
+
"epoch": 1.52,
|
| 1745 |
+
"step": 950
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"loss": 0.4197529315948486,
|
| 1749 |
+
"grad_norm": 0.4546315670013428,
|
| 1750 |
+
"learning_rate": 2.3280000000000004e-06,
|
| 1751 |
+
"entropy": 0.8600580797472503,
|
| 1752 |
+
"num_tokens": 12897107.0,
|
| 1753 |
+
"logits/chosen": -1.5123304301188327,
|
| 1754 |
+
"logits/rejected": -1.2960668561291118,
|
| 1755 |
+
"mean_token_accuracy": 0.6949498526751995,
|
| 1756 |
+
"rewards/chosen": -1.043558939262016,
|
| 1757 |
+
"rewards/rejected": -2.7619683917961084,
|
| 1758 |
+
"rewards/accuracies": 0.84375,
|
| 1759 |
+
"rewards/margins": 1.7184094509109855,
|
| 1760 |
+
"logps/chosen": -64.66679666042327,
|
| 1761 |
+
"logps/rejected": -76.41976501941681,
|
| 1762 |
+
"epoch": 1.536,
|
| 1763 |
+
"step": 960
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"loss": 0.437890625,
|
| 1767 |
+
"grad_norm": 0.2756805121898651,
|
| 1768 |
+
"learning_rate": 2.2480000000000003e-06,
|
| 1769 |
+
"entropy": 0.8089202568226028,
|
| 1770 |
+
"num_tokens": 13030353.0,
|
| 1771 |
+
"logits/chosen": -1.7310479211277259,
|
| 1772 |
+
"logits/rejected": -1.41339340249286,
|
| 1773 |
+
"mean_token_accuracy": 0.6968138926662505,
|
| 1774 |
+
"rewards/chosen": -0.9847447354724863,
|
| 1775 |
+
"rewards/rejected": -2.6614407726796343,
|
| 1776 |
+
"rewards/accuracies": 0.83125,
|
| 1777 |
+
"rewards/margins": 1.6766960403416307,
|
| 1778 |
+
"logps/chosen": -58.44614028930664,
|
| 1779 |
+
"logps/rejected": -77.09000945091248,
|
| 1780 |
+
"epoch": 1.552,
|
| 1781 |
+
"step": 970
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"loss": 0.3848176717758179,
|
| 1785 |
+
"grad_norm": 0.14892172813415527,
|
| 1786 |
+
"learning_rate": 2.1680000000000002e-06,
|
| 1787 |
+
"entropy": 0.8586067174313939,
|
| 1788 |
+
"num_tokens": 13164375.0,
|
| 1789 |
+
"logits/chosen": -1.6485429765390862,
|
| 1790 |
+
"logits/rejected": -1.3191338729892919,
|
| 1791 |
+
"mean_token_accuracy": 0.7080312158912421,
|
| 1792 |
+
"rewards/chosen": -1.0914479763247074,
|
| 1793 |
+
"rewards/rejected": -2.8948664830531925,
|
| 1794 |
+
"rewards/accuracies": 0.81875,
|
| 1795 |
+
"rewards/margins": 1.8034184967167675,
|
| 1796 |
+
"logps/chosen": -57.48373928070068,
|
| 1797 |
+
"logps/rejected": -83.9749594092369,
|
| 1798 |
+
"epoch": 1.568,
|
| 1799 |
+
"step": 980
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"loss": 0.29979939460754396,
|
| 1803 |
+
"grad_norm": 0.21152755618095398,
|
| 1804 |
+
"learning_rate": 2.088e-06,
|
| 1805 |
+
"entropy": 0.8144882260909071,
|
| 1806 |
+
"num_tokens": 13298403.0,
|
| 1807 |
+
"logits/chosen": -1.698445047222234,
|
| 1808 |
+
"logits/rejected": -1.4396970964011087,
|
| 1809 |
+
"mean_token_accuracy": 0.6979767033830285,
|
| 1810 |
+
"rewards/chosen": -1.070426406210754,
|
| 1811 |
+
"rewards/rejected": -3.0210326331667603,
|
| 1812 |
+
"rewards/accuracies": 0.88125,
|
| 1813 |
+
"rewards/margins": 1.9506062217988074,
|
| 1814 |
+
"logps/chosen": -63.166816246509555,
|
| 1815 |
+
"logps/rejected": -81.8851042509079,
|
| 1816 |
+
"epoch": 1.584,
|
| 1817 |
+
"step": 990
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"loss": 0.35019617080688475,
|
| 1821 |
+
"grad_norm": 0.3046397268772125,
|
| 1822 |
+
"learning_rate": 2.008e-06,
|
| 1823 |
+
"entropy": 0.8480100489337928,
|
| 1824 |
+
"num_tokens": 13433900.0,
|
| 1825 |
+
"logits/chosen": -1.6347868792855649,
|
| 1826 |
+
"logits/rejected": -1.3061482556403887,
|
| 1827 |
+
"mean_token_accuracy": 0.7166428156197071,
|
| 1828 |
+
"rewards/chosen": -1.1963652066508075,
|
| 1829 |
+
"rewards/rejected": -3.097625444867299,
|
| 1830 |
+
"rewards/accuracies": 0.8375,
|
| 1831 |
+
"rewards/margins": 1.9012602401897312,
|
| 1832 |
+
"logps/chosen": -65.1461256980896,
|
| 1833 |
+
"logps/rejected": -91.62114884853364,
|
| 1834 |
+
"epoch": 1.6,
|
| 1835 |
+
"step": 1000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"loss": 0.4143568515777588,
|
| 1839 |
+
"grad_norm": 0.45840078592300415,
|
| 1840 |
+
"learning_rate": 1.928e-06,
|
| 1841 |
+
"entropy": 0.8242917212191969,
|
| 1842 |
+
"num_tokens": 13567714.0,
|
| 1843 |
+
"logits/chosen": -1.642167009439039,
|
| 1844 |
+
"logits/rejected": -1.2664307786708329,
|
| 1845 |
+
"mean_token_accuracy": 0.7156291794031858,
|
| 1846 |
+
"rewards/chosen": -1.0634639814918045,
|
| 1847 |
+
"rewards/rejected": -2.652825303049758,
|
| 1848 |
+
"rewards/accuracies": 0.83125,
|
| 1849 |
+
"rewards/margins": 1.5893613207619637,
|
| 1850 |
+
"logps/chosen": -57.849858009815215,
|
| 1851 |
+
"logps/rejected": -77.98368090391159,
|
| 1852 |
+
"epoch": 1.616,
|
| 1853 |
+
"step": 1010
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"loss": 0.37326362133026125,
|
| 1857 |
+
"grad_norm": 0.1788681149482727,
|
| 1858 |
+
"learning_rate": 1.8480000000000001e-06,
|
| 1859 |
+
"entropy": 0.9456939667841653,
|
| 1860 |
+
"num_tokens": 13702663.0,
|
| 1861 |
+
"logits/chosen": -1.460861215199083,
|
| 1862 |
+
"logits/rejected": -1.172372178526072,
|
| 1863 |
+
"mean_token_accuracy": 0.6930852761492133,
|
| 1864 |
+
"rewards/chosen": -0.8584173520317563,
|
| 1865 |
+
"rewards/rejected": -2.7061082574073225,
|
| 1866 |
+
"rewards/accuracies": 0.85625,
|
| 1867 |
+
"rewards/margins": 1.8476909057237207,
|
| 1868 |
+
"logps/chosen": -68.6634070277214,
|
| 1869 |
+
"logps/rejected": -81.45435483455658,
|
| 1870 |
+
"epoch": 1.6320000000000001,
|
| 1871 |
+
"step": 1020
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"loss": 0.4709066390991211,
|
| 1875 |
+
"grad_norm": 0.228103369474411,
|
| 1876 |
+
"learning_rate": 1.7680000000000003e-06,
|
| 1877 |
+
"entropy": 0.8447275205573532,
|
| 1878 |
+
"num_tokens": 13836837.0,
|
| 1879 |
+
"logits/chosen": -1.571289342886968,
|
| 1880 |
+
"logits/rejected": -1.3598711901562217,
|
| 1881 |
+
"mean_token_accuracy": 0.6932502503506839,
|
| 1882 |
+
"rewards/chosen": -1.1255047489772552,
|
| 1883 |
+
"rewards/rejected": -2.439122271069209,
|
| 1884 |
+
"rewards/accuracies": 0.81875,
|
| 1885 |
+
"rewards/margins": 1.3136175224557518,
|
| 1886 |
+
"logps/chosen": -63.986464536190034,
|
| 1887 |
+
"logps/rejected": -74.33912942409515,
|
| 1888 |
+
"epoch": 1.6480000000000001,
|
| 1889 |
+
"step": 1030
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"loss": 0.3415141820907593,
|
| 1893 |
+
"grad_norm": 0.2817641496658325,
|
| 1894 |
+
"learning_rate": 1.6880000000000002e-06,
|
| 1895 |
+
"entropy": 0.8753047725534998,
|
| 1896 |
+
"num_tokens": 13971505.0,
|
| 1897 |
+
"logits/chosen": -1.6041345006828691,
|
| 1898 |
+
"logits/rejected": -1.2299033571160378,
|
| 1899 |
+
"mean_token_accuracy": 0.712097577098757,
|
| 1900 |
+
"rewards/chosen": -1.010352850565687,
|
| 1901 |
+
"rewards/rejected": -2.877846031897934,
|
| 1902 |
+
"rewards/accuracies": 0.8375,
|
| 1903 |
+
"rewards/margins": 1.8674931800924242,
|
| 1904 |
+
"logps/chosen": -70.5002592921257,
|
| 1905 |
+
"logps/rejected": -82.22316973209381,
|
| 1906 |
+
"epoch": 1.6640000000000001,
|
| 1907 |
+
"step": 1040
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"loss": 0.3020626068115234,
|
| 1911 |
+
"grad_norm": 0.1538437157869339,
|
| 1912 |
+
"learning_rate": 1.608e-06,
|
| 1913 |
+
"entropy": 0.8325463001674507,
|
| 1914 |
+
"num_tokens": 14105815.0,
|
| 1915 |
+
"logits/chosen": -1.7390800318885187,
|
| 1916 |
+
"logits/rejected": -1.4494792389981166,
|
| 1917 |
+
"mean_token_accuracy": 0.7042280109599233,
|
| 1918 |
+
"rewards/chosen": -0.9141291501142405,
|
| 1919 |
+
"rewards/rejected": -2.9765128223691137,
|
| 1920 |
+
"rewards/accuracies": 0.8625,
|
| 1921 |
+
"rewards/margins": 2.0623836696147917,
|
| 1922 |
+
"logps/chosen": -61.52977831363678,
|
| 1923 |
+
"logps/rejected": -81.60874741077423,
|
| 1924 |
+
"epoch": 1.6800000000000002,
|
| 1925 |
+
"step": 1050
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"loss": 0.3109905481338501,
|
| 1929 |
+
"grad_norm": 0.15473701059818268,
|
| 1930 |
+
"learning_rate": 1.528e-06,
|
| 1931 |
+
"entropy": 0.8021208431367995,
|
| 1932 |
+
"num_tokens": 14238856.0,
|
| 1933 |
+
"logits/chosen": -1.6674065144174803,
|
| 1934 |
+
"logits/rejected": -1.3240702207297095,
|
| 1935 |
+
"mean_token_accuracy": 0.7186633998528122,
|
| 1936 |
+
"rewards/chosen": -0.7751999277505093,
|
| 1937 |
+
"rewards/rejected": -2.7401396270201075,
|
| 1938 |
+
"rewards/accuracies": 0.88125,
|
| 1939 |
+
"rewards/margins": 1.964939699228853,
|
| 1940 |
+
"logps/chosen": -54.86852971315384,
|
| 1941 |
+
"logps/rejected": -73.9770770072937,
|
| 1942 |
+
"epoch": 1.696,
|
| 1943 |
+
"step": 1060
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"loss": 0.4520918846130371,
|
| 1947 |
+
"grad_norm": 0.23978354036808014,
|
| 1948 |
+
"learning_rate": 1.4480000000000002e-06,
|
| 1949 |
+
"entropy": 0.8903392211504979,
|
| 1950 |
+
"num_tokens": 14373384.0,
|
| 1951 |
+
"logits/chosen": -1.4482252148769281,
|
| 1952 |
+
"logits/rejected": -1.242930829412274,
|
| 1953 |
+
"mean_token_accuracy": 0.7004984401166439,
|
| 1954 |
+
"rewards/chosen": -0.9901666876743548,
|
| 1955 |
+
"rewards/rejected": -2.4927260723896323,
|
| 1956 |
+
"rewards/accuracies": 0.81875,
|
| 1957 |
+
"rewards/margins": 1.502559380698949,
|
| 1958 |
+
"logps/chosen": -65.76897095441818,
|
| 1959 |
+
"logps/rejected": -78.67869209051132,
|
| 1960 |
+
"epoch": 1.712,
|
| 1961 |
+
"step": 1070
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"loss": 0.3588923692703247,
|
| 1965 |
+
"grad_norm": 0.1725820004940033,
|
| 1966 |
+
"learning_rate": 1.368e-06,
|
| 1967 |
+
"entropy": 0.8564222747285385,
|
| 1968 |
+
"num_tokens": 14507731.0,
|
| 1969 |
+
"logits/chosen": -1.6029525860382414,
|
| 1970 |
+
"logits/rejected": -1.3248457362553194,
|
| 1971 |
+
"mean_token_accuracy": 0.705509422160685,
|
| 1972 |
+
"rewards/chosen": -0.92755253186333,
|
| 1973 |
+
"rewards/rejected": -2.6779576923698185,
|
| 1974 |
+
"rewards/accuracies": 0.85625,
|
| 1975 |
+
"rewards/margins": 1.7504051598720252,
|
| 1976 |
+
"logps/chosen": -62.447744405269624,
|
| 1977 |
+
"logps/rejected": -78.27735936641693,
|
| 1978 |
+
"epoch": 1.728,
|
| 1979 |
+
"step": 1080
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"loss": 0.35362203121185304,
|
| 1983 |
+
"grad_norm": 0.28334617614746094,
|
| 1984 |
+
"learning_rate": 1.288e-06,
|
| 1985 |
+
"entropy": 0.861250405901228,
|
| 1986 |
+
"num_tokens": 14642108.0,
|
| 1987 |
+
"logits/chosen": -1.7742959843989723,
|
| 1988 |
+
"logits/rejected": -1.3623826689832286,
|
| 1989 |
+
"mean_token_accuracy": 0.686492515169084,
|
| 1990 |
+
"rewards/chosen": -1.0328600911307149,
|
| 1991 |
+
"rewards/rejected": -3.0031035180203616,
|
| 1992 |
+
"rewards/accuracies": 0.825,
|
| 1993 |
+
"rewards/margins": 1.9702434232458472,
|
| 1994 |
+
"logps/chosen": -62.426606476306915,
|
| 1995 |
+
"logps/rejected": -83.86972826719284,
|
| 1996 |
+
"epoch": 1.744,
|
| 1997 |
+
"step": 1090
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"loss": 0.34779436588287355,
|
| 2001 |
+
"grad_norm": 0.3788210153579712,
|
| 2002 |
+
"learning_rate": 1.2080000000000001e-06,
|
| 2003 |
+
"entropy": 0.8534026202658425,
|
| 2004 |
+
"num_tokens": 14775766.0,
|
| 2005 |
+
"logits/chosen": -1.8180023300144117,
|
| 2006 |
+
"logits/rejected": -1.4538507613172182,
|
| 2007 |
+
"mean_token_accuracy": 0.6866250839084387,
|
| 2008 |
+
"rewards/chosen": -1.1147863603779116,
|
| 2009 |
+
"rewards/rejected": -3.0679479777580125,
|
| 2010 |
+
"rewards/accuracies": 0.85,
|
| 2011 |
+
"rewards/margins": 1.9531616102904081,
|
| 2012 |
+
"logps/chosen": -65.53060195446014,
|
| 2013 |
+
"logps/rejected": -82.48647910356522,
|
| 2014 |
+
"epoch": 1.76,
|
| 2015 |
+
"step": 1100
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"loss": 0.37132487297058103,
|
| 2019 |
+
"grad_norm": 0.19326536357402802,
|
| 2020 |
+
"learning_rate": 1.128e-06,
|
| 2021 |
+
"entropy": 0.9107772405259311,
|
| 2022 |
+
"num_tokens": 14910427.0,
|
| 2023 |
+
"logits/chosen": -1.641197768452253,
|
| 2024 |
+
"logits/rejected": -1.304703577911116,
|
| 2025 |
+
"mean_token_accuracy": 0.7103016471490264,
|
| 2026 |
+
"rewards/chosen": -0.8659923097206047,
|
| 2027 |
+
"rewards/rejected": -2.6321759003447367,
|
| 2028 |
+
"rewards/accuracies": 0.84375,
|
| 2029 |
+
"rewards/margins": 1.7661835953593255,
|
| 2030 |
+
"logps/chosen": -65.0176889181137,
|
| 2031 |
+
"logps/rejected": -79.6421341061592,
|
| 2032 |
+
"epoch": 1.776,
|
| 2033 |
+
"step": 1110
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"loss": 0.3679664611816406,
|
| 2037 |
+
"grad_norm": 0.4326045513153076,
|
| 2038 |
+
"learning_rate": 1.0480000000000002e-06,
|
| 2039 |
+
"entropy": 0.7961695769336075,
|
| 2040 |
+
"num_tokens": 15044039.0,
|
| 2041 |
+
"logits/chosen": -1.9227093861865228,
|
| 2042 |
+
"logits/rejected": -1.5155804088366798,
|
| 2043 |
+
"mean_token_accuracy": 0.6913959834724664,
|
| 2044 |
+
"rewards/chosen": -1.066448407387361,
|
| 2045 |
+
"rewards/rejected": -3.0380996730877086,
|
| 2046 |
+
"rewards/accuracies": 0.825,
|
| 2047 |
+
"rewards/margins": 1.9716512618586421,
|
| 2048 |
+
"logps/chosen": -61.214910900592805,
|
| 2049 |
+
"logps/rejected": -81.22428269386292,
|
| 2050 |
+
"epoch": 1.792,
|
| 2051 |
+
"step": 1120
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"loss": 0.36129958629608155,
|
| 2055 |
+
"grad_norm": 0.22744089365005493,
|
| 2056 |
+
"learning_rate": 9.68e-07,
|
| 2057 |
+
"entropy": 0.8296155019488651,
|
| 2058 |
+
"num_tokens": 15178463.0,
|
| 2059 |
+
"logits/chosen": -1.6523843235216673,
|
| 2060 |
+
"logits/rejected": -1.3291273159510866,
|
| 2061 |
+
"mean_token_accuracy": 0.7189529817551374,
|
| 2062 |
+
"rewards/chosen": -1.2038293910562061,
|
| 2063 |
+
"rewards/rejected": -2.9502542468369937,
|
| 2064 |
+
"rewards/accuracies": 0.825,
|
| 2065 |
+
"rewards/margins": 1.7464248487725853,
|
| 2066 |
+
"logps/chosen": -60.435161435604094,
|
| 2067 |
+
"logps/rejected": -84.77885848283768,
|
| 2068 |
+
"epoch": 1.808,
|
| 2069 |
+
"step": 1130
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"loss": 0.45287551879882815,
|
| 2073 |
+
"grad_norm": 0.3803448975086212,
|
| 2074 |
+
"learning_rate": 8.880000000000001e-07,
|
| 2075 |
+
"entropy": 0.8621988683065865,
|
| 2076 |
+
"num_tokens": 15312634.0,
|
| 2077 |
+
"logits/chosen": -1.8584417536818054,
|
| 2078 |
+
"logits/rejected": -1.3626020514261226,
|
| 2079 |
+
"mean_token_accuracy": 0.6986214431002736,
|
| 2080 |
+
"rewards/chosen": -1.2058987704804167,
|
| 2081 |
+
"rewards/rejected": -2.929702739935601,
|
| 2082 |
+
"rewards/accuracies": 0.7875,
|
| 2083 |
+
"rewards/margins": 1.7238039686344564,
|
| 2084 |
+
"logps/chosen": -63.66816487312317,
|
| 2085 |
+
"logps/rejected": -86.74580391645432,
|
| 2086 |
+
"epoch": 1.8239999999999998,
|
| 2087 |
+
"step": 1140
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"loss": 0.3780132055282593,
|
| 2091 |
+
"grad_norm": 0.20091857016086578,
|
| 2092 |
+
"learning_rate": 8.08e-07,
|
| 2093 |
+
"entropy": 0.7919891161145642,
|
| 2094 |
+
"num_tokens": 15447212.0,
|
| 2095 |
+
"logits/chosen": -1.8383007431632599,
|
| 2096 |
+
"logits/rejected": -1.4778489711588703,
|
| 2097 |
+
"mean_token_accuracy": 0.716796095110476,
|
| 2098 |
+
"rewards/chosen": -1.0603335103718563,
|
| 2099 |
+
"rewards/rejected": -2.9484992748941297,
|
| 2100 |
+
"rewards/accuracies": 0.83125,
|
| 2101 |
+
"rewards/margins": 1.8881657619029284,
|
| 2102 |
+
"logps/chosen": -64.96314210891724,
|
| 2103 |
+
"logps/rejected": -80.53740651607514,
|
| 2104 |
+
"epoch": 1.8399999999999999,
|
| 2105 |
+
"step": 1150
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"loss": 0.3855898380279541,
|
| 2109 |
+
"grad_norm": 0.3632994294166565,
|
| 2110 |
+
"learning_rate": 7.280000000000001e-07,
|
| 2111 |
+
"entropy": 0.8882737910782452,
|
| 2112 |
+
"num_tokens": 15582198.0,
|
| 2113 |
+
"logits/chosen": -1.6513791891715037,
|
| 2114 |
+
"logits/rejected": -1.3469994636936868,
|
| 2115 |
+
"mean_token_accuracy": 0.6885129513218999,
|
| 2116 |
+
"rewards/chosen": -1.0803626882761819,
|
| 2117 |
+
"rewards/rejected": -2.767381212627515,
|
| 2118 |
+
"rewards/accuracies": 0.83125,
|
| 2119 |
+
"rewards/margins": 1.6870185302570462,
|
| 2120 |
+
"logps/chosen": -70.96726630926132,
|
| 2121 |
+
"logps/rejected": -80.29995436668396,
|
| 2122 |
+
"epoch": 1.8559999999999999,
|
| 2123 |
+
"step": 1160
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"loss": 0.3437692642211914,
|
| 2127 |
+
"grad_norm": 0.09972692281007767,
|
| 2128 |
+
"learning_rate": 6.48e-07,
|
| 2129 |
+
"entropy": 0.8239364031498553,
|
| 2130 |
+
"num_tokens": 15716912.0,
|
| 2131 |
+
"logits/chosen": -1.727000862036285,
|
| 2132 |
+
"logits/rejected": -1.3723281799777793,
|
| 2133 |
+
"mean_token_accuracy": 0.6879225814715028,
|
| 2134 |
+
"rewards/chosen": -1.195139434072189,
|
| 2135 |
+
"rewards/rejected": -3.18661116566509,
|
| 2136 |
+
"rewards/accuracies": 0.85625,
|
| 2137 |
+
"rewards/margins": 1.9914717345498503,
|
| 2138 |
+
"logps/chosen": -65.99867243766785,
|
| 2139 |
+
"logps/rejected": -84.84569313526154,
|
| 2140 |
+
"epoch": 1.8719999999999999,
|
| 2141 |
+
"step": 1170
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"loss": 0.42089314460754396,
|
| 2145 |
+
"grad_norm": 0.555910587310791,
|
| 2146 |
+
"learning_rate": 5.680000000000001e-07,
|
| 2147 |
+
"entropy": 0.8244564271823037,
|
| 2148 |
+
"num_tokens": 15851192.0,
|
| 2149 |
+
"logits/chosen": -1.6690832434550242,
|
| 2150 |
+
"logits/rejected": -1.4840573209303312,
|
| 2151 |
+
"mean_token_accuracy": 0.7006650499068201,
|
| 2152 |
+
"rewards/chosen": -1.1593663250401733,
|
| 2153 |
+
"rewards/rejected": -2.811464017489925,
|
| 2154 |
+
"rewards/accuracies": 0.8125,
|
| 2155 |
+
"rewards/margins": 1.6520976910367609,
|
| 2156 |
+
"logps/chosen": -66.40225726366043,
|
| 2157 |
+
"logps/rejected": -75.05194103717804,
|
| 2158 |
+
"epoch": 1.888,
|
| 2159 |
+
"step": 1180
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"loss": 0.3379099130630493,
|
| 2163 |
+
"grad_norm": 0.22328400611877441,
|
| 2164 |
+
"learning_rate": 4.88e-07,
|
| 2165 |
+
"entropy": 0.8442109470954164,
|
| 2166 |
+
"num_tokens": 15986206.0,
|
| 2167 |
+
"logits/chosen": -1.7052544516735195,
|
| 2168 |
+
"logits/rejected": -1.3585597492703503,
|
| 2169 |
+
"mean_token_accuracy": 0.7097283221781254,
|
| 2170 |
+
"rewards/chosen": -1.0867546882946044,
|
| 2171 |
+
"rewards/rejected": -3.0426087318570354,
|
| 2172 |
+
"rewards/accuracies": 0.85,
|
| 2173 |
+
"rewards/margins": 1.9558540463447571,
|
| 2174 |
+
"logps/chosen": -71.25543893575669,
|
| 2175 |
+
"logps/rejected": -80.6093442440033,
|
| 2176 |
+
"epoch": 1.904,
|
| 2177 |
+
"step": 1190
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"loss": 0.4386793613433838,
|
| 2181 |
+
"grad_norm": 0.1305859386920929,
|
| 2182 |
+
"learning_rate": 4.0800000000000005e-07,
|
| 2183 |
+
"entropy": 0.8031474264746066,
|
| 2184 |
+
"num_tokens": 16119991.0,
|
| 2185 |
+
"logits/chosen": -1.7559797354444115,
|
| 2186 |
+
"logits/rejected": -1.4829372292152603,
|
| 2187 |
+
"mean_token_accuracy": 0.6981100890785455,
|
| 2188 |
+
"rewards/chosen": -1.213875091643422,
|
| 2189 |
+
"rewards/rejected": -2.8735405456507577,
|
| 2190 |
+
"rewards/accuracies": 0.85,
|
| 2191 |
+
"rewards/margins": 1.6596654396969825,
|
| 2192 |
+
"logps/chosen": -65.33509893417359,
|
| 2193 |
+
"logps/rejected": -75.79733974933625,
|
| 2194 |
+
"epoch": 1.92,
|
| 2195 |
+
"step": 1200
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"loss": 0.36259419918060304,
|
| 2199 |
+
"grad_norm": 0.09996015578508377,
|
| 2200 |
+
"learning_rate": 3.280000000000001e-07,
|
| 2201 |
+
"entropy": 0.8677655028168374,
|
| 2202 |
+
"num_tokens": 16254379.0,
|
| 2203 |
+
"logits/chosen": -1.648042040777673,
|
| 2204 |
+
"logits/rejected": -1.3085691421244126,
|
| 2205 |
+
"mean_token_accuracy": 0.7121957898139953,
|
| 2206 |
+
"rewards/chosen": -1.1004725129532744,
|
| 2207 |
+
"rewards/rejected": -2.985144932055846,
|
| 2208 |
+
"rewards/accuracies": 0.85625,
|
| 2209 |
+
"rewards/margins": 1.884672424197197,
|
| 2210 |
+
"logps/chosen": -63.39742431640625,
|
| 2211 |
+
"logps/rejected": -85.63640072345734,
|
| 2212 |
+
"epoch": 1.936,
|
| 2213 |
+
"step": 1210
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"loss": 0.4340504169464111,
|
| 2217 |
+
"grad_norm": 0.42559272050857544,
|
| 2218 |
+
"learning_rate": 2.48e-07,
|
| 2219 |
+
"entropy": 0.9353518724892638,
|
| 2220 |
+
"num_tokens": 16389460.0,
|
| 2221 |
+
"logits/chosen": -1.3213048164257681,
|
| 2222 |
+
"logits/rejected": -1.1537002555080433,
|
| 2223 |
+
"mean_token_accuracy": 0.6816943414509297,
|
| 2224 |
+
"rewards/chosen": -1.0376092607082683,
|
| 2225 |
+
"rewards/rejected": -2.6352516803657635,
|
| 2226 |
+
"rewards/accuracies": 0.81875,
|
| 2227 |
+
"rewards/margins": 1.5976424139924348,
|
| 2228 |
+
"logps/chosen": -70.72122876644134,
|
| 2229 |
+
"logps/rejected": -80.32262324094772,
|
| 2230 |
+
"epoch": 1.952,
|
| 2231 |
+
"step": 1220
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"loss": 0.36806693077087405,
|
| 2235 |
+
"grad_norm": 0.2418236881494522,
|
| 2236 |
+
"learning_rate": 1.68e-07,
|
| 2237 |
+
"entropy": 0.8688882581103826,
|
| 2238 |
+
"num_tokens": 16523707.0,
|
| 2239 |
+
"logits/chosen": -1.6066775225775856,
|
| 2240 |
+
"logits/rejected": -1.321082453123463,
|
| 2241 |
+
"mean_token_accuracy": 0.6974495501257479,
|
| 2242 |
+
"rewards/chosen": -1.0087569263385376,
|
| 2243 |
+
"rewards/rejected": -2.7120864556869493,
|
| 2244 |
+
"rewards/accuracies": 0.8375,
|
| 2245 |
+
"rewards/margins": 1.7033295256085694,
|
| 2246 |
+
"logps/chosen": -66.87134427428245,
|
| 2247 |
+
"logps/rejected": -79.92380349636078,
|
| 2248 |
+
"epoch": 1.968,
|
| 2249 |
+
"step": 1230
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"loss": 0.4246889591217041,
|
| 2253 |
+
"grad_norm": 0.3003811538219452,
|
| 2254 |
+
"learning_rate": 8.800000000000001e-08,
|
| 2255 |
+
"entropy": 0.8781503778780462,
|
| 2256 |
+
"num_tokens": 16659265.0,
|
| 2257 |
+
"logits/chosen": -1.631281403137044,
|
| 2258 |
+
"logits/rejected": -1.2568794515214976,
|
| 2259 |
+
"mean_token_accuracy": 0.6956031711772084,
|
| 2260 |
+
"rewards/chosen": -1.1406315140135121,
|
| 2261 |
+
"rewards/rejected": -2.8570110524073242,
|
| 2262 |
+
"rewards/accuracies": 0.83125,
|
| 2263 |
+
"rewards/margins": 1.7163795358501375,
|
| 2264 |
+
"logps/chosen": -70.35265641212463,
|
| 2265 |
+
"logps/rejected": -83.55104370117188,
|
| 2266 |
+
"epoch": 1.984,
|
| 2267 |
+
"step": 1240
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"loss": 0.38119938373565676,
|
| 2271 |
+
"grad_norm": 0.2904031574726105,
|
| 2272 |
+
"learning_rate": 8e-09,
|
| 2273 |
+
"entropy": 0.8772105067997472,
|
| 2274 |
+
"num_tokens": 16794078.0,
|
| 2275 |
+
"logits/chosen": -1.5495304974756305,
|
| 2276 |
+
"logits/rejected": -1.2771007258995903,
|
| 2277 |
+
"mean_token_accuracy": 0.7142244290560484,
|
| 2278 |
+
"rewards/chosen": -1.032747484119318,
|
| 2279 |
+
"rewards/rejected": -2.8586006118333898,
|
| 2280 |
+
"rewards/accuracies": 0.8,
|
| 2281 |
+
"rewards/margins": 1.8258531308267265,
|
| 2282 |
+
"logps/chosen": -65.70732997655868,
|
| 2283 |
+
"logps/rejected": -81.15303890705108,
|
| 2284 |
+
"epoch": 2.0,
|
| 2285 |
+
"step": 1250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"train_runtime": 5990.5564,
|
| 2289 |
+
"train_samples_per_second": 3.339,
|
| 2290 |
+
"train_steps_per_second": 0.209,
|
| 2291 |
+
"total_flos": 1.3612284112597402e+17,
|
| 2292 |
+
"train_loss": 0.4133561391830444,
|
| 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": 23,
|
| 2300 |
+
"validation_monitor_indices": [],
|
| 2301 |
+
"early_stopping_patience": null
|
| 2302 |
+
}
|