Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- D_full_corrected_seed22/adapter_config.json +43 -0
- D_full_corrected_seed22/adapter_model.safetensors +3 -0
- D_full_corrected_seed22/chat_template.jinja +54 -0
- D_full_corrected_seed22/tokenizer.json +3 -0
- D_full_corrected_seed22/tokenizer_config.json +29 -0
- D_full_corrected_seed22/training_summary.json +2302 -0
.gitattributes
CHANGED
|
@@ -35,3 +35,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
D_full_emf1_seed23/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 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
|
D_full_corrected_seed22/adapter_config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"megatron_config": null,
|
| 23 |
+
"megatron_core": "megatron.core",
|
| 24 |
+
"modules_to_save": null,
|
| 25 |
+
"peft_type": "LORA",
|
| 26 |
+
"peft_version": "0.18.1",
|
| 27 |
+
"qalora_group_size": 16,
|
| 28 |
+
"r": 16,
|
| 29 |
+
"rank_pattern": {},
|
| 30 |
+
"revision": null,
|
| 31 |
+
"target_modules": [
|
| 32 |
+
"k_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"v_proj",
|
| 35 |
+
"q_proj"
|
| 36 |
+
],
|
| 37 |
+
"target_parameters": null,
|
| 38 |
+
"task_type": "CAUSAL_LM",
|
| 39 |
+
"trainable_token_indices": null,
|
| 40 |
+
"use_dora": false,
|
| 41 |
+
"use_qalora": false,
|
| 42 |
+
"use_rslora": false
|
| 43 |
+
}
|
D_full_corrected_seed22/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:774ff089fc0f49a6afa930b0e75ae3a848e49c322bf44f47bec4b0058613489a
|
| 3 |
+
size 17462432
|
D_full_corrected_seed22/chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
D_full_corrected_seed22/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b4360dd6a184650ffc48056c2569bc603f896c5adfe94b10f1c79f809638aa5
|
| 3 |
+
size 11422166
|
D_full_corrected_seed22/tokenizer_config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": true,
|
| 24 |
+
"model_max_length": 131072,
|
| 25 |
+
"pad_token": "<|endoftext|>",
|
| 26 |
+
"split_special_tokens": false,
|
| 27 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 28 |
+
"unk_token": null
|
| 29 |
+
}
|
D_full_corrected_seed22/training_summary.json
ADDED
|
@@ -0,0 +1,2302 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_name": "Qwen/Qwen2.5-1.5B-Instruct",
|
| 3 |
+
"resolved_model": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 4 |
+
"num_train_records": 10000,
|
| 5 |
+
"num_validation_records_available": 0,
|
| 6 |
+
"training_config": {
|
| 7 |
+
"lora_rank": 16,
|
| 8 |
+
"lora_alpha": 32,
|
| 9 |
+
"learning_rate": 1e-05,
|
| 10 |
+
"num_epochs": 2,
|
| 11 |
+
"per_device_batch_size": 1,
|
| 12 |
+
"per_device_eval_batch_size": 1,
|
| 13 |
+
"gradient_accumulation_steps": 16,
|
| 14 |
+
"beta": 0.1,
|
| 15 |
+
"max_length": 512,
|
| 16 |
+
"max_prompt_length": 384,
|
| 17 |
+
"seed": 22,
|
| 18 |
+
"save_steps": 250,
|
| 19 |
+
"save_total_limit": 2,
|
| 20 |
+
"resume_from_checkpoint": null,
|
| 21 |
+
"precision_dtype": "torch.bfloat16",
|
| 22 |
+
"bf16": true,
|
| 23 |
+
"fp16": false,
|
| 24 |
+
"tokenizer_add_bos_token": false
|
| 25 |
+
},
|
| 26 |
+
"global_step": 1250,
|
| 27 |
+
"best_metric": null,
|
| 28 |
+
"best_model_checkpoint": null,
|
| 29 |
+
"train_metrics": {
|
| 30 |
+
"train_runtime": 5988.37,
|
| 31 |
+
"train_samples_per_second": 3.34,
|
| 32 |
+
"train_steps_per_second": 0.209,
|
| 33 |
+
"total_flos": 1.3612284112597402e+17,
|
| 34 |
+
"train_loss": 0.41622718410491943
|
| 35 |
+
},
|
| 36 |
+
"log_history": [
|
| 37 |
+
{
|
| 38 |
+
"loss": 0.687640380859375,
|
| 39 |
+
"grad_norm": 0.1060064360499382,
|
| 40 |
+
"learning_rate": 9.928e-06,
|
| 41 |
+
"entropy": 0.8565638590604067,
|
| 42 |
+
"num_tokens": 134743.0,
|
| 43 |
+
"logits/chosen": -0.9762120088020098,
|
| 44 |
+
"logits/rejected": -0.9701892853926462,
|
| 45 |
+
"mean_token_accuracy": 0.73006621170789,
|
| 46 |
+
"rewards/chosen": 0.0025454330782849867,
|
| 47 |
+
"rewards/rejected": -0.008883580170868299,
|
| 48 |
+
"rewards/accuracies": 0.5625,
|
| 49 |
+
"rewards/margins": 0.011429013135057176,
|
| 50 |
+
"logps/chosen": -57.36903718709946,
|
| 51 |
+
"logps/rejected": -54.170119202136995,
|
| 52 |
+
"epoch": 0.016,
|
| 53 |
+
"step": 10
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"loss": 0.6806084156036377,
|
| 57 |
+
"grad_norm": 0.11952455341815948,
|
| 58 |
+
"learning_rate": 9.848000000000001e-06,
|
| 59 |
+
"entropy": 0.8075785945402458,
|
| 60 |
+
"num_tokens": 269598.0,
|
| 61 |
+
"logits/chosen": -0.9017632693546302,
|
| 62 |
+
"logits/rejected": -0.9173937418731543,
|
| 63 |
+
"mean_token_accuracy": 0.7327126525342464,
|
| 64 |
+
"rewards/chosen": -0.01022402061316825,
|
| 65 |
+
"rewards/rejected": -0.036268199003461635,
|
| 66 |
+
"rewards/accuracies": 0.69375,
|
| 67 |
+
"rewards/margins": 0.02604417833717889,
|
| 68 |
+
"logps/chosen": -57.75539029836655,
|
| 69 |
+
"logps/rejected": -51.10095466375351,
|
| 70 |
+
"epoch": 0.032,
|
| 71 |
+
"step": 20
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"loss": 0.6647059917449951,
|
| 75 |
+
"grad_norm": 0.11816244572401047,
|
| 76 |
+
"learning_rate": 9.768e-06,
|
| 77 |
+
"entropy": 0.830331975640729,
|
| 78 |
+
"num_tokens": 404333.0,
|
| 79 |
+
"logits/chosen": -1.0667518885409581,
|
| 80 |
+
"logits/rejected": -0.9044131975420534,
|
| 81 |
+
"mean_token_accuracy": 0.7304680673405528,
|
| 82 |
+
"rewards/chosen": -0.03215525241475916,
|
| 83 |
+
"rewards/rejected": -0.09206540035329454,
|
| 84 |
+
"rewards/accuracies": 0.775,
|
| 85 |
+
"rewards/margins": 0.05991014770988841,
|
| 86 |
+
"logps/chosen": -53.41478923559189,
|
| 87 |
+
"logps/rejected": -58.32939873933792,
|
| 88 |
+
"epoch": 0.048,
|
| 89 |
+
"step": 30
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"loss": 0.6360646724700928,
|
| 93 |
+
"grad_norm": 0.11358040571212769,
|
| 94 |
+
"learning_rate": 9.688000000000001e-06,
|
| 95 |
+
"entropy": 0.8251496440032497,
|
| 96 |
+
"num_tokens": 538701.0,
|
| 97 |
+
"logits/chosen": -1.002133708392577,
|
| 98 |
+
"logits/rejected": -0.9339561312410426,
|
| 99 |
+
"mean_token_accuracy": 0.737026778422296,
|
| 100 |
+
"rewards/chosen": -0.0684578845683177,
|
| 101 |
+
"rewards/rejected": -0.1924211388861295,
|
| 102 |
+
"rewards/accuracies": 0.84375,
|
| 103 |
+
"rewards/margins": 0.12396325423032976,
|
| 104 |
+
"logps/chosen": -51.08748831748962,
|
| 105 |
+
"logps/rejected": -52.612894320487975,
|
| 106 |
+
"epoch": 0.064,
|
| 107 |
+
"step": 40
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"loss": 0.6088774681091309,
|
| 111 |
+
"grad_norm": 0.1321704089641571,
|
| 112 |
+
"learning_rate": 9.608e-06,
|
| 113 |
+
"entropy": 0.8403675278299488,
|
| 114 |
+
"num_tokens": 673524.0,
|
| 115 |
+
"logits/chosen": -0.9297918675282997,
|
| 116 |
+
"logits/rejected": -0.8459581713260917,
|
| 117 |
+
"mean_token_accuracy": 0.7066086169332266,
|
| 118 |
+
"rewards/chosen": -0.1453652457725184,
|
| 119 |
+
"rewards/rejected": -0.34192078668165776,
|
| 120 |
+
"rewards/accuracies": 0.76875,
|
| 121 |
+
"rewards/margins": 0.19655554031487554,
|
| 122 |
+
"logps/chosen": -60.78850121498108,
|
| 123 |
+
"logps/rejected": -58.992058432102205,
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"step": 50
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"loss": 0.5826151847839356,
|
| 129 |
+
"grad_norm": 0.11255195736885071,
|
| 130 |
+
"learning_rate": 9.528000000000001e-06,
|
| 131 |
+
"entropy": 0.8106186802498996,
|
| 132 |
+
"num_tokens": 808013.0,
|
| 133 |
+
"logits/chosen": -0.9986341208486369,
|
| 134 |
+
"logits/rejected": -0.8505068748346585,
|
| 135 |
+
"mean_token_accuracy": 0.7220808529295027,
|
| 136 |
+
"rewards/chosen": -0.24798050145618616,
|
| 137 |
+
"rewards/rejected": -0.5257746596296784,
|
| 138 |
+
"rewards/accuracies": 0.8125,
|
| 139 |
+
"rewards/margins": 0.27779415745753794,
|
| 140 |
+
"logps/chosen": -53.93427662849426,
|
| 141 |
+
"logps/rejected": -60.885983073711394,
|
| 142 |
+
"epoch": 0.096,
|
| 143 |
+
"step": 60
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"loss": 0.5510534763336181,
|
| 147 |
+
"grad_norm": 0.12085709720849991,
|
| 148 |
+
"learning_rate": 9.448e-06,
|
| 149 |
+
"entropy": 0.809669681743253,
|
| 150 |
+
"num_tokens": 942494.0,
|
| 151 |
+
"logits/chosen": -1.007010350692191,
|
| 152 |
+
"logits/rejected": -0.9043599711007537,
|
| 153 |
+
"mean_token_accuracy": 0.7091988798230886,
|
| 154 |
+
"rewards/chosen": -0.3718697968673951,
|
| 155 |
+
"rewards/rejected": -0.7532227354706265,
|
| 156 |
+
"rewards/accuracies": 0.79375,
|
| 157 |
+
"rewards/margins": 0.38135293698869643,
|
| 158 |
+
"logps/chosen": -63.80851249694824,
|
| 159 |
+
"logps/rejected": -57.738853216171265,
|
| 160 |
+
"epoch": 0.112,
|
| 161 |
+
"step": 70
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"loss": 0.5266392707824707,
|
| 165 |
+
"grad_norm": 0.1074029728770256,
|
| 166 |
+
"learning_rate": 9.368e-06,
|
| 167 |
+
"entropy": 0.822804095945321,
|
| 168 |
+
"num_tokens": 1076339.0,
|
| 169 |
+
"logits/chosen": -1.0472984386066868,
|
| 170 |
+
"logits/rejected": -0.8199694184864057,
|
| 171 |
+
"mean_token_accuracy": 0.718729289341718,
|
| 172 |
+
"rewards/chosen": -0.5515795681139934,
|
| 173 |
+
"rewards/rejected": -1.0349310390651225,
|
| 174 |
+
"rewards/accuracies": 0.76875,
|
| 175 |
+
"rewards/margins": 0.483351469039917,
|
| 176 |
+
"logps/chosen": -53.268451523780826,
|
| 177 |
+
"logps/rejected": -66.34326626062393,
|
| 178 |
+
"epoch": 0.128,
|
| 179 |
+
"step": 80
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"loss": 0.5478773593902588,
|
| 183 |
+
"grad_norm": 0.16121409833431244,
|
| 184 |
+
"learning_rate": 9.288e-06,
|
| 185 |
+
"entropy": 0.7804107027535793,
|
| 186 |
+
"num_tokens": 1210662.0,
|
| 187 |
+
"logits/chosen": -1.1064176386508637,
|
| 188 |
+
"logits/rejected": -0.9639100541229437,
|
| 189 |
+
"mean_token_accuracy": 0.7135749334469438,
|
| 190 |
+
"rewards/chosen": -0.72423874346714,
|
| 191 |
+
"rewards/rejected": -1.227196942118462,
|
| 192 |
+
"rewards/accuracies": 0.75,
|
| 193 |
+
"rewards/margins": 0.5029582012444734,
|
| 194 |
+
"logps/chosen": -62.34713068008423,
|
| 195 |
+
"logps/rejected": -62.778570568561555,
|
| 196 |
+
"epoch": 0.144,
|
| 197 |
+
"step": 90
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"loss": 0.5251368999481201,
|
| 201 |
+
"grad_norm": 0.12899401783943176,
|
| 202 |
+
"learning_rate": 9.208e-06,
|
| 203 |
+
"entropy": 0.7937996337132063,
|
| 204 |
+
"num_tokens": 1345120.0,
|
| 205 |
+
"logits/chosen": -1.1104136710992427,
|
| 206 |
+
"logits/rejected": -0.9978494176751795,
|
| 207 |
+
"mean_token_accuracy": 0.7030816830694675,
|
| 208 |
+
"rewards/chosen": -0.8510917660896666,
|
| 209 |
+
"rewards/rejected": -1.4670812282594852,
|
| 210 |
+
"rewards/accuracies": 0.75,
|
| 211 |
+
"rewards/margins": 0.6159894620068371,
|
| 212 |
+
"logps/chosen": -62.66893690824509,
|
| 213 |
+
"logps/rejected": -65.58398360013962,
|
| 214 |
+
"epoch": 0.16,
|
| 215 |
+
"step": 100
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"loss": 0.4492000102996826,
|
| 219 |
+
"grad_norm": 0.1406930536031723,
|
| 220 |
+
"learning_rate": 9.128e-06,
|
| 221 |
+
"entropy": 0.740898304147413,
|
| 222 |
+
"num_tokens": 1479281.0,
|
| 223 |
+
"logits/chosen": -1.2706884151573026,
|
| 224 |
+
"logits/rejected": -1.124058169901609,
|
| 225 |
+
"mean_token_accuracy": 0.7055532153695822,
|
| 226 |
+
"rewards/chosen": -0.8618847452569753,
|
| 227 |
+
"rewards/rejected": -1.6921822651755065,
|
| 228 |
+
"rewards/accuracies": 0.81875,
|
| 229 |
+
"rewards/margins": 0.8302975169382989,
|
| 230 |
+
"logps/chosen": -62.41115806102753,
|
| 231 |
+
"logps/rejected": -69.8843122601509,
|
| 232 |
+
"epoch": 0.176,
|
| 233 |
+
"step": 110
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"loss": 0.5015526294708252,
|
| 237 |
+
"grad_norm": 0.19250263273715973,
|
| 238 |
+
"learning_rate": 9.048e-06,
|
| 239 |
+
"entropy": 0.7727364487247541,
|
| 240 |
+
"num_tokens": 1613851.0,
|
| 241 |
+
"logits/chosen": -1.2708980964300172,
|
| 242 |
+
"logits/rejected": -1.1060235384368617,
|
| 243 |
+
"mean_token_accuracy": 0.7058340331539512,
|
| 244 |
+
"rewards/chosen": -1.1178167212291783,
|
| 245 |
+
"rewards/rejected": -1.9005187458358705,
|
| 246 |
+
"rewards/accuracies": 0.7625,
|
| 247 |
+
"rewards/margins": 0.7827020254917443,
|
| 248 |
+
"logps/chosen": -68.03103417158127,
|
| 249 |
+
"logps/rejected": -74.40341311693192,
|
| 250 |
+
"epoch": 0.192,
|
| 251 |
+
"step": 120
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"loss": 0.46552538871765137,
|
| 255 |
+
"grad_norm": 0.10027917474508286,
|
| 256 |
+
"learning_rate": 8.968000000000001e-06,
|
| 257 |
+
"entropy": 0.7614070464973338,
|
| 258 |
+
"num_tokens": 1748368.0,
|
| 259 |
+
"logits/chosen": -1.3760571993000903,
|
| 260 |
+
"logits/rejected": -1.143091703786983,
|
| 261 |
+
"mean_token_accuracy": 0.7143012292683124,
|
| 262 |
+
"rewards/chosen": -1.0495449592126533,
|
| 263 |
+
"rewards/rejected": -2.0446871710009873,
|
| 264 |
+
"rewards/accuracies": 0.79375,
|
| 265 |
+
"rewards/margins": 0.9951422128826379,
|
| 266 |
+
"logps/chosen": -64.35170601606369,
|
| 267 |
+
"logps/rejected": -76.99314639568328,
|
| 268 |
+
"epoch": 0.208,
|
| 269 |
+
"step": 130
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"loss": 0.436888313293457,
|
| 273 |
+
"grad_norm": 0.14637508988380432,
|
| 274 |
+
"learning_rate": 8.888e-06,
|
| 275 |
+
"entropy": 0.8021846901858225,
|
| 276 |
+
"num_tokens": 1883496.0,
|
| 277 |
+
"logits/chosen": -1.3044026544607896,
|
| 278 |
+
"logits/rejected": -1.080970107989037,
|
| 279 |
+
"mean_token_accuracy": 0.7037744291126729,
|
| 280 |
+
"rewards/chosen": -1.1526502840570174,
|
| 281 |
+
"rewards/rejected": -2.3245541878044604,
|
| 282 |
+
"rewards/accuracies": 0.80625,
|
| 283 |
+
"rewards/margins": 1.171903900266625,
|
| 284 |
+
"logps/chosen": -68.41839230060577,
|
| 285 |
+
"logps/rejected": -81.16507298946381,
|
| 286 |
+
"epoch": 0.224,
|
| 287 |
+
"step": 140
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"loss": 0.48706469535827634,
|
| 291 |
+
"grad_norm": 0.1303538978099823,
|
| 292 |
+
"learning_rate": 8.808000000000001e-06,
|
| 293 |
+
"entropy": 0.7772014890098944,
|
| 294 |
+
"num_tokens": 2017285.0,
|
| 295 |
+
"logits/chosen": -1.5057701161807828,
|
| 296 |
+
"logits/rejected": -1.192066047114015,
|
| 297 |
+
"mean_token_accuracy": 0.7042934297583997,
|
| 298 |
+
"rewards/chosen": -1.2129425552935573,
|
| 299 |
+
"rewards/rejected": -2.1681987981253767,
|
| 300 |
+
"rewards/accuracies": 0.76875,
|
| 301 |
+
"rewards/margins": 0.9552562347613275,
|
| 302 |
+
"logps/chosen": -60.23983110189438,
|
| 303 |
+
"logps/rejected": -75.58166043758392,
|
| 304 |
+
"epoch": 0.24,
|
| 305 |
+
"step": 150
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"loss": 0.5098927021026611,
|
| 309 |
+
"grad_norm": 0.14422298967838287,
|
| 310 |
+
"learning_rate": 8.728e-06,
|
| 311 |
+
"entropy": 0.7005686257762136,
|
| 312 |
+
"num_tokens": 2150045.0,
|
| 313 |
+
"logits/chosen": -1.4558571597747738,
|
| 314 |
+
"logits/rejected": -1.2716753484632168,
|
| 315 |
+
"mean_token_accuracy": 0.6975366961210966,
|
| 316 |
+
"rewards/chosen": -1.2024810665636323,
|
| 317 |
+
"rewards/rejected": -1.950557324080728,
|
| 318 |
+
"rewards/accuracies": 0.8,
|
| 319 |
+
"rewards/margins": 0.7480762558057904,
|
| 320 |
+
"logps/chosen": -59.91471302509308,
|
| 321 |
+
"logps/rejected": -63.08780847787857,
|
| 322 |
+
"epoch": 0.256,
|
| 323 |
+
"step": 160
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"loss": 0.5322403907775879,
|
| 327 |
+
"grad_norm": 0.20070654153823853,
|
| 328 |
+
"learning_rate": 8.648000000000001e-06,
|
| 329 |
+
"entropy": 0.8089895805169363,
|
| 330 |
+
"num_tokens": 2284628.0,
|
| 331 |
+
"logits/chosen": -1.3888271276307687,
|
| 332 |
+
"logits/rejected": -1.1132356974860957,
|
| 333 |
+
"mean_token_accuracy": 0.6912105225026608,
|
| 334 |
+
"rewards/chosen": -1.359617133741267,
|
| 335 |
+
"rewards/rejected": -2.1929144957102835,
|
| 336 |
+
"rewards/accuracies": 0.725,
|
| 337 |
+
"rewards/margins": 0.833297361060977,
|
| 338 |
+
"logps/chosen": -67.08332798480987,
|
| 339 |
+
"logps/rejected": -77.29410004615784,
|
| 340 |
+
"epoch": 0.272,
|
| 341 |
+
"step": 170
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"loss": 0.4481250286102295,
|
| 345 |
+
"grad_norm": 0.15456373989582062,
|
| 346 |
+
"learning_rate": 8.568e-06,
|
| 347 |
+
"entropy": 0.7180120348231867,
|
| 348 |
+
"num_tokens": 2418154.0,
|
| 349 |
+
"logits/chosen": -1.4233392348074516,
|
| 350 |
+
"logits/rejected": -1.1661942213333962,
|
| 351 |
+
"mean_token_accuracy": 0.7062221618369222,
|
| 352 |
+
"rewards/chosen": -1.1455491209286266,
|
| 353 |
+
"rewards/rejected": -2.2020440079271792,
|
| 354 |
+
"rewards/accuracies": 0.775,
|
| 355 |
+
"rewards/margins": 1.0564948851242661,
|
| 356 |
+
"logps/chosen": -60.87348268032074,
|
| 357 |
+
"logps/rejected": -72.09199923276901,
|
| 358 |
+
"epoch": 0.288,
|
| 359 |
+
"step": 180
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"loss": 0.4390109539031982,
|
| 363 |
+
"grad_norm": 0.16209152340888977,
|
| 364 |
+
"learning_rate": 8.488e-06,
|
| 365 |
+
"entropy": 0.7668704556766898,
|
| 366 |
+
"num_tokens": 2551816.0,
|
| 367 |
+
"logits/chosen": -1.247304855758104,
|
| 368 |
+
"logits/rejected": -1.0460455366518644,
|
| 369 |
+
"mean_token_accuracy": 0.7004121996462345,
|
| 370 |
+
"rewards/chosen": -1.2016246612009127,
|
| 371 |
+
"rewards/rejected": -2.227105630352162,
|
| 372 |
+
"rewards/accuracies": 0.84375,
|
| 373 |
+
"rewards/margins": 1.025480966223404,
|
| 374 |
+
"logps/chosen": -67.76577914953232,
|
| 375 |
+
"logps/rejected": -70.5734748005867,
|
| 376 |
+
"epoch": 0.304,
|
| 377 |
+
"step": 190
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"loss": 0.47978858947753905,
|
| 381 |
+
"grad_norm": 0.16654466092586517,
|
| 382 |
+
"learning_rate": 8.408e-06,
|
| 383 |
+
"entropy": 0.849099058099091,
|
| 384 |
+
"num_tokens": 2686525.0,
|
| 385 |
+
"logits/chosen": -1.2724904274761242,
|
| 386 |
+
"logits/rejected": -0.9709807192389887,
|
| 387 |
+
"mean_token_accuracy": 0.6885419142432511,
|
| 388 |
+
"rewards/chosen": -1.4079357270762558,
|
| 389 |
+
"rewards/rejected": -2.426362790848361,
|
| 390 |
+
"rewards/accuracies": 0.79375,
|
| 391 |
+
"rewards/margins": 1.0184270616620779,
|
| 392 |
+
"logps/chosen": -67.60490100383758,
|
| 393 |
+
"logps/rejected": -79.58337481021881,
|
| 394 |
+
"epoch": 0.32,
|
| 395 |
+
"step": 200
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"loss": 0.4578042984008789,
|
| 399 |
+
"grad_norm": 0.299275666475296,
|
| 400 |
+
"learning_rate": 8.328e-06,
|
| 401 |
+
"entropy": 0.8540181846939958,
|
| 402 |
+
"num_tokens": 2821412.0,
|
| 403 |
+
"logits/chosen": -1.1864729170701347,
|
| 404 |
+
"logits/rejected": -0.9097683475095948,
|
| 405 |
+
"mean_token_accuracy": 0.6739382394123823,
|
| 406 |
+
"rewards/chosen": -1.4339511796366424,
|
| 407 |
+
"rewards/rejected": -2.5972983174491673,
|
| 408 |
+
"rewards/accuracies": 0.7625,
|
| 409 |
+
"rewards/margins": 1.1633471305016427,
|
| 410 |
+
"logps/chosen": -68.65028207302093,
|
| 411 |
+
"logps/rejected": -81.62369173765182,
|
| 412 |
+
"epoch": 0.336,
|
| 413 |
+
"step": 210
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"loss": 0.48204550743103025,
|
| 417 |
+
"grad_norm": 0.12565219402313232,
|
| 418 |
+
"learning_rate": 8.248e-06,
|
| 419 |
+
"entropy": 0.8143817926684278,
|
| 420 |
+
"num_tokens": 2956627.0,
|
| 421 |
+
"logits/chosen": -1.2639339689459161,
|
| 422 |
+
"logits/rejected": -0.9821035544699896,
|
| 423 |
+
"mean_token_accuracy": 0.6959543541073799,
|
| 424 |
+
"rewards/chosen": -1.2881334296544082,
|
| 425 |
+
"rewards/rejected": -2.4724910952383654,
|
| 426 |
+
"rewards/accuracies": 0.81875,
|
| 427 |
+
"rewards/margins": 1.1843576611950994,
|
| 428 |
+
"logps/chosen": -65.9024406671524,
|
| 429 |
+
"logps/rejected": -82.27175046205521,
|
| 430 |
+
"epoch": 0.352,
|
| 431 |
+
"step": 220
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"loss": 0.417206859588623,
|
| 435 |
+
"grad_norm": 0.1738692671060562,
|
| 436 |
+
"learning_rate": 8.168e-06,
|
| 437 |
+
"entropy": 0.8036025390363648,
|
| 438 |
+
"num_tokens": 3090619.0,
|
| 439 |
+
"logits/chosen": -1.2377200494324052,
|
| 440 |
+
"logits/rejected": -0.9386820201339596,
|
| 441 |
+
"mean_token_accuracy": 0.6922742452472448,
|
| 442 |
+
"rewards/chosen": -1.1464276203303598,
|
| 443 |
+
"rewards/rejected": -2.262399673461914,
|
| 444 |
+
"rewards/accuracies": 0.8125,
|
| 445 |
+
"rewards/margins": 1.1159720475785435,
|
| 446 |
+
"logps/chosen": -61.8274448633194,
|
| 447 |
+
"logps/rejected": -73.91179895401001,
|
| 448 |
+
"epoch": 0.368,
|
| 449 |
+
"step": 230
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"loss": 0.43752446174621584,
|
| 453 |
+
"grad_norm": 0.23912598192691803,
|
| 454 |
+
"learning_rate": 8.088e-06,
|
| 455 |
+
"entropy": 0.8557709640779649,
|
| 456 |
+
"num_tokens": 3225094.0,
|
| 457 |
+
"logits/chosen": -1.207070475111356,
|
| 458 |
+
"logits/rejected": -1.0002472519775074,
|
| 459 |
+
"mean_token_accuracy": 0.6910314219072461,
|
| 460 |
+
"rewards/chosen": -1.3286611944553441,
|
| 461 |
+
"rewards/rejected": -2.4035517456475644,
|
| 462 |
+
"rewards/accuracies": 0.81875,
|
| 463 |
+
"rewards/margins": 1.0748905479907989,
|
| 464 |
+
"logps/chosen": -70.01648730039597,
|
| 465 |
+
"logps/rejected": -78.90722476243972,
|
| 466 |
+
"epoch": 0.384,
|
| 467 |
+
"step": 240
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"loss": 0.5086671352386475,
|
| 471 |
+
"grad_norm": 0.11762253940105438,
|
| 472 |
+
"learning_rate": 8.008e-06,
|
| 473 |
+
"entropy": 0.7838629280799069,
|
| 474 |
+
"num_tokens": 3360182.0,
|
| 475 |
+
"logits/chosen": -1.3827117551449377,
|
| 476 |
+
"logits/rejected": -1.1543053421574434,
|
| 477 |
+
"mean_token_accuracy": 0.6989276727661491,
|
| 478 |
+
"rewards/chosen": -1.30892960667843,
|
| 479 |
+
"rewards/rejected": -2.295319798681885,
|
| 480 |
+
"rewards/accuracies": 0.7625,
|
| 481 |
+
"rewards/margins": 0.9863901949487627,
|
| 482 |
+
"logps/chosen": -69.07811139822006,
|
| 483 |
+
"logps/rejected": -74.69848490953446,
|
| 484 |
+
"epoch": 0.4,
|
| 485 |
+
"step": 250
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"loss": 0.4053815364837646,
|
| 489 |
+
"grad_norm": 0.13419465720653534,
|
| 490 |
+
"learning_rate": 7.928e-06,
|
| 491 |
+
"entropy": 0.7642714152345433,
|
| 492 |
+
"num_tokens": 3494544.0,
|
| 493 |
+
"logits/chosen": -1.4618686984229707,
|
| 494 |
+
"logits/rejected": -1.1528846002086408,
|
| 495 |
+
"mean_token_accuracy": 0.6982316529378295,
|
| 496 |
+
"rewards/chosen": -1.1799098454997874,
|
| 497 |
+
"rewards/rejected": -2.3433254302304705,
|
| 498 |
+
"rewards/accuracies": 0.81875,
|
| 499 |
+
"rewards/margins": 1.1634155815467238,
|
| 500 |
+
"logps/chosen": -64.5714376449585,
|
| 501 |
+
"logps/rejected": -75.23598402738571,
|
| 502 |
+
"epoch": 0.416,
|
| 503 |
+
"step": 260
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"loss": 0.42232260704040525,
|
| 507 |
+
"grad_norm": 0.1312924027442932,
|
| 508 |
+
"learning_rate": 7.848000000000002e-06,
|
| 509 |
+
"entropy": 0.807637597725261,
|
| 510 |
+
"num_tokens": 3628021.0,
|
| 511 |
+
"logits/chosen": -1.394175250623325,
|
| 512 |
+
"logits/rejected": -1.0996939451010583,
|
| 513 |
+
"mean_token_accuracy": 0.6931433760561049,
|
| 514 |
+
"rewards/chosen": -1.2270895831286908,
|
| 515 |
+
"rewards/rejected": -2.40001835685689,
|
| 516 |
+
"rewards/accuracies": 0.81875,
|
| 517 |
+
"rewards/margins": 1.172928768955171,
|
| 518 |
+
"logps/chosen": -65.94223583936692,
|
| 519 |
+
"logps/rejected": -71.70688098669052,
|
| 520 |
+
"epoch": 0.432,
|
| 521 |
+
"step": 270
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"loss": 0.4506784439086914,
|
| 525 |
+
"grad_norm": 0.15292952954769135,
|
| 526 |
+
"learning_rate": 7.768e-06,
|
| 527 |
+
"entropy": 0.8501870285661426,
|
| 528 |
+
"num_tokens": 3762769.0,
|
| 529 |
+
"logits/chosen": -1.3672126839983112,
|
| 530 |
+
"logits/rejected": -1.022872153260842,
|
| 531 |
+
"mean_token_accuracy": 0.7088376709260047,
|
| 532 |
+
"rewards/chosen": -1.2467133824247867,
|
| 533 |
+
"rewards/rejected": -2.436103739729151,
|
| 534 |
+
"rewards/accuracies": 0.78125,
|
| 535 |
+
"rewards/margins": 1.1893903512507678,
|
| 536 |
+
"logps/chosen": -63.230743324756624,
|
| 537 |
+
"logps/rejected": -81.51095448732376,
|
| 538 |
+
"epoch": 0.448,
|
| 539 |
+
"step": 280
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"loss": 0.4967855453491211,
|
| 543 |
+
"grad_norm": 0.09499868005514145,
|
| 544 |
+
"learning_rate": 7.688000000000002e-06,
|
| 545 |
+
"entropy": 0.7845427984429989,
|
| 546 |
+
"num_tokens": 3896554.0,
|
| 547 |
+
"logits/chosen": -1.4888776858543409,
|
| 548 |
+
"logits/rejected": -1.2712349530477007,
|
| 549 |
+
"mean_token_accuracy": 0.6925253708846867,
|
| 550 |
+
"rewards/chosen": -1.3171632731740828,
|
| 551 |
+
"rewards/rejected": -2.286881530005485,
|
| 552 |
+
"rewards/accuracies": 0.78125,
|
| 553 |
+
"rewards/margins": 0.9697182591073215,
|
| 554 |
+
"logps/chosen": -68.97013219594956,
|
| 555 |
+
"logps/rejected": -71.86911202669144,
|
| 556 |
+
"epoch": 0.464,
|
| 557 |
+
"step": 290
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"loss": 0.4359022617340088,
|
| 561 |
+
"grad_norm": 0.14849627017974854,
|
| 562 |
+
"learning_rate": 7.608000000000001e-06,
|
| 563 |
+
"entropy": 0.7481403223006055,
|
| 564 |
+
"num_tokens": 4030330.0,
|
| 565 |
+
"logits/chosen": -1.620492465090626,
|
| 566 |
+
"logits/rejected": -1.3052264396252262,
|
| 567 |
+
"mean_token_accuracy": 0.7040407178923488,
|
| 568 |
+
"rewards/chosen": -1.2250386900850572,
|
| 569 |
+
"rewards/rejected": -2.5307369497139005,
|
| 570 |
+
"rewards/accuracies": 0.83125,
|
| 571 |
+
"rewards/margins": 1.3056982582435013,
|
| 572 |
+
"logps/chosen": -65.0528399348259,
|
| 573 |
+
"logps/rejected": -76.93463238477707,
|
| 574 |
+
"epoch": 0.48,
|
| 575 |
+
"step": 300
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"loss": 0.41805324554443357,
|
| 579 |
+
"grad_norm": 0.14763571321964264,
|
| 580 |
+
"learning_rate": 7.528000000000001e-06,
|
| 581 |
+
"entropy": 0.7143904635333456,
|
| 582 |
+
"num_tokens": 4163071.0,
|
| 583 |
+
"logits/chosen": -1.6273476694178992,
|
| 584 |
+
"logits/rejected": -1.301192233033906,
|
| 585 |
+
"mean_token_accuracy": 0.719860578328371,
|
| 586 |
+
"rewards/chosen": -1.1687166114104912,
|
| 587 |
+
"rewards/rejected": -2.3252038951730354,
|
| 588 |
+
"rewards/accuracies": 0.84375,
|
| 589 |
+
"rewards/margins": 1.1564872874878347,
|
| 590 |
+
"logps/chosen": -61.4238782286644,
|
| 591 |
+
"logps/rejected": -72.35806931257248,
|
| 592 |
+
"epoch": 0.496,
|
| 593 |
+
"step": 310
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"loss": 0.3954430103302002,
|
| 597 |
+
"grad_norm": 0.12442152202129364,
|
| 598 |
+
"learning_rate": 7.4480000000000005e-06,
|
| 599 |
+
"entropy": 0.793784411335946,
|
| 600 |
+
"num_tokens": 4297650.0,
|
| 601 |
+
"logits/chosen": -1.477777992964485,
|
| 602 |
+
"logits/rejected": -1.1953127751466863,
|
| 603 |
+
"mean_token_accuracy": 0.7043021106161177,
|
| 604 |
+
"rewards/chosen": -0.9529725074127782,
|
| 605 |
+
"rewards/rejected": -2.1367164697730914,
|
| 606 |
+
"rewards/accuracies": 0.875,
|
| 607 |
+
"rewards/margins": 1.1837439638096838,
|
| 608 |
+
"logps/chosen": -68.56799573898316,
|
| 609 |
+
"logps/rejected": -72.77112766504288,
|
| 610 |
+
"epoch": 0.512,
|
| 611 |
+
"step": 320
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"loss": 0.40878615379333494,
|
| 615 |
+
"grad_norm": 0.36143746972084045,
|
| 616 |
+
"learning_rate": 7.3680000000000004e-06,
|
| 617 |
+
"entropy": 0.7595908994378988,
|
| 618 |
+
"num_tokens": 4431577.0,
|
| 619 |
+
"logits/chosen": -1.7031231925665729,
|
| 620 |
+
"logits/rejected": -1.3380552516991515,
|
| 621 |
+
"mean_token_accuracy": 0.6962206202559174,
|
| 622 |
+
"rewards/chosen": -0.9892913607123773,
|
| 623 |
+
"rewards/rejected": -2.246308311697794,
|
| 624 |
+
"rewards/accuracies": 0.83125,
|
| 625 |
+
"rewards/margins": 1.2570169498212636,
|
| 626 |
+
"logps/chosen": -64.16500892639161,
|
| 627 |
+
"logps/rejected": -75.47238931655883,
|
| 628 |
+
"epoch": 0.528,
|
| 629 |
+
"step": 330
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"loss": 0.4154834270477295,
|
| 633 |
+
"grad_norm": 0.14866887032985687,
|
| 634 |
+
"learning_rate": 7.288e-06,
|
| 635 |
+
"entropy": 0.6968162566889078,
|
| 636 |
+
"num_tokens": 4565362.0,
|
| 637 |
+
"logits/chosen": -1.779194939359829,
|
| 638 |
+
"logits/rejected": -1.433946542656388,
|
| 639 |
+
"mean_token_accuracy": 0.7016559707000851,
|
| 640 |
+
"rewards/chosen": -1.1913962944876402,
|
| 641 |
+
"rewards/rejected": -2.540257385186851,
|
| 642 |
+
"rewards/accuracies": 0.8125,
|
| 643 |
+
"rewards/margins": 1.34886109624058,
|
| 644 |
+
"logps/chosen": -64.73984076976777,
|
| 645 |
+
"logps/rejected": -75.1862884759903,
|
| 646 |
+
"epoch": 0.544,
|
| 647 |
+
"step": 340
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"loss": 0.4844213485717773,
|
| 651 |
+
"grad_norm": 0.2365846037864685,
|
| 652 |
+
"learning_rate": 7.208e-06,
|
| 653 |
+
"entropy": 0.7582426005683374,
|
| 654 |
+
"num_tokens": 4699278.0,
|
| 655 |
+
"logits/chosen": -1.5956848837508553,
|
| 656 |
+
"logits/rejected": -1.3307987742757406,
|
| 657 |
+
"mean_token_accuracy": 0.7080934265628457,
|
| 658 |
+
"rewards/chosen": -1.1894279563051895,
|
| 659 |
+
"rewards/rejected": -2.359440893511055,
|
| 660 |
+
"rewards/accuracies": 0.8,
|
| 661 |
+
"rewards/margins": 1.1700129341334105,
|
| 662 |
+
"logps/chosen": -61.314231514930725,
|
| 663 |
+
"logps/rejected": -74.14340764284134,
|
| 664 |
+
"epoch": 0.56,
|
| 665 |
+
"step": 350
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"loss": 0.3694411516189575,
|
| 669 |
+
"grad_norm": 0.12817151844501495,
|
| 670 |
+
"learning_rate": 7.128e-06,
|
| 671 |
+
"entropy": 0.7540120068471878,
|
| 672 |
+
"num_tokens": 4834112.0,
|
| 673 |
+
"logits/chosen": -1.5378615143259888,
|
| 674 |
+
"logits/rejected": -1.3836281190488564,
|
| 675 |
+
"mean_token_accuracy": 0.7073088637553155,
|
| 676 |
+
"rewards/chosen": -1.0913370640162612,
|
| 677 |
+
"rewards/rejected": -2.4583144369447836,
|
| 678 |
+
"rewards/accuracies": 0.85,
|
| 679 |
+
"rewards/margins": 1.3669773700647057,
|
| 680 |
+
"logps/chosen": -75.5698697090149,
|
| 681 |
+
"logps/rejected": -72.76329650878907,
|
| 682 |
+
"epoch": 0.576,
|
| 683 |
+
"step": 360
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"loss": 0.46155486106872556,
|
| 687 |
+
"grad_norm": 0.19920872151851654,
|
| 688 |
+
"learning_rate": 7.048e-06,
|
| 689 |
+
"entropy": 0.8591178697883152,
|
| 690 |
+
"num_tokens": 4968874.0,
|
| 691 |
+
"logits/chosen": -1.3770929342653395,
|
| 692 |
+
"logits/rejected": -1.1350725351971858,
|
| 693 |
+
"mean_token_accuracy": 0.7090006841346621,
|
| 694 |
+
"rewards/chosen": -1.027018013925408,
|
| 695 |
+
"rewards/rejected": -2.083010801524506,
|
| 696 |
+
"rewards/accuracies": 0.79375,
|
| 697 |
+
"rewards/margins": 1.0559927898459136,
|
| 698 |
+
"logps/chosen": -63.01605747938156,
|
| 699 |
+
"logps/rejected": -76.09901772737503,
|
| 700 |
+
"epoch": 0.592,
|
| 701 |
+
"step": 370
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"loss": 0.4007772922515869,
|
| 705 |
+
"grad_norm": 0.126731276512146,
|
| 706 |
+
"learning_rate": 6.968e-06,
|
| 707 |
+
"entropy": 0.7402204329788219,
|
| 708 |
+
"num_tokens": 5102712.0,
|
| 709 |
+
"logits/chosen": -1.5890271734186694,
|
| 710 |
+
"logits/rejected": -1.2882636864367343,
|
| 711 |
+
"mean_token_accuracy": 0.6976638314314186,
|
| 712 |
+
"rewards/chosen": -0.9160357586602913,
|
| 713 |
+
"rewards/rejected": -2.2116100109182297,
|
| 714 |
+
"rewards/accuracies": 0.81875,
|
| 715 |
+
"rewards/margins": 1.2955742529593408,
|
| 716 |
+
"logps/chosen": -59.917374193668365,
|
| 717 |
+
"logps/rejected": -72.02031389474868,
|
| 718 |
+
"epoch": 0.608,
|
| 719 |
+
"step": 380
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"loss": 0.5412398338317871,
|
| 723 |
+
"grad_norm": 0.2466248720884323,
|
| 724 |
+
"learning_rate": 6.888e-06,
|
| 725 |
+
"entropy": 0.7976110359537415,
|
| 726 |
+
"num_tokens": 5236746.0,
|
| 727 |
+
"logits/chosen": -1.4715698894187053,
|
| 728 |
+
"logits/rejected": -1.2727028168607453,
|
| 729 |
+
"mean_token_accuracy": 0.7017135294154286,
|
| 730 |
+
"rewards/chosen": -1.1942087264207657,
|
| 731 |
+
"rewards/rejected": -2.24681163968271,
|
| 732 |
+
"rewards/accuracies": 0.75625,
|
| 733 |
+
"rewards/margins": 1.0526029143482447,
|
| 734 |
+
"logps/chosen": -67.42821979522705,
|
| 735 |
+
"logps/rejected": -71.59138869047165,
|
| 736 |
+
"epoch": 0.624,
|
| 737 |
+
"step": 390
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"loss": 0.4620323181152344,
|
| 741 |
+
"grad_norm": 0.19288523495197296,
|
| 742 |
+
"learning_rate": 6.808e-06,
|
| 743 |
+
"entropy": 0.8328921512700618,
|
| 744 |
+
"num_tokens": 5371265.0,
|
| 745 |
+
"logits/chosen": -1.3978239761455913,
|
| 746 |
+
"logits/rejected": -1.0700948427929158,
|
| 747 |
+
"mean_token_accuracy": 0.709115880727768,
|
| 748 |
+
"rewards/chosen": -0.9376408788230037,
|
| 749 |
+
"rewards/rejected": -2.036768500949256,
|
| 750 |
+
"rewards/accuracies": 0.79375,
|
| 751 |
+
"rewards/margins": 1.0991276198532431,
|
| 752 |
+
"logps/chosen": -63.57323570251465,
|
| 753 |
+
"logps/rejected": -70.56169172525406,
|
| 754 |
+
"epoch": 0.64,
|
| 755 |
+
"step": 400
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"loss": 0.37686054706573485,
|
| 759 |
+
"grad_norm": 0.22674216330051422,
|
| 760 |
+
"learning_rate": 6.728e-06,
|
| 761 |
+
"entropy": 0.810622811812209,
|
| 762 |
+
"num_tokens": 5505311.0,
|
| 763 |
+
"logits/chosen": -1.3580103877547358,
|
| 764 |
+
"logits/rejected": -1.1514336147547608,
|
| 765 |
+
"mean_token_accuracy": 0.7109507186338305,
|
| 766 |
+
"rewards/chosen": -0.5899793192191283,
|
| 767 |
+
"rewards/rejected": -1.8998081584577449,
|
| 768 |
+
"rewards/accuracies": 0.875,
|
| 769 |
+
"rewards/margins": 1.3098288293927909,
|
| 770 |
+
"logps/chosen": -62.02449452877045,
|
| 771 |
+
"logps/rejected": -69.4373902440071,
|
| 772 |
+
"epoch": 0.656,
|
| 773 |
+
"step": 410
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"loss": 0.4350727558135986,
|
| 777 |
+
"grad_norm": 0.16856394708156586,
|
| 778 |
+
"learning_rate": 6.648e-06,
|
| 779 |
+
"entropy": 0.8629972039605491,
|
| 780 |
+
"num_tokens": 5639727.0,
|
| 781 |
+
"logits/chosen": -1.296001624679018,
|
| 782 |
+
"logits/rejected": -1.0425315305055807,
|
| 783 |
+
"mean_token_accuracy": 0.6885800404474139,
|
| 784 |
+
"rewards/chosen": -0.7213470239425078,
|
| 785 |
+
"rewards/rejected": -1.8970674871001392,
|
| 786 |
+
"rewards/accuracies": 0.79375,
|
| 787 |
+
"rewards/margins": 1.1757204541005195,
|
| 788 |
+
"logps/chosen": -63.914617967605594,
|
| 789 |
+
"logps/rejected": -72.86851428747177,
|
| 790 |
+
"epoch": 0.672,
|
| 791 |
+
"step": 420
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"loss": 0.4224717140197754,
|
| 795 |
+
"grad_norm": 0.1479523628950119,
|
| 796 |
+
"learning_rate": 6.568000000000001e-06,
|
| 797 |
+
"entropy": 0.807172019200516,
|
| 798 |
+
"num_tokens": 5775347.0,
|
| 799 |
+
"logits/chosen": -1.5484924158169868,
|
| 800 |
+
"logits/rejected": -1.2025112233438988,
|
| 801 |
+
"mean_token_accuracy": 0.7156462514773011,
|
| 802 |
+
"rewards/chosen": -0.9871690425672568,
|
| 803 |
+
"rewards/rejected": -2.2504361010855063,
|
| 804 |
+
"rewards/accuracies": 0.8125,
|
| 805 |
+
"rewards/margins": 1.263267059624195,
|
| 806 |
+
"logps/chosen": -71.64003691673278,
|
| 807 |
+
"logps/rejected": -77.72584683895111,
|
| 808 |
+
"epoch": 0.688,
|
| 809 |
+
"step": 430
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"loss": 0.3901363849639893,
|
| 813 |
+
"grad_norm": 0.2484031617641449,
|
| 814 |
+
"learning_rate": 6.488000000000001e-06,
|
| 815 |
+
"entropy": 0.7301724591321544,
|
| 816 |
+
"num_tokens": 5908570.0,
|
| 817 |
+
"logits/chosen": -1.6231685453664735,
|
| 818 |
+
"logits/rejected": -1.2750791023731196,
|
| 819 |
+
"mean_token_accuracy": 0.7146034015342593,
|
| 820 |
+
"rewards/chosen": -1.1554993306839605,
|
| 821 |
+
"rewards/rejected": -2.4748502201691736,
|
| 822 |
+
"rewards/accuracies": 0.8375,
|
| 823 |
+
"rewards/margins": 1.3193508887663483,
|
| 824 |
+
"logps/chosen": -61.40522866249084,
|
| 825 |
+
"logps/rejected": -72.07903642654419,
|
| 826 |
+
"epoch": 0.704,
|
| 827 |
+
"step": 440
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"loss": 0.44475975036621096,
|
| 831 |
+
"grad_norm": 0.14690765738487244,
|
| 832 |
+
"learning_rate": 6.408000000000001e-06,
|
| 833 |
+
"entropy": 0.8033796272822655,
|
| 834 |
+
"num_tokens": 6042989.0,
|
| 835 |
+
"logits/chosen": -1.5474439165267901,
|
| 836 |
+
"logits/rejected": -1.2024063732889139,
|
| 837 |
+
"mean_token_accuracy": 0.7229112152010202,
|
| 838 |
+
"rewards/chosen": -1.2543478403938935,
|
| 839 |
+
"rewards/rejected": -2.721182713645976,
|
| 840 |
+
"rewards/accuracies": 0.8,
|
| 841 |
+
"rewards/margins": 1.4668348756618799,
|
| 842 |
+
"logps/chosen": -64.07742089033127,
|
| 843 |
+
"logps/rejected": -82.54657541513443,
|
| 844 |
+
"epoch": 0.72,
|
| 845 |
+
"step": 450
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"loss": 0.3836177110671997,
|
| 849 |
+
"grad_norm": 0.15559780597686768,
|
| 850 |
+
"learning_rate": 6.328000000000001e-06,
|
| 851 |
+
"entropy": 0.7930325305147562,
|
| 852 |
+
"num_tokens": 6177897.0,
|
| 853 |
+
"logits/chosen": -1.4986191571534253,
|
| 854 |
+
"logits/rejected": -1.2517336747528183,
|
| 855 |
+
"mean_token_accuracy": 0.7092798817902803,
|
| 856 |
+
"rewards/chosen": -1.14846060044365,
|
| 857 |
+
"rewards/rejected": -2.6847293067490683,
|
| 858 |
+
"rewards/accuracies": 0.86875,
|
| 859 |
+
"rewards/margins": 1.5362687051296233,
|
| 860 |
+
"logps/chosen": -67.3670264840126,
|
| 861 |
+
"logps/rejected": -78.67413110733033,
|
| 862 |
+
"epoch": 0.736,
|
| 863 |
+
"step": 460
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"loss": 0.40950717926025393,
|
| 867 |
+
"grad_norm": 0.1694619208574295,
|
| 868 |
+
"learning_rate": 6.248000000000001e-06,
|
| 869 |
+
"entropy": 0.8649296773452079,
|
| 870 |
+
"num_tokens": 6312714.0,
|
| 871 |
+
"logits/chosen": -1.2895246077624505,
|
| 872 |
+
"logits/rejected": -1.0161292065638676,
|
| 873 |
+
"mean_token_accuracy": 0.6956680539995432,
|
| 874 |
+
"rewards/chosen": -0.9786080905782001,
|
| 875 |
+
"rewards/rejected": -2.165545484540053,
|
| 876 |
+
"rewards/accuracies": 0.79375,
|
| 877 |
+
"rewards/margins": 1.1869373918976636,
|
| 878 |
+
"logps/chosen": -71.72884435653687,
|
| 879 |
+
"logps/rejected": -73.86714477539063,
|
| 880 |
+
"epoch": 0.752,
|
| 881 |
+
"step": 470
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"loss": 0.38074028491973877,
|
| 885 |
+
"grad_norm": 0.16199514269828796,
|
| 886 |
+
"learning_rate": 6.168000000000001e-06,
|
| 887 |
+
"entropy": 0.7845575341256336,
|
| 888 |
+
"num_tokens": 6446811.0,
|
| 889 |
+
"logits/chosen": -1.5487570258189929,
|
| 890 |
+
"logits/rejected": -1.3019611828115045,
|
| 891 |
+
"mean_token_accuracy": 0.7036560578271747,
|
| 892 |
+
"rewards/chosen": -1.0273570286633913,
|
| 893 |
+
"rewards/rejected": -2.410348759838962,
|
| 894 |
+
"rewards/accuracies": 0.81875,
|
| 895 |
+
"rewards/margins": 1.382991727720946,
|
| 896 |
+
"logps/chosen": -61.84894493818283,
|
| 897 |
+
"logps/rejected": -74.51683121919632,
|
| 898 |
+
"epoch": 0.768,
|
| 899 |
+
"step": 480
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"loss": 0.44433164596557617,
|
| 903 |
+
"grad_norm": 0.2861591875553131,
|
| 904 |
+
"learning_rate": 6.088000000000001e-06,
|
| 905 |
+
"entropy": 0.8558074278931599,
|
| 906 |
+
"num_tokens": 6581587.0,
|
| 907 |
+
"logits/chosen": -1.474127254445022,
|
| 908 |
+
"logits/rejected": -1.3042647255554116,
|
| 909 |
+
"mean_token_accuracy": 0.6997112531214953,
|
| 910 |
+
"rewards/chosen": -1.139366758624692,
|
| 911 |
+
"rewards/rejected": -2.3755088145378975,
|
| 912 |
+
"rewards/accuracies": 0.8,
|
| 913 |
+
"rewards/margins": 1.236142060905695,
|
| 914 |
+
"logps/chosen": -66.69998366832733,
|
| 915 |
+
"logps/rejected": -77.3253685593605,
|
| 916 |
+
"epoch": 0.784,
|
| 917 |
+
"step": 490
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"loss": 0.3971534729003906,
|
| 921 |
+
"grad_norm": 0.21812406182289124,
|
| 922 |
+
"learning_rate": 6.008000000000001e-06,
|
| 923 |
+
"entropy": 0.8011515848615091,
|
| 924 |
+
"num_tokens": 6716289.0,
|
| 925 |
+
"logits/chosen": -1.5570773973414094,
|
| 926 |
+
"logits/rejected": -1.1872032439366709,
|
| 927 |
+
"mean_token_accuracy": 0.7197606116533279,
|
| 928 |
+
"rewards/chosen": -1.1305481396848336,
|
| 929 |
+
"rewards/rejected": -2.501478673432575,
|
| 930 |
+
"rewards/accuracies": 0.825,
|
| 931 |
+
"rewards/margins": 1.3709305252879858,
|
| 932 |
+
"logps/chosen": -65.16791021823883,
|
| 933 |
+
"logps/rejected": -78.89157961606979,
|
| 934 |
+
"epoch": 0.8,
|
| 935 |
+
"step": 500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"loss": 0.37973556518554685,
|
| 939 |
+
"grad_norm": 0.16790038347244263,
|
| 940 |
+
"learning_rate": 5.928000000000001e-06,
|
| 941 |
+
"entropy": 0.8486743798508541,
|
| 942 |
+
"num_tokens": 6850716.0,
|
| 943 |
+
"logits/chosen": -1.647078239252084,
|
| 944 |
+
"logits/rejected": -1.2117462973476156,
|
| 945 |
+
"mean_token_accuracy": 0.6988467565737665,
|
| 946 |
+
"rewards/chosen": -0.9271791904306156,
|
| 947 |
+
"rewards/rejected": -2.5509458231390454,
|
| 948 |
+
"rewards/accuracies": 0.88125,
|
| 949 |
+
"rewards/margins": 1.6237666252069176,
|
| 950 |
+
"logps/chosen": -59.44525392055512,
|
| 951 |
+
"logps/rejected": -82.1239993095398,
|
| 952 |
+
"epoch": 0.816,
|
| 953 |
+
"step": 510
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"loss": 0.3395850658416748,
|
| 957 |
+
"grad_norm": 0.18018585443496704,
|
| 958 |
+
"learning_rate": 5.848000000000001e-06,
|
| 959 |
+
"entropy": 0.8265948085230775,
|
| 960 |
+
"num_tokens": 6984942.0,
|
| 961 |
+
"logits/chosen": -1.5910515463457346,
|
| 962 |
+
"logits/rejected": -1.128224583204076,
|
| 963 |
+
"mean_token_accuracy": 0.7097872287034989,
|
| 964 |
+
"rewards/chosen": -0.9600080527772661,
|
| 965 |
+
"rewards/rejected": -2.5357889306993457,
|
| 966 |
+
"rewards/accuracies": 0.86875,
|
| 967 |
+
"rewards/margins": 1.5757808735594154,
|
| 968 |
+
"logps/chosen": -64.7915377497673,
|
| 969 |
+
"logps/rejected": -79.26469901800155,
|
| 970 |
+
"epoch": 0.832,
|
| 971 |
+
"step": 520
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"loss": 0.35488090515136717,
|
| 975 |
+
"grad_norm": 0.2233273833990097,
|
| 976 |
+
"learning_rate": 5.7680000000000005e-06,
|
| 977 |
+
"entropy": 0.8343491271720268,
|
| 978 |
+
"num_tokens": 7118390.0,
|
| 979 |
+
"logits/chosen": -1.6220010549732222,
|
| 980 |
+
"logits/rejected": -1.2938580946166651,
|
| 981 |
+
"mean_token_accuracy": 0.7028924554586411,
|
| 982 |
+
"rewards/chosen": -1.0165848374832422,
|
| 983 |
+
"rewards/rejected": -2.6679536976735108,
|
| 984 |
+
"rewards/accuracies": 0.85625,
|
| 985 |
+
"rewards/margins": 1.6513688609353268,
|
| 986 |
+
"logps/chosen": -62.88960087299347,
|
| 987 |
+
"logps/rejected": -75.79626542329788,
|
| 988 |
+
"epoch": 0.848,
|
| 989 |
+
"step": 530
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"loss": 0.4319893836975098,
|
| 993 |
+
"grad_norm": 0.11507096141576767,
|
| 994 |
+
"learning_rate": 5.6880000000000004e-06,
|
| 995 |
+
"entropy": 0.8100111907624523,
|
| 996 |
+
"num_tokens": 7252607.0,
|
| 997 |
+
"logits/chosen": -1.690550336613653,
|
| 998 |
+
"logits/rejected": -1.296816083476357,
|
| 999 |
+
"mean_token_accuracy": 0.7133562916889786,
|
| 1000 |
+
"rewards/chosen": -1.126295023932471,
|
| 1001 |
+
"rewards/rejected": -2.788479746971279,
|
| 1002 |
+
"rewards/accuracies": 0.81875,
|
| 1003 |
+
"rewards/margins": 1.6621847297996282,
|
| 1004 |
+
"logps/chosen": -61.42004044055939,
|
| 1005 |
+
"logps/rejected": -82.76795139312745,
|
| 1006 |
+
"epoch": 0.864,
|
| 1007 |
+
"step": 540
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"loss": 0.40495963096618653,
|
| 1011 |
+
"grad_norm": 0.17571839690208435,
|
| 1012 |
+
"learning_rate": 5.608e-06,
|
| 1013 |
+
"entropy": 0.7233784958618343,
|
| 1014 |
+
"num_tokens": 7386208.0,
|
| 1015 |
+
"logits/chosen": -1.8208090889419204,
|
| 1016 |
+
"logits/rejected": -1.542335199851309,
|
| 1017 |
+
"mean_token_accuracy": 0.6927881988696754,
|
| 1018 |
+
"rewards/chosen": -1.2034590220835526,
|
| 1019 |
+
"rewards/rejected": -2.8114285852760075,
|
| 1020 |
+
"rewards/accuracies": 0.8375,
|
| 1021 |
+
"rewards/margins": 1.607969574816525,
|
| 1022 |
+
"logps/chosen": -60.88411279916763,
|
| 1023 |
+
"logps/rejected": -79.19290779829025,
|
| 1024 |
+
"epoch": 0.88,
|
| 1025 |
+
"step": 550
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"loss": 0.35812089443206785,
|
| 1029 |
+
"grad_norm": 0.17074806988239288,
|
| 1030 |
+
"learning_rate": 5.528e-06,
|
| 1031 |
+
"entropy": 0.8315804405428935,
|
| 1032 |
+
"num_tokens": 7521519.0,
|
| 1033 |
+
"logits/chosen": -1.7844075335921168,
|
| 1034 |
+
"logits/rejected": -1.4025306371099586,
|
| 1035 |
+
"mean_token_accuracy": 0.7045917056500912,
|
| 1036 |
+
"rewards/chosen": -1.1598298101627733,
|
| 1037 |
+
"rewards/rejected": -3.0565326128853485,
|
| 1038 |
+
"rewards/accuracies": 0.85625,
|
| 1039 |
+
"rewards/margins": 1.8967028062790632,
|
| 1040 |
+
"logps/chosen": -66.83561842441559,
|
| 1041 |
+
"logps/rejected": -92.35079349279404,
|
| 1042 |
+
"epoch": 0.896,
|
| 1043 |
+
"step": 560
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"loss": 0.4332554817199707,
|
| 1047 |
+
"grad_norm": 0.2333141267299652,
|
| 1048 |
+
"learning_rate": 5.448e-06,
|
| 1049 |
+
"entropy": 0.8866394167125691,
|
| 1050 |
+
"num_tokens": 7656864.0,
|
| 1051 |
+
"logits/chosen": -1.6045995118119563,
|
| 1052 |
+
"logits/rejected": -1.3098578342414633,
|
| 1053 |
+
"mean_token_accuracy": 0.6903222082182765,
|
| 1054 |
+
"rewards/chosen": -1.2476856299676,
|
| 1055 |
+
"rewards/rejected": -2.8476344041526316,
|
| 1056 |
+
"rewards/accuracies": 0.7875,
|
| 1057 |
+
"rewards/margins": 1.5999487713910638,
|
| 1058 |
+
"logps/chosen": -73.24417120218277,
|
| 1059 |
+
"logps/rejected": -85.07146887779236,
|
| 1060 |
+
"epoch": 0.912,
|
| 1061 |
+
"step": 570
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"loss": 0.3921473503112793,
|
| 1065 |
+
"grad_norm": 0.08647091686725616,
|
| 1066 |
+
"learning_rate": 5.368000000000001e-06,
|
| 1067 |
+
"entropy": 0.8643649270874448,
|
| 1068 |
+
"num_tokens": 7792899.0,
|
| 1069 |
+
"logits/chosen": -1.4821103055613285,
|
| 1070 |
+
"logits/rejected": -1.2834978774386792,
|
| 1071 |
+
"mean_token_accuracy": 0.679491494782269,
|
| 1072 |
+
"rewards/chosen": -1.2919643745910434,
|
| 1073 |
+
"rewards/rejected": -2.864986840263009,
|
| 1074 |
+
"rewards/accuracies": 0.84375,
|
| 1075 |
+
"rewards/margins": 1.5730224633589387,
|
| 1076 |
+
"logps/chosen": -79.60574653148652,
|
| 1077 |
+
"logps/rejected": -82.35489063262939,
|
| 1078 |
+
"epoch": 0.928,
|
| 1079 |
+
"step": 580
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"loss": 0.35202584266662595,
|
| 1083 |
+
"grad_norm": 0.1588275283575058,
|
| 1084 |
+
"learning_rate": 5.288000000000001e-06,
|
| 1085 |
+
"entropy": 0.8073152711498551,
|
| 1086 |
+
"num_tokens": 7926779.0,
|
| 1087 |
+
"logits/chosen": -1.7572898405260537,
|
| 1088 |
+
"logits/rejected": -1.3840750039245335,
|
| 1089 |
+
"mean_token_accuracy": 0.7003073239699006,
|
| 1090 |
+
"rewards/chosen": -0.9911620355036576,
|
| 1091 |
+
"rewards/rejected": -2.889724260215007,
|
| 1092 |
+
"rewards/accuracies": 0.85,
|
| 1093 |
+
"rewards/margins": 1.8985622256062924,
|
| 1094 |
+
"logps/chosen": -63.91571959257126,
|
| 1095 |
+
"logps/rejected": -79.23539482355118,
|
| 1096 |
+
"epoch": 0.944,
|
| 1097 |
+
"step": 590
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"loss": 0.4021899700164795,
|
| 1101 |
+
"grad_norm": 0.2877083122730255,
|
| 1102 |
+
"learning_rate": 5.208000000000001e-06,
|
| 1103 |
+
"entropy": 0.7111593881221779,
|
| 1104 |
+
"num_tokens": 8060079.0,
|
| 1105 |
+
"logits/chosen": -1.8737773265139275,
|
| 1106 |
+
"logits/rejected": -1.5103133544722522,
|
| 1107 |
+
"mean_token_accuracy": 0.7314831869676709,
|
| 1108 |
+
"rewards/chosen": -1.071392911881412,
|
| 1109 |
+
"rewards/rejected": -2.6491885555908086,
|
| 1110 |
+
"rewards/accuracies": 0.875,
|
| 1111 |
+
"rewards/margins": 1.5777956407517195,
|
| 1112 |
+
"logps/chosen": -56.149746000766754,
|
| 1113 |
+
"logps/rejected": -73.58239371776581,
|
| 1114 |
+
"epoch": 0.96,
|
| 1115 |
+
"step": 600
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"loss": 0.4523749828338623,
|
| 1119 |
+
"grad_norm": 0.20730054378509521,
|
| 1120 |
+
"learning_rate": 5.128000000000001e-06,
|
| 1121 |
+
"entropy": 0.8179327869209374,
|
| 1122 |
+
"num_tokens": 8194850.0,
|
| 1123 |
+
"logits/chosen": -1.557446959658919,
|
| 1124 |
+
"logits/rejected": -1.2361516245111352,
|
| 1125 |
+
"mean_token_accuracy": 0.6996246555820107,
|
| 1126 |
+
"rewards/chosen": -0.9787147940834984,
|
| 1127 |
+
"rewards/rejected": -2.3619479497894647,
|
| 1128 |
+
"rewards/accuracies": 0.81875,
|
| 1129 |
+
"rewards/margins": 1.3832331511192024,
|
| 1130 |
+
"logps/chosen": -61.125505900383,
|
| 1131 |
+
"logps/rejected": -76.67065613269806,
|
| 1132 |
+
"epoch": 0.976,
|
| 1133 |
+
"step": 610
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"loss": 0.40350823402404784,
|
| 1137 |
+
"grad_norm": 0.2933248281478882,
|
| 1138 |
+
"learning_rate": 5.048000000000001e-06,
|
| 1139 |
+
"entropy": 0.8629961145110429,
|
| 1140 |
+
"num_tokens": 8329816.0,
|
| 1141 |
+
"logits/chosen": -1.6289873809292001,
|
| 1142 |
+
"logits/rejected": -1.344169554894789,
|
| 1143 |
+
"mean_token_accuracy": 0.6997346790507436,
|
| 1144 |
+
"rewards/chosen": -0.9943598300509621,
|
| 1145 |
+
"rewards/rejected": -2.568173354433384,
|
| 1146 |
+
"rewards/accuracies": 0.8125,
|
| 1147 |
+
"rewards/margins": 1.5738135130144655,
|
| 1148 |
+
"logps/chosen": -65.94150141477584,
|
| 1149 |
+
"logps/rejected": -81.87862153053284,
|
| 1150 |
+
"epoch": 0.992,
|
| 1151 |
+
"step": 620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"loss": 0.3900306224822998,
|
| 1155 |
+
"grad_norm": 0.23495233058929443,
|
| 1156 |
+
"learning_rate": 4.9680000000000005e-06,
|
| 1157 |
+
"entropy": 0.8978026660814067,
|
| 1158 |
+
"num_tokens": 8464146.0,
|
| 1159 |
+
"logits/chosen": -1.4666583790779915,
|
| 1160 |
+
"logits/rejected": -1.2711647147767078,
|
| 1161 |
+
"mean_token_accuracy": 0.681163523439318,
|
| 1162 |
+
"rewards/chosen": -0.927846639166819,
|
| 1163 |
+
"rewards/rejected": -2.4021925135632047,
|
| 1164 |
+
"rewards/accuracies": 0.81875,
|
| 1165 |
+
"rewards/margins": 1.4743458745069802,
|
| 1166 |
+
"logps/chosen": -65.02851099967957,
|
| 1167 |
+
"logps/rejected": -75.92052472829819,
|
| 1168 |
+
"epoch": 1.008,
|
| 1169 |
+
"step": 630
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"loss": 0.36000044345855714,
|
| 1173 |
+
"grad_norm": 0.20741842687129974,
|
| 1174 |
+
"learning_rate": 4.8880000000000005e-06,
|
| 1175 |
+
"entropy": 0.83651371775195,
|
| 1176 |
+
"num_tokens": 8598165.0,
|
| 1177 |
+
"logits/chosen": -1.75481545405402,
|
| 1178 |
+
"logits/rejected": -1.2945304508994626,
|
| 1179 |
+
"mean_token_accuracy": 0.696090385876596,
|
| 1180 |
+
"rewards/chosen": -1.0052818381387625,
|
| 1181 |
+
"rewards/rejected": -2.635122027853504,
|
| 1182 |
+
"rewards/accuracies": 0.86875,
|
| 1183 |
+
"rewards/margins": 1.6298401849344373,
|
| 1184 |
+
"logps/chosen": -61.1220197558403,
|
| 1185 |
+
"logps/rejected": -80.07098776102066,
|
| 1186 |
+
"epoch": 1.024,
|
| 1187 |
+
"step": 640
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"loss": 0.3603890657424927,
|
| 1191 |
+
"grad_norm": 0.13139435648918152,
|
| 1192 |
+
"learning_rate": 4.808e-06,
|
| 1193 |
+
"entropy": 0.8451052400683693,
|
| 1194 |
+
"num_tokens": 8733182.0,
|
| 1195 |
+
"logits/chosen": -1.4909471858334125,
|
| 1196 |
+
"logits/rejected": -1.169245322622007,
|
| 1197 |
+
"mean_token_accuracy": 0.711542940698564,
|
| 1198 |
+
"rewards/chosen": -1.1369427774789074,
|
| 1199 |
+
"rewards/rejected": -2.7133941426873207,
|
| 1200 |
+
"rewards/accuracies": 0.8875,
|
| 1201 |
+
"rewards/margins": 1.5764513722620905,
|
| 1202 |
+
"logps/chosen": -68.56432271003723,
|
| 1203 |
+
"logps/rejected": -82.01030631065369,
|
| 1204 |
+
"epoch": 1.04,
|
| 1205 |
+
"step": 650
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"loss": 0.3818248748779297,
|
| 1209 |
+
"grad_norm": 0.2262716144323349,
|
| 1210 |
+
"learning_rate": 4.728e-06,
|
| 1211 |
+
"entropy": 0.8377733555389568,
|
| 1212 |
+
"num_tokens": 8866532.0,
|
| 1213 |
+
"logits/chosen": -1.5367347222131402,
|
| 1214 |
+
"logits/rejected": -1.099738407975035,
|
| 1215 |
+
"mean_token_accuracy": 0.7022423831745982,
|
| 1216 |
+
"rewards/chosen": -1.2749694007623475,
|
| 1217 |
+
"rewards/rejected": -3.0242153271101415,
|
| 1218 |
+
"rewards/accuracies": 0.85625,
|
| 1219 |
+
"rewards/margins": 1.7492459254339336,
|
| 1220 |
+
"logps/chosen": -63.86832449436188,
|
| 1221 |
+
"logps/rejected": -83.10683650970459,
|
| 1222 |
+
"epoch": 1.056,
|
| 1223 |
+
"step": 660
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"loss": 0.4226510047912598,
|
| 1227 |
+
"grad_norm": 0.14209647476673126,
|
| 1228 |
+
"learning_rate": 4.648e-06,
|
| 1229 |
+
"entropy": 0.8949731356144184,
|
| 1230 |
+
"num_tokens": 9002058.0,
|
| 1231 |
+
"logits/chosen": -1.3276259144604574,
|
| 1232 |
+
"logits/rejected": -1.0882318472209296,
|
| 1233 |
+
"mean_token_accuracy": 0.6937677466310561,
|
| 1234 |
+
"rewards/chosen": -1.3226296437671408,
|
| 1235 |
+
"rewards/rejected": -2.8028671694453804,
|
| 1236 |
+
"rewards/accuracies": 0.81875,
|
| 1237 |
+
"rewards/margins": 1.4802375258877873,
|
| 1238 |
+
"logps/chosen": -76.20145353078843,
|
| 1239 |
+
"logps/rejected": -83.99533534049988,
|
| 1240 |
+
"epoch": 1.072,
|
| 1241 |
+
"step": 670
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"loss": 0.37326576709747317,
|
| 1245 |
+
"grad_norm": 0.15859447419643402,
|
| 1246 |
+
"learning_rate": 4.568e-06,
|
| 1247 |
+
"entropy": 0.849602137284819,
|
| 1248 |
+
"num_tokens": 9136559.0,
|
| 1249 |
+
"logits/chosen": -1.5246000942091014,
|
| 1250 |
+
"logits/rejected": -1.1591778776100774,
|
| 1251 |
+
"mean_token_accuracy": 0.698604229837656,
|
| 1252 |
+
"rewards/chosen": -1.0672220802895027,
|
| 1253 |
+
"rewards/rejected": -2.744057681580307,
|
| 1254 |
+
"rewards/accuracies": 0.81875,
|
| 1255 |
+
"rewards/margins": 1.6768356027081608,
|
| 1256 |
+
"logps/chosen": -68.01634130477905,
|
| 1257 |
+
"logps/rejected": -79.27202568054199,
|
| 1258 |
+
"epoch": 1.088,
|
| 1259 |
+
"step": 680
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"loss": 0.37404346466064453,
|
| 1263 |
+
"grad_norm": 0.24664947390556335,
|
| 1264 |
+
"learning_rate": 4.488e-06,
|
| 1265 |
+
"entropy": 0.7337783641763963,
|
| 1266 |
+
"num_tokens": 9269437.0,
|
| 1267 |
+
"logits/chosen": -1.7975371256399726,
|
| 1268 |
+
"logits/rejected": -1.4473784742934162,
|
| 1269 |
+
"mean_token_accuracy": 0.7255065334960819,
|
| 1270 |
+
"rewards/chosen": -1.1141222154546995,
|
| 1271 |
+
"rewards/rejected": -2.7036397748812306,
|
| 1272 |
+
"rewards/accuracies": 0.8625,
|
| 1273 |
+
"rewards/margins": 1.5895175519399345,
|
| 1274 |
+
"logps/chosen": -59.54515955448151,
|
| 1275 |
+
"logps/rejected": -72.48874368667603,
|
| 1276 |
+
"epoch": 1.104,
|
| 1277 |
+
"step": 690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"loss": 0.32263295650482177,
|
| 1281 |
+
"grad_norm": 0.11260554194450378,
|
| 1282 |
+
"learning_rate": 4.408000000000001e-06,
|
| 1283 |
+
"entropy": 0.896336995443562,
|
| 1284 |
+
"num_tokens": 9404559.0,
|
| 1285 |
+
"logits/chosen": -1.393970769973714,
|
| 1286 |
+
"logits/rejected": -1.1303411869798095,
|
| 1287 |
+
"mean_token_accuracy": 0.6889882240444422,
|
| 1288 |
+
"rewards/chosen": -0.8930940314894542,
|
| 1289 |
+
"rewards/rejected": -2.7461632952094077,
|
| 1290 |
+
"rewards/accuracies": 0.86875,
|
| 1291 |
+
"rewards/margins": 1.8530692683532834,
|
| 1292 |
+
"logps/chosen": -68.77362660169601,
|
| 1293 |
+
"logps/rejected": -83.58723495006561,
|
| 1294 |
+
"epoch": 1.12,
|
| 1295 |
+
"step": 700
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"loss": 0.3827756643295288,
|
| 1299 |
+
"grad_norm": 0.23789574205875397,
|
| 1300 |
+
"learning_rate": 4.328000000000001e-06,
|
| 1301 |
+
"entropy": 0.8888725781347603,
|
| 1302 |
+
"num_tokens": 9539800.0,
|
| 1303 |
+
"logits/chosen": -1.3627809429008482,
|
| 1304 |
+
"logits/rejected": -0.9781940445978725,
|
| 1305 |
+
"mean_token_accuracy": 0.6972703387029469,
|
| 1306 |
+
"rewards/chosen": -0.8857813907903619,
|
| 1307 |
+
"rewards/rejected": -2.463674101768993,
|
| 1308 |
+
"rewards/accuracies": 0.86875,
|
| 1309 |
+
"rewards/margins": 1.577892707195133,
|
| 1310 |
+
"logps/chosen": -63.77985138893128,
|
| 1311 |
+
"logps/rejected": -79.29511383771896,
|
| 1312 |
+
"epoch": 1.1360000000000001,
|
| 1313 |
+
"step": 710
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"loss": 0.4144582748413086,
|
| 1317 |
+
"grad_norm": 0.163747638463974,
|
| 1318 |
+
"learning_rate": 4.248000000000001e-06,
|
| 1319 |
+
"entropy": 0.8231887853733497,
|
| 1320 |
+
"num_tokens": 9673582.0,
|
| 1321 |
+
"logits/chosen": -1.6634044705240583,
|
| 1322 |
+
"logits/rejected": -1.219379832809559,
|
| 1323 |
+
"mean_token_accuracy": 0.7126298141665757,
|
| 1324 |
+
"rewards/chosen": -0.919659434541245,
|
| 1325 |
+
"rewards/rejected": -2.3605332392355196,
|
| 1326 |
+
"rewards/accuracies": 0.7875,
|
| 1327 |
+
"rewards/margins": 1.440873800171539,
|
| 1328 |
+
"logps/chosen": -56.74032578468323,
|
| 1329 |
+
"logps/rejected": -75.13354351520539,
|
| 1330 |
+
"epoch": 1.152,
|
| 1331 |
+
"step": 720
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"loss": 0.36356205940246583,
|
| 1335 |
+
"grad_norm": 0.21325920522212982,
|
| 1336 |
+
"learning_rate": 4.168000000000001e-06,
|
| 1337 |
+
"entropy": 0.8396360178128817,
|
| 1338 |
+
"num_tokens": 9807723.0,
|
| 1339 |
+
"logits/chosen": -1.6067492277562274,
|
| 1340 |
+
"logits/rejected": -1.2729465928708659,
|
| 1341 |
+
"mean_token_accuracy": 0.7002692798152566,
|
| 1342 |
+
"rewards/chosen": -0.9061545075150207,
|
| 1343 |
+
"rewards/rejected": -2.6099032792961223,
|
| 1344 |
+
"rewards/accuracies": 0.8125,
|
| 1345 |
+
"rewards/margins": 1.7037487668916582,
|
| 1346 |
+
"logps/chosen": -58.46343905925751,
|
| 1347 |
+
"logps/rejected": -80.48531986474991,
|
| 1348 |
+
"epoch": 1.168,
|
| 1349 |
+
"step": 730
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"loss": 0.37539641857147216,
|
| 1353 |
+
"grad_norm": 0.20330095291137695,
|
| 1354 |
+
"learning_rate": 4.0880000000000005e-06,
|
| 1355 |
+
"entropy": 0.7463710659649223,
|
| 1356 |
+
"num_tokens": 9940521.0,
|
| 1357 |
+
"logits/chosen": -1.8096148756643575,
|
| 1358 |
+
"logits/rejected": -1.3828069653615978,
|
| 1359 |
+
"mean_token_accuracy": 0.7085271723568439,
|
| 1360 |
+
"rewards/chosen": -0.8202605500380742,
|
| 1361 |
+
"rewards/rejected": -2.4309751536580735,
|
| 1362 |
+
"rewards/accuracies": 0.825,
|
| 1363 |
+
"rewards/margins": 1.6107146009802817,
|
| 1364 |
+
"logps/chosen": -56.036488878726956,
|
| 1365 |
+
"logps/rejected": -69.77100689411164,
|
| 1366 |
+
"epoch": 1.184,
|
| 1367 |
+
"step": 740
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"loss": 0.4134565830230713,
|
| 1371 |
+
"grad_norm": 0.18556374311447144,
|
| 1372 |
+
"learning_rate": 4.008e-06,
|
| 1373 |
+
"entropy": 0.9190443507628515,
|
| 1374 |
+
"num_tokens": 10075349.0,
|
| 1375 |
+
"logits/chosen": -1.3740209851402565,
|
| 1376 |
+
"logits/rejected": -1.107910639583531,
|
| 1377 |
+
"mean_token_accuracy": 0.6791403977200389,
|
| 1378 |
+
"rewards/chosen": -1.0011425132019212,
|
| 1379 |
+
"rewards/rejected": -2.464534604479559,
|
| 1380 |
+
"rewards/accuracies": 0.825,
|
| 1381 |
+
"rewards/margins": 1.4633920882362872,
|
| 1382 |
+
"logps/chosen": -66.21626259088517,
|
| 1383 |
+
"logps/rejected": -77.55921986103058,
|
| 1384 |
+
"epoch": 1.2,
|
| 1385 |
+
"step": 750
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"loss": 0.3304694414138794,
|
| 1389 |
+
"grad_norm": 0.13154134154319763,
|
| 1390 |
+
"learning_rate": 3.928e-06,
|
| 1391 |
+
"entropy": 0.9318983974168077,
|
| 1392 |
+
"num_tokens": 10210309.0,
|
| 1393 |
+
"logits/chosen": -1.470085552602296,
|
| 1394 |
+
"logits/rejected": -1.1098111260703036,
|
| 1395 |
+
"mean_token_accuracy": 0.6740563122555614,
|
| 1396 |
+
"rewards/chosen": -0.9196080261957832,
|
| 1397 |
+
"rewards/rejected": -2.8224395443685353,
|
| 1398 |
+
"rewards/accuracies": 0.90625,
|
| 1399 |
+
"rewards/margins": 1.9028315196745098,
|
| 1400 |
+
"logps/chosen": -67.99445941448212,
|
| 1401 |
+
"logps/rejected": -84.53982878923416,
|
| 1402 |
+
"epoch": 1.216,
|
| 1403 |
+
"step": 760
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"loss": 0.365576171875,
|
| 1407 |
+
"grad_norm": 0.23938694596290588,
|
| 1408 |
+
"learning_rate": 3.848e-06,
|
| 1409 |
+
"entropy": 0.852998117939569,
|
| 1410 |
+
"num_tokens": 10344571.0,
|
| 1411 |
+
"logits/chosen": -1.6343794581170346,
|
| 1412 |
+
"logits/rejected": -1.1391983971011674,
|
| 1413 |
+
"mean_token_accuracy": 0.7156095307320356,
|
| 1414 |
+
"rewards/chosen": -1.0998709766048704,
|
| 1415 |
+
"rewards/rejected": -2.813695689383894,
|
| 1416 |
+
"rewards/accuracies": 0.84375,
|
| 1417 |
+
"rewards/margins": 1.71382470196113,
|
| 1418 |
+
"logps/chosen": -56.721822392940524,
|
| 1419 |
+
"logps/rejected": -86.82191110849381,
|
| 1420 |
+
"epoch": 1.232,
|
| 1421 |
+
"step": 770
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"loss": 0.37106683254241946,
|
| 1425 |
+
"grad_norm": 0.2882203459739685,
|
| 1426 |
+
"learning_rate": 3.7680000000000006e-06,
|
| 1427 |
+
"entropy": 0.7641319791146088,
|
| 1428 |
+
"num_tokens": 10478286.0,
|
| 1429 |
+
"logits/chosen": -1.9411937331572156,
|
| 1430 |
+
"logits/rejected": -1.4751655271730548,
|
| 1431 |
+
"mean_token_accuracy": 0.7104108802042901,
|
| 1432 |
+
"rewards/chosen": -1.266980020480696,
|
| 1433 |
+
"rewards/rejected": -3.165134116273839,
|
| 1434 |
+
"rewards/accuracies": 0.8375,
|
| 1435 |
+
"rewards/margins": 1.8981541025917976,
|
| 1436 |
+
"logps/chosen": -61.775586903095245,
|
| 1437 |
+
"logps/rejected": -82.27877287864685,
|
| 1438 |
+
"epoch": 1.248,
|
| 1439 |
+
"step": 780
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"loss": 0.39991624355316163,
|
| 1443 |
+
"grad_norm": 0.36986157298088074,
|
| 1444 |
+
"learning_rate": 3.6880000000000005e-06,
|
| 1445 |
+
"entropy": 0.8686222603835632,
|
| 1446 |
+
"num_tokens": 10612286.0,
|
| 1447 |
+
"logits/chosen": -1.6299915497340305,
|
| 1448 |
+
"logits/rejected": -1.2540868199616586,
|
| 1449 |
+
"mean_token_accuracy": 0.6993465088307858,
|
| 1450 |
+
"rewards/chosen": -1.324579149678175,
|
| 1451 |
+
"rewards/rejected": -3.0413625931367276,
|
| 1452 |
+
"rewards/accuracies": 0.81875,
|
| 1453 |
+
"rewards/margins": 1.7167834496125578,
|
| 1454 |
+
"logps/chosen": -66.68241719007491,
|
| 1455 |
+
"logps/rejected": -82.70180885791778,
|
| 1456 |
+
"epoch": 1.264,
|
| 1457 |
+
"step": 790
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"loss": 0.4275864601135254,
|
| 1461 |
+
"grad_norm": 0.15313459932804108,
|
| 1462 |
+
"learning_rate": 3.6080000000000004e-06,
|
| 1463 |
+
"entropy": 0.8286645252257585,
|
| 1464 |
+
"num_tokens": 10747462.0,
|
| 1465 |
+
"logits/chosen": -1.6437166513479284,
|
| 1466 |
+
"logits/rejected": -1.321237043633697,
|
| 1467 |
+
"mean_token_accuracy": 0.6946176150813699,
|
| 1468 |
+
"rewards/chosen": -1.1271761629614048,
|
| 1469 |
+
"rewards/rejected": -2.741164942184696,
|
| 1470 |
+
"rewards/accuracies": 0.79375,
|
| 1471 |
+
"rewards/margins": 1.6139887776225805,
|
| 1472 |
+
"logps/chosen": -70.66578447818756,
|
| 1473 |
+
"logps/rejected": -82.14218564033509,
|
| 1474 |
+
"epoch": 1.28,
|
| 1475 |
+
"step": 800
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"loss": 0.3684597730636597,
|
| 1479 |
+
"grad_norm": 0.13916034996509552,
|
| 1480 |
+
"learning_rate": 3.5280000000000004e-06,
|
| 1481 |
+
"entropy": 0.8530381735425181,
|
| 1482 |
+
"num_tokens": 10882594.0,
|
| 1483 |
+
"logits/chosen": -1.5982792033852526,
|
| 1484 |
+
"logits/rejected": -1.3680311549140272,
|
| 1485 |
+
"mean_token_accuracy": 0.6905860897153616,
|
| 1486 |
+
"rewards/chosen": -1.05685164804745,
|
| 1487 |
+
"rewards/rejected": -2.84542333453428,
|
| 1488 |
+
"rewards/accuracies": 0.8625,
|
| 1489 |
+
"rewards/margins": 1.7885716904886066,
|
| 1490 |
+
"logps/chosen": -69.83115639686585,
|
| 1491 |
+
"logps/rejected": -84.56852403879165,
|
| 1492 |
+
"epoch": 1.296,
|
| 1493 |
+
"step": 810
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"loss": 0.42574124336242675,
|
| 1497 |
+
"grad_norm": 0.2668042778968811,
|
| 1498 |
+
"learning_rate": 3.4480000000000003e-06,
|
| 1499 |
+
"entropy": 0.8458031455113086,
|
| 1500 |
+
"num_tokens": 11016931.0,
|
| 1501 |
+
"logits/chosen": -1.5238562132702138,
|
| 1502 |
+
"logits/rejected": -1.315186408558361,
|
| 1503 |
+
"mean_token_accuracy": 0.6900694493204356,
|
| 1504 |
+
"rewards/chosen": -1.2684904914320214,
|
| 1505 |
+
"rewards/rejected": -2.8179244715720415,
|
| 1506 |
+
"rewards/accuracies": 0.78125,
|
| 1507 |
+
"rewards/margins": 1.549433980602771,
|
| 1508 |
+
"logps/chosen": -72.86792405843735,
|
| 1509 |
+
"logps/rejected": -78.78973067998886,
|
| 1510 |
+
"epoch": 1.312,
|
| 1511 |
+
"step": 820
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"loss": 0.38849568367004395,
|
| 1515 |
+
"grad_norm": 0.21854253113269806,
|
| 1516 |
+
"learning_rate": 3.368e-06,
|
| 1517 |
+
"entropy": 0.7808890710002743,
|
| 1518 |
+
"num_tokens": 11150609.0,
|
| 1519 |
+
"logits/chosen": -1.7333184588359392,
|
| 1520 |
+
"logits/rejected": -1.3160499906125902,
|
| 1521 |
+
"mean_token_accuracy": 0.7088964313268662,
|
| 1522 |
+
"rewards/chosen": -1.051935706997756,
|
| 1523 |
+
"rewards/rejected": -2.7482495688600466,
|
| 1524 |
+
"rewards/accuracies": 0.84375,
|
| 1525 |
+
"rewards/margins": 1.696313863270916,
|
| 1526 |
+
"logps/chosen": -60.777068567276004,
|
| 1527 |
+
"logps/rejected": -78.34404106140137,
|
| 1528 |
+
"epoch": 1.328,
|
| 1529 |
+
"step": 830
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"loss": 0.2801020383834839,
|
| 1533 |
+
"grad_norm": 0.14788253605365753,
|
| 1534 |
+
"learning_rate": 3.288e-06,
|
| 1535 |
+
"entropy": 0.8135352808632887,
|
| 1536 |
+
"num_tokens": 11284845.0,
|
| 1537 |
+
"logits/chosen": -1.7096239866010419,
|
| 1538 |
+
"logits/rejected": -1.3006975907050757,
|
| 1539 |
+
"mean_token_accuracy": 0.7158975720405578,
|
| 1540 |
+
"rewards/chosen": -1.0535932141996454,
|
| 1541 |
+
"rewards/rejected": -3.0869329798792022,
|
| 1542 |
+
"rewards/accuracies": 0.8875,
|
| 1543 |
+
"rewards/margins": 2.0333397705107927,
|
| 1544 |
+
"logps/chosen": -59.187955629825595,
|
| 1545 |
+
"logps/rejected": -85.36171028614044,
|
| 1546 |
+
"epoch": 1.3439999999999999,
|
| 1547 |
+
"step": 840
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"loss": 0.34232614040374754,
|
| 1551 |
+
"grad_norm": 0.2614403963088989,
|
| 1552 |
+
"learning_rate": 3.208e-06,
|
| 1553 |
+
"entropy": 0.7222817947855219,
|
| 1554 |
+
"num_tokens": 11418114.0,
|
| 1555 |
+
"logits/chosen": -1.9495841034100554,
|
| 1556 |
+
"logits/rejected": -1.5302492821805944,
|
| 1557 |
+
"mean_token_accuracy": 0.7231298649683595,
|
| 1558 |
+
"rewards/chosen": -0.9936948938993737,
|
| 1559 |
+
"rewards/rejected": -2.912717835023068,
|
| 1560 |
+
"rewards/accuracies": 0.81875,
|
| 1561 |
+
"rewards/margins": 1.9190229372587055,
|
| 1562 |
+
"logps/chosen": -56.314193999767305,
|
| 1563 |
+
"logps/rejected": -76.84246274232865,
|
| 1564 |
+
"epoch": 1.3599999999999999,
|
| 1565 |
+
"step": 850
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"loss": 0.38190374374389646,
|
| 1569 |
+
"grad_norm": 0.17502538859844208,
|
| 1570 |
+
"learning_rate": 3.1280000000000004e-06,
|
| 1571 |
+
"entropy": 0.7970137125259498,
|
| 1572 |
+
"num_tokens": 11552146.0,
|
| 1573 |
+
"logits/chosen": -1.6720444613561587,
|
| 1574 |
+
"logits/rejected": -1.3887348601897287,
|
| 1575 |
+
"mean_token_accuracy": 0.7104262910783291,
|
| 1576 |
+
"rewards/chosen": -1.0953017962921876,
|
| 1577 |
+
"rewards/rejected": -2.871660438575782,
|
| 1578 |
+
"rewards/accuracies": 0.8375,
|
| 1579 |
+
"rewards/margins": 1.7763586389832198,
|
| 1580 |
+
"logps/chosen": -60.417977035045624,
|
| 1581 |
+
"logps/rejected": -80.406764960289,
|
| 1582 |
+
"epoch": 1.376,
|
| 1583 |
+
"step": 860
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"loss": 0.28182373046875,
|
| 1587 |
+
"grad_norm": 0.2861666679382324,
|
| 1588 |
+
"learning_rate": 3.0480000000000003e-06,
|
| 1589 |
+
"entropy": 0.8064172255253652,
|
| 1590 |
+
"num_tokens": 11686051.0,
|
| 1591 |
+
"logits/chosen": -1.6329583906103573,
|
| 1592 |
+
"logits/rejected": -1.3650252362412618,
|
| 1593 |
+
"mean_token_accuracy": 0.7067547043785453,
|
| 1594 |
+
"rewards/chosen": -1.0566333419817966,
|
| 1595 |
+
"rewards/rejected": -3.1455142412742134,
|
| 1596 |
+
"rewards/accuracies": 0.8875,
|
| 1597 |
+
"rewards/margins": 2.088880904391408,
|
| 1598 |
+
"logps/chosen": -64.32867197990417,
|
| 1599 |
+
"logps/rejected": -79.8797638297081,
|
| 1600 |
+
"epoch": 1.392,
|
| 1601 |
+
"step": 870
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"loss": 0.4359275817871094,
|
| 1605 |
+
"grad_norm": 0.39904382824897766,
|
| 1606 |
+
"learning_rate": 2.9680000000000002e-06,
|
| 1607 |
+
"entropy": 0.8518196220771642,
|
| 1608 |
+
"num_tokens": 11820419.0,
|
| 1609 |
+
"logits/chosen": -1.5591272336223294,
|
| 1610 |
+
"logits/rejected": -1.2953753406100752,
|
| 1611 |
+
"mean_token_accuracy": 0.6923748327419161,
|
| 1612 |
+
"rewards/chosen": -1.0741606087132822,
|
| 1613 |
+
"rewards/rejected": -2.7385556782130154,
|
| 1614 |
+
"rewards/accuracies": 0.83125,
|
| 1615 |
+
"rewards/margins": 1.6643950667232275,
|
| 1616 |
+
"logps/chosen": -68.02462786436081,
|
| 1617 |
+
"logps/rejected": -78.2056318640709,
|
| 1618 |
+
"epoch": 1.408,
|
| 1619 |
+
"step": 880
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"loss": 0.2997459411621094,
|
| 1623 |
+
"grad_norm": 0.24769429862499237,
|
| 1624 |
+
"learning_rate": 2.888e-06,
|
| 1625 |
+
"entropy": 0.883172089169966,
|
| 1626 |
+
"num_tokens": 11954898.0,
|
| 1627 |
+
"logits/chosen": -1.5578397765491054,
|
| 1628 |
+
"logits/rejected": -1.1913887642008372,
|
| 1629 |
+
"mean_token_accuracy": 0.6889681628905236,
|
| 1630 |
+
"rewards/chosen": -1.1197501636808738,
|
| 1631 |
+
"rewards/rejected": -3.1189077521674333,
|
| 1632 |
+
"rewards/accuracies": 0.9125,
|
| 1633 |
+
"rewards/margins": 1.9991575853899122,
|
| 1634 |
+
"logps/chosen": -66.62000039815902,
|
| 1635 |
+
"logps/rejected": -88.68679538965225,
|
| 1636 |
+
"epoch": 1.424,
|
| 1637 |
+
"step": 890
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"loss": 0.29968535900115967,
|
| 1641 |
+
"grad_norm": 0.18391333520412445,
|
| 1642 |
+
"learning_rate": 2.808e-06,
|
| 1643 |
+
"entropy": 0.8217648260528222,
|
| 1644 |
+
"num_tokens": 12089403.0,
|
| 1645 |
+
"logits/chosen": -1.654199654113943,
|
| 1646 |
+
"logits/rejected": -1.3743922696577775,
|
| 1647 |
+
"mean_token_accuracy": 0.7038281874731183,
|
| 1648 |
+
"rewards/chosen": -1.0770810046582482,
|
| 1649 |
+
"rewards/rejected": -3.0150956384284653,
|
| 1650 |
+
"rewards/accuracies": 0.86875,
|
| 1651 |
+
"rewards/margins": 1.9380146319046616,
|
| 1652 |
+
"logps/chosen": -68.25891151428223,
|
| 1653 |
+
"logps/rejected": -83.41171008348465,
|
| 1654 |
+
"epoch": 1.44,
|
| 1655 |
+
"step": 900
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"loss": 0.35481507778167726,
|
| 1659 |
+
"grad_norm": 0.1317698210477829,
|
| 1660 |
+
"learning_rate": 2.728e-06,
|
| 1661 |
+
"entropy": 0.8032614279203699,
|
| 1662 |
+
"num_tokens": 12223715.0,
|
| 1663 |
+
"logits/chosen": -1.8449737532260895,
|
| 1664 |
+
"logits/rejected": -1.4833183673721464,
|
| 1665 |
+
"mean_token_accuracy": 0.7075733580626548,
|
| 1666 |
+
"rewards/chosen": -1.1848695925727952,
|
| 1667 |
+
"rewards/rejected": -3.049835895402066,
|
| 1668 |
+
"rewards/accuracies": 0.8375,
|
| 1669 |
+
"rewards/margins": 1.8649663005489856,
|
| 1670 |
+
"logps/chosen": -63.349848651885985,
|
| 1671 |
+
"logps/rejected": -80.95961456298828,
|
| 1672 |
+
"epoch": 1.456,
|
| 1673 |
+
"step": 910
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"loss": 0.42683234214782717,
|
| 1677 |
+
"grad_norm": 0.1465081423521042,
|
| 1678 |
+
"learning_rate": 2.648e-06,
|
| 1679 |
+
"entropy": 0.8238454858452314,
|
| 1680 |
+
"num_tokens": 12357865.0,
|
| 1681 |
+
"logits/chosen": -1.7008559115211324,
|
| 1682 |
+
"logits/rejected": -1.357567183494433,
|
| 1683 |
+
"mean_token_accuracy": 0.7087827803567052,
|
| 1684 |
+
"rewards/chosen": -1.1621609168796567,
|
| 1685 |
+
"rewards/rejected": -2.8480011661653406,
|
| 1686 |
+
"rewards/accuracies": 0.825,
|
| 1687 |
+
"rewards/margins": 1.685840255022049,
|
| 1688 |
+
"logps/chosen": -63.283623594045636,
|
| 1689 |
+
"logps/rejected": -82.64039599895477,
|
| 1690 |
+
"epoch": 1.472,
|
| 1691 |
+
"step": 920
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"loss": 0.3974130392074585,
|
| 1695 |
+
"grad_norm": 0.16524887084960938,
|
| 1696 |
+
"learning_rate": 2.568e-06,
|
| 1697 |
+
"entropy": 0.8038198169349926,
|
| 1698 |
+
"num_tokens": 12492128.0,
|
| 1699 |
+
"logits/chosen": -1.698162074907906,
|
| 1700 |
+
"logits/rejected": -1.3623355208033048,
|
| 1701 |
+
"mean_token_accuracy": 0.6874856553971768,
|
| 1702 |
+
"rewards/chosen": -1.1360894276527689,
|
| 1703 |
+
"rewards/rejected": -2.871986218681559,
|
| 1704 |
+
"rewards/accuracies": 0.83125,
|
| 1705 |
+
"rewards/margins": 1.7358967858366667,
|
| 1706 |
+
"logps/chosen": -66.65757706165314,
|
| 1707 |
+
"logps/rejected": -80.95530601739884,
|
| 1708 |
+
"epoch": 1.488,
|
| 1709 |
+
"step": 930
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"loss": 0.38668370246887207,
|
| 1713 |
+
"grad_norm": 0.3068690001964569,
|
| 1714 |
+
"learning_rate": 2.488e-06,
|
| 1715 |
+
"entropy": 0.8237791604944504,
|
| 1716 |
+
"num_tokens": 12626077.0,
|
| 1717 |
+
"logits/chosen": -1.6233650559790422,
|
| 1718 |
+
"logits/rejected": -1.3661072660262759,
|
| 1719 |
+
"mean_token_accuracy": 0.6761062476783991,
|
| 1720 |
+
"rewards/chosen": -1.2516383008420235,
|
| 1721 |
+
"rewards/rejected": -2.9294361690990627,
|
| 1722 |
+
"rewards/accuracies": 0.8625,
|
| 1723 |
+
"rewards/margins": 1.6777978722006082,
|
| 1724 |
+
"logps/chosen": -65.11999430656434,
|
| 1725 |
+
"logps/rejected": -78.18645598888398,
|
| 1726 |
+
"epoch": 1.504,
|
| 1727 |
+
"step": 940
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"loss": 0.313944673538208,
|
| 1731 |
+
"grad_norm": 0.1565018594264984,
|
| 1732 |
+
"learning_rate": 2.408e-06,
|
| 1733 |
+
"entropy": 0.8455988334477297,
|
| 1734 |
+
"num_tokens": 12760098.0,
|
| 1735 |
+
"logits/chosen": -1.6097542143043664,
|
| 1736 |
+
"logits/rejected": -1.276364359775962,
|
| 1737 |
+
"mean_token_accuracy": 0.7134172804653645,
|
| 1738 |
+
"rewards/chosen": -1.2715510466623527,
|
| 1739 |
+
"rewards/rejected": -3.1404986954294145,
|
| 1740 |
+
"rewards/accuracies": 0.9,
|
| 1741 |
+
"rewards/margins": 1.868947643879801,
|
| 1742 |
+
"logps/chosen": -67.0497816324234,
|
| 1743 |
+
"logps/rejected": -84.85053882598876,
|
| 1744 |
+
"epoch": 1.52,
|
| 1745 |
+
"step": 950
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"loss": 0.3781593322753906,
|
| 1749 |
+
"grad_norm": 0.182264506816864,
|
| 1750 |
+
"learning_rate": 2.3280000000000004e-06,
|
| 1751 |
+
"entropy": 0.8393823139776941,
|
| 1752 |
+
"num_tokens": 12894652.0,
|
| 1753 |
+
"logits/chosen": -1.7312270020692502,
|
| 1754 |
+
"logits/rejected": -1.3582874506659626,
|
| 1755 |
+
"mean_token_accuracy": 0.7012655016966164,
|
| 1756 |
+
"rewards/chosen": -1.2354033031384462,
|
| 1757 |
+
"rewards/rejected": -3.0771941147511823,
|
| 1758 |
+
"rewards/accuracies": 0.81875,
|
| 1759 |
+
"rewards/margins": 1.8417908100411295,
|
| 1760 |
+
"logps/chosen": -65.62636283636093,
|
| 1761 |
+
"logps/rejected": -85.22261914014817,
|
| 1762 |
+
"epoch": 1.536,
|
| 1763 |
+
"step": 960
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"loss": 0.40849480628967283,
|
| 1767 |
+
"grad_norm": 0.1598474234342575,
|
| 1768 |
+
"learning_rate": 2.2480000000000003e-06,
|
| 1769 |
+
"entropy": 0.8594467148970579,
|
| 1770 |
+
"num_tokens": 13030162.0,
|
| 1771 |
+
"logits/chosen": -1.6824241357599359,
|
| 1772 |
+
"logits/rejected": -1.364666390483536,
|
| 1773 |
+
"mean_token_accuracy": 0.6912161948159337,
|
| 1774 |
+
"rewards/chosen": -1.2346445021947148,
|
| 1775 |
+
"rewards/rejected": -3.1347126386594026,
|
| 1776 |
+
"rewards/accuracies": 0.825,
|
| 1777 |
+
"rewards/margins": 1.900068132765591,
|
| 1778 |
+
"logps/chosen": -69.26320999860764,
|
| 1779 |
+
"logps/rejected": -92.15415328741074,
|
| 1780 |
+
"epoch": 1.552,
|
| 1781 |
+
"step": 970
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"loss": 0.31421561241149903,
|
| 1785 |
+
"grad_norm": 0.29455259442329407,
|
| 1786 |
+
"learning_rate": 2.1680000000000002e-06,
|
| 1787 |
+
"entropy": 0.9111468735121889,
|
| 1788 |
+
"num_tokens": 13165409.0,
|
| 1789 |
+
"logits/chosen": -1.4505250904980647,
|
| 1790 |
+
"logits/rejected": -1.136234173291695,
|
| 1791 |
+
"mean_token_accuracy": 0.6673261382617056,
|
| 1792 |
+
"rewards/chosen": -1.1774217841426435,
|
| 1793 |
+
"rewards/rejected": -3.0821035149041562,
|
| 1794 |
+
"rewards/accuracies": 0.8625,
|
| 1795 |
+
"rewards/margins": 1.9046817258931696,
|
| 1796 |
+
"logps/chosen": -71.22281358242034,
|
| 1797 |
+
"logps/rejected": -87.25895103216172,
|
| 1798 |
+
"epoch": 1.568,
|
| 1799 |
+
"step": 980
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"loss": 0.41238603591918943,
|
| 1803 |
+
"grad_norm": 0.2690979838371277,
|
| 1804 |
+
"learning_rate": 2.088e-06,
|
| 1805 |
+
"entropy": 0.8509069160441868,
|
| 1806 |
+
"num_tokens": 13299607.0,
|
| 1807 |
+
"logits/chosen": -1.5380459391326322,
|
| 1808 |
+
"logits/rejected": -1.3527591640050916,
|
| 1809 |
+
"mean_token_accuracy": 0.6808761429972947,
|
| 1810 |
+
"rewards/chosen": -1.2909708148508798,
|
| 1811 |
+
"rewards/rejected": -2.9686779485433363,
|
| 1812 |
+
"rewards/accuracies": 0.825,
|
| 1813 |
+
"rewards/margins": 1.677707129716873,
|
| 1814 |
+
"logps/chosen": -68.10802036523819,
|
| 1815 |
+
"logps/rejected": -78.50427184104919,
|
| 1816 |
+
"epoch": 1.584,
|
| 1817 |
+
"step": 990
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"loss": 0.2818290233612061,
|
| 1821 |
+
"grad_norm": 0.48030605912208557,
|
| 1822 |
+
"learning_rate": 2.008e-06,
|
| 1823 |
+
"entropy": 0.7622321485774591,
|
| 1824 |
+
"num_tokens": 13434502.0,
|
| 1825 |
+
"logits/chosen": -1.829241864435301,
|
| 1826 |
+
"logits/rejected": -1.4082243740070777,
|
| 1827 |
+
"mean_token_accuracy": 0.7344102315604687,
|
| 1828 |
+
"rewards/chosen": -1.2174823874112917,
|
| 1829 |
+
"rewards/rejected": -3.3832782270852477,
|
| 1830 |
+
"rewards/accuracies": 0.9125,
|
| 1831 |
+
"rewards/margins": 2.1657958355732263,
|
| 1832 |
+
"logps/chosen": -65.61667195558547,
|
| 1833 |
+
"logps/rejected": -86.77593697309494,
|
| 1834 |
+
"epoch": 1.6,
|
| 1835 |
+
"step": 1000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"loss": 0.3375999927520752,
|
| 1839 |
+
"grad_norm": 0.12456448376178741,
|
| 1840 |
+
"learning_rate": 1.928e-06,
|
| 1841 |
+
"entropy": 0.8247544016325264,
|
| 1842 |
+
"num_tokens": 13568600.0,
|
| 1843 |
+
"logits/chosen": -1.7053266821514785,
|
| 1844 |
+
"logits/rejected": -1.4006955106894101,
|
| 1845 |
+
"mean_token_accuracy": 0.7055992158129811,
|
| 1846 |
+
"rewards/chosen": -1.1076750126376282,
|
| 1847 |
+
"rewards/rejected": -3.069399724388495,
|
| 1848 |
+
"rewards/accuracies": 0.8625,
|
| 1849 |
+
"rewards/margins": 1.961724712792784,
|
| 1850 |
+
"logps/chosen": -66.50230821371079,
|
| 1851 |
+
"logps/rejected": -79.85716701745987,
|
| 1852 |
+
"epoch": 1.616,
|
| 1853 |
+
"step": 1010
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"loss": 0.2663018226623535,
|
| 1857 |
+
"grad_norm": 0.12037217617034912,
|
| 1858 |
+
"learning_rate": 1.8480000000000001e-06,
|
| 1859 |
+
"entropy": 0.8061382247862638,
|
| 1860 |
+
"num_tokens": 13702787.0,
|
| 1861 |
+
"logits/chosen": -1.7181343884757247,
|
| 1862 |
+
"logits/rejected": -1.2938053464715578,
|
| 1863 |
+
"mean_token_accuracy": 0.7232431828044354,
|
| 1864 |
+
"rewards/chosen": -0.9457807264756412,
|
| 1865 |
+
"rewards/rejected": -3.189152976346668,
|
| 1866 |
+
"rewards/accuracies": 0.89375,
|
| 1867 |
+
"rewards/margins": 2.2433722466230392,
|
| 1868 |
+
"logps/chosen": -65.12261333465577,
|
| 1869 |
+
"logps/rejected": -82.98033672571182,
|
| 1870 |
+
"epoch": 1.6320000000000001,
|
| 1871 |
+
"step": 1020
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"loss": 0.36628768444061277,
|
| 1875 |
+
"grad_norm": 0.2802140414714813,
|
| 1876 |
+
"learning_rate": 1.7680000000000003e-06,
|
| 1877 |
+
"entropy": 0.9013640963385114,
|
| 1878 |
+
"num_tokens": 13837799.0,
|
| 1879 |
+
"logits/chosen": -1.5675454347315843,
|
| 1880 |
+
"logits/rejected": -1.2078320249373804,
|
| 1881 |
+
"mean_token_accuracy": 0.7057046189904213,
|
| 1882 |
+
"rewards/chosen": -1.1604038240911905,
|
| 1883 |
+
"rewards/rejected": -2.967831138102338,
|
| 1884 |
+
"rewards/accuracies": 0.83125,
|
| 1885 |
+
"rewards/margins": 1.8074273100122809,
|
| 1886 |
+
"logps/chosen": -71.22785445451737,
|
| 1887 |
+
"logps/rejected": -84.38925257921218,
|
| 1888 |
+
"epoch": 1.6480000000000001,
|
| 1889 |
+
"step": 1030
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"loss": 0.34422104358673095,
|
| 1893 |
+
"grad_norm": 0.292550653219223,
|
| 1894 |
+
"learning_rate": 1.6880000000000002e-06,
|
| 1895 |
+
"entropy": 0.8182750114472583,
|
| 1896 |
+
"num_tokens": 13972529.0,
|
| 1897 |
+
"logits/chosen": -1.7217927195292735,
|
| 1898 |
+
"logits/rejected": -1.3031066883714544,
|
| 1899 |
+
"mean_token_accuracy": 0.7254631980322301,
|
| 1900 |
+
"rewards/chosen": -1.0691149362013674,
|
| 1901 |
+
"rewards/rejected": -3.089101183600724,
|
| 1902 |
+
"rewards/accuracies": 0.875,
|
| 1903 |
+
"rewards/margins": 2.019986251555383,
|
| 1904 |
+
"logps/chosen": -64.04030301570893,
|
| 1905 |
+
"logps/rejected": -84.32700728178024,
|
| 1906 |
+
"epoch": 1.6640000000000001,
|
| 1907 |
+
"step": 1040
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"loss": 0.33436872959136965,
|
| 1911 |
+
"grad_norm": 0.1531970351934433,
|
| 1912 |
+
"learning_rate": 1.608e-06,
|
| 1913 |
+
"entropy": 0.9578034371705144,
|
| 1914 |
+
"num_tokens": 14107336.0,
|
| 1915 |
+
"logits/chosen": -1.4119801823201867,
|
| 1916 |
+
"logits/rejected": -1.1656953772424705,
|
| 1917 |
+
"mean_token_accuracy": 0.6918271525762976,
|
| 1918 |
+
"rewards/chosen": -0.9633726783911698,
|
| 1919 |
+
"rewards/rejected": -2.8910770336631684,
|
| 1920 |
+
"rewards/accuracies": 0.8625,
|
| 1921 |
+
"rewards/margins": 1.9277043485082686,
|
| 1922 |
+
"logps/chosen": -77.73689750432968,
|
| 1923 |
+
"logps/rejected": -80.27133039236068,
|
| 1924 |
+
"epoch": 1.6800000000000002,
|
| 1925 |
+
"step": 1050
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"loss": 0.455459451675415,
|
| 1929 |
+
"grad_norm": 0.14618733525276184,
|
| 1930 |
+
"learning_rate": 1.528e-06,
|
| 1931 |
+
"entropy": 0.8247536150622181,
|
| 1932 |
+
"num_tokens": 14241540.0,
|
| 1933 |
+
"logits/chosen": -1.6757442822998638,
|
| 1934 |
+
"logits/rejected": -1.2824111642346598,
|
| 1935 |
+
"mean_token_accuracy": 0.7151812262833118,
|
| 1936 |
+
"rewards/chosen": -1.0835218713560608,
|
| 1937 |
+
"rewards/rejected": -2.7469264089362695,
|
| 1938 |
+
"rewards/accuracies": 0.8125,
|
| 1939 |
+
"rewards/margins": 1.6634045344311743,
|
| 1940 |
+
"logps/chosen": -61.41530600786209,
|
| 1941 |
+
"logps/rejected": -79.09610180854797,
|
| 1942 |
+
"epoch": 1.696,
|
| 1943 |
+
"step": 1060
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"loss": 0.2985008716583252,
|
| 1947 |
+
"grad_norm": 0.13027869164943695,
|
| 1948 |
+
"learning_rate": 1.4480000000000002e-06,
|
| 1949 |
+
"entropy": 0.9044246163713978,
|
| 1950 |
+
"num_tokens": 14376121.0,
|
| 1951 |
+
"logits/chosen": -1.5742238024916015,
|
| 1952 |
+
"logits/rejected": -1.2785115461054921,
|
| 1953 |
+
"mean_token_accuracy": 0.6957173220813274,
|
| 1954 |
+
"rewards/chosen": -1.0939421178074553,
|
| 1955 |
+
"rewards/rejected": -3.234512549918145,
|
| 1956 |
+
"rewards/accuracies": 0.89375,
|
| 1957 |
+
"rewards/margins": 2.1405704393051566,
|
| 1958 |
+
"logps/chosen": -71.98274400830269,
|
| 1959 |
+
"logps/rejected": -83.31159448623657,
|
| 1960 |
+
"epoch": 1.712,
|
| 1961 |
+
"step": 1070
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"loss": 0.4502802848815918,
|
| 1965 |
+
"grad_norm": 0.3486226797103882,
|
| 1966 |
+
"learning_rate": 1.368e-06,
|
| 1967 |
+
"entropy": 0.8118856262910412,
|
| 1968 |
+
"num_tokens": 14510553.0,
|
| 1969 |
+
"logits/chosen": -1.7772147720399374,
|
| 1970 |
+
"logits/rejected": -1.458009069817984,
|
| 1971 |
+
"mean_token_accuracy": 0.7012517482042313,
|
| 1972 |
+
"rewards/chosen": -1.1378889516170603,
|
| 1973 |
+
"rewards/rejected": -2.878922158543719,
|
| 1974 |
+
"rewards/accuracies": 0.81875,
|
| 1975 |
+
"rewards/margins": 1.7410332140512765,
|
| 1976 |
+
"logps/chosen": -66.70538790225983,
|
| 1977 |
+
"logps/rejected": -79.41604725122451,
|
| 1978 |
+
"epoch": 1.728,
|
| 1979 |
+
"step": 1080
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"loss": 0.3594567537307739,
|
| 1983 |
+
"grad_norm": 0.14351795613765717,
|
| 1984 |
+
"learning_rate": 1.288e-06,
|
| 1985 |
+
"entropy": 0.8375704119505827,
|
| 1986 |
+
"num_tokens": 14644507.0,
|
| 1987 |
+
"logits/chosen": -1.8039575345884757,
|
| 1988 |
+
"logits/rejected": -1.4812283851337462,
|
| 1989 |
+
"mean_token_accuracy": 0.7144409637898207,
|
| 1990 |
+
"rewards/chosen": -0.9819474776973948,
|
| 1991 |
+
"rewards/rejected": -2.906684129824862,
|
| 1992 |
+
"rewards/accuracies": 0.825,
|
| 1993 |
+
"rewards/margins": 1.9247366478666663,
|
| 1994 |
+
"logps/chosen": -60.533272886276244,
|
| 1995 |
+
"logps/rejected": -84.3676023721695,
|
| 1996 |
+
"epoch": 1.744,
|
| 1997 |
+
"step": 1090
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"loss": 0.3065323829650879,
|
| 2001 |
+
"grad_norm": 0.11794531345367432,
|
| 2002 |
+
"learning_rate": 1.2080000000000001e-06,
|
| 2003 |
+
"entropy": 0.875045446419972,
|
| 2004 |
+
"num_tokens": 14779558.0,
|
| 2005 |
+
"logits/chosen": -1.6532927212634152,
|
| 2006 |
+
"logits/rejected": -1.2562811780241412,
|
| 2007 |
+
"mean_token_accuracy": 0.6935827617533505,
|
| 2008 |
+
"rewards/chosen": -1.0809898364299442,
|
| 2009 |
+
"rewards/rejected": -3.053542599780485,
|
| 2010 |
+
"rewards/accuracies": 0.875,
|
| 2011 |
+
"rewards/margins": 1.9725527653470636,
|
| 2012 |
+
"logps/chosen": -66.53190513849259,
|
| 2013 |
+
"logps/rejected": -85.30938122272491,
|
| 2014 |
+
"epoch": 1.76,
|
| 2015 |
+
"step": 1100
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"loss": 0.43490915298461913,
|
| 2019 |
+
"grad_norm": 0.35169699788093567,
|
| 2020 |
+
"learning_rate": 1.128e-06,
|
| 2021 |
+
"entropy": 0.8724702712381258,
|
| 2022 |
+
"num_tokens": 14914286.0,
|
| 2023 |
+
"logits/chosen": -1.633896747832146,
|
| 2024 |
+
"logits/rejected": -1.236257593037082,
|
| 2025 |
+
"mean_token_accuracy": 0.7060686960816384,
|
| 2026 |
+
"rewards/chosen": -0.9823191232244426,
|
| 2027 |
+
"rewards/rejected": -2.7769972719892393,
|
| 2028 |
+
"rewards/accuracies": 0.76875,
|
| 2029 |
+
"rewards/margins": 1.79467814497184,
|
| 2030 |
+
"logps/chosen": -62.539447128772736,
|
| 2031 |
+
"logps/rejected": -81.97587960958481,
|
| 2032 |
+
"epoch": 1.776,
|
| 2033 |
+
"step": 1110
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"loss": 0.36343390941619874,
|
| 2037 |
+
"grad_norm": 0.21236824989318848,
|
| 2038 |
+
"learning_rate": 1.0480000000000002e-06,
|
| 2039 |
+
"entropy": 0.8476358034400618,
|
| 2040 |
+
"num_tokens": 15048840.0,
|
| 2041 |
+
"logits/chosen": -1.6855269579061596,
|
| 2042 |
+
"logits/rejected": -1.410021104757498,
|
| 2043 |
+
"mean_token_accuracy": 0.7133221762254834,
|
| 2044 |
+
"rewards/chosen": -0.8813168017921271,
|
| 2045 |
+
"rewards/rejected": -2.8057729384163395,
|
| 2046 |
+
"rewards/accuracies": 0.8625,
|
| 2047 |
+
"rewards/margins": 1.9244561415165662,
|
| 2048 |
+
"logps/chosen": -64.20361061096192,
|
| 2049 |
+
"logps/rejected": -79.37510650157928,
|
| 2050 |
+
"epoch": 1.792,
|
| 2051 |
+
"step": 1120
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"loss": 0.44675116539001464,
|
| 2055 |
+
"grad_norm": 0.10052505135536194,
|
| 2056 |
+
"learning_rate": 9.68e-07,
|
| 2057 |
+
"entropy": 0.8899719029432163,
|
| 2058 |
+
"num_tokens": 15184276.0,
|
| 2059 |
+
"logits/chosen": -1.4263961653062525,
|
| 2060 |
+
"logits/rejected": -1.24636328183648,
|
| 2061 |
+
"mean_token_accuracy": 0.6905126156285405,
|
| 2062 |
+
"rewards/chosen": -1.0598846327117826,
|
| 2063 |
+
"rewards/rejected": -2.65966563107213,
|
| 2064 |
+
"rewards/accuracies": 0.78125,
|
| 2065 |
+
"rewards/margins": 1.5997809943335597,
|
| 2066 |
+
"logps/chosen": -71.35373864173889,
|
| 2067 |
+
"logps/rejected": -82.29206101894378,
|
| 2068 |
+
"epoch": 1.808,
|
| 2069 |
+
"step": 1130
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"loss": 0.47310647964477537,
|
| 2073 |
+
"grad_norm": 0.4083561897277832,
|
| 2074 |
+
"learning_rate": 8.880000000000001e-07,
|
| 2075 |
+
"entropy": 0.8037759118797112,
|
| 2076 |
+
"num_tokens": 15317721.0,
|
| 2077 |
+
"logits/chosen": -1.5803079508150808,
|
| 2078 |
+
"logits/rejected": -1.402057939879517,
|
| 2079 |
+
"mean_token_accuracy": 0.7061841377057135,
|
| 2080 |
+
"rewards/chosen": -1.08754569782177,
|
| 2081 |
+
"rewards/rejected": -2.4734331390354782,
|
| 2082 |
+
"rewards/accuracies": 0.7875,
|
| 2083 |
+
"rewards/margins": 1.3858874359168112,
|
| 2084 |
+
"logps/chosen": -59.80233359336853,
|
| 2085 |
+
"logps/rejected": -72.9841728925705,
|
| 2086 |
+
"epoch": 1.8239999999999998,
|
| 2087 |
+
"step": 1140
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"loss": 0.36058506965637205,
|
| 2091 |
+
"grad_norm": 0.11218813806772232,
|
| 2092 |
+
"learning_rate": 8.08e-07,
|
| 2093 |
+
"entropy": 0.8560581791069126,
|
| 2094 |
+
"num_tokens": 15452553.0,
|
| 2095 |
+
"logits/chosen": -1.630976294843057,
|
| 2096 |
+
"logits/rejected": -1.248477545650383,
|
| 2097 |
+
"mean_token_accuracy": 0.6922427358105778,
|
| 2098 |
+
"rewards/chosen": -1.114829145767726,
|
| 2099 |
+
"rewards/rejected": -2.877474607457407,
|
| 2100 |
+
"rewards/accuracies": 0.85625,
|
| 2101 |
+
"rewards/margins": 1.7626454542391001,
|
| 2102 |
+
"logps/chosen": -64.00351672172546,
|
| 2103 |
+
"logps/rejected": -84.10348170995712,
|
| 2104 |
+
"epoch": 1.8399999999999999,
|
| 2105 |
+
"step": 1150
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"loss": 0.35589578151702883,
|
| 2109 |
+
"grad_norm": 0.22038227319717407,
|
| 2110 |
+
"learning_rate": 7.280000000000001e-07,
|
| 2111 |
+
"entropy": 0.8272235481563257,
|
| 2112 |
+
"num_tokens": 15586655.0,
|
| 2113 |
+
"logits/chosen": -1.8085838271541335,
|
| 2114 |
+
"logits/rejected": -1.3973146749628618,
|
| 2115 |
+
"mean_token_accuracy": 0.7218052752315998,
|
| 2116 |
+
"rewards/chosen": -1.1222759875148767,
|
| 2117 |
+
"rewards/rejected": -3.0347228790167717,
|
| 2118 |
+
"rewards/accuracies": 0.825,
|
| 2119 |
+
"rewards/margins": 1.912446884578094,
|
| 2120 |
+
"logps/chosen": -60.28894137740135,
|
| 2121 |
+
"logps/rejected": -83.4263523697853,
|
| 2122 |
+
"epoch": 1.8559999999999999,
|
| 2123 |
+
"step": 1160
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"loss": 0.4880672931671143,
|
| 2127 |
+
"grad_norm": 0.39379969239234924,
|
| 2128 |
+
"learning_rate": 6.48e-07,
|
| 2129 |
+
"entropy": 0.7986266901803901,
|
| 2130 |
+
"num_tokens": 15720172.0,
|
| 2131 |
+
"logits/chosen": -1.8278137696870522,
|
| 2132 |
+
"logits/rejected": -1.391921900675239,
|
| 2133 |
+
"mean_token_accuracy": 0.6994905389845372,
|
| 2134 |
+
"rewards/chosen": -1.1727065812981892,
|
| 2135 |
+
"rewards/rejected": -2.6952040486503392,
|
| 2136 |
+
"rewards/accuracies": 0.7625,
|
| 2137 |
+
"rewards/margins": 1.522497471049428,
|
| 2138 |
+
"logps/chosen": -58.29740512371063,
|
| 2139 |
+
"logps/rejected": -81.62007702589035,
|
| 2140 |
+
"epoch": 1.8719999999999999,
|
| 2141 |
+
"step": 1170
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"loss": 0.45793957710266114,
|
| 2145 |
+
"grad_norm": 0.23936687409877777,
|
| 2146 |
+
"learning_rate": 5.680000000000001e-07,
|
| 2147 |
+
"entropy": 0.8128495428711175,
|
| 2148 |
+
"num_tokens": 15853755.0,
|
| 2149 |
+
"logits/chosen": -1.5949269701598194,
|
| 2150 |
+
"logits/rejected": -1.4100056303290278,
|
| 2151 |
+
"mean_token_accuracy": 0.7196939526125788,
|
| 2152 |
+
"rewards/chosen": -1.171289337798953,
|
| 2153 |
+
"rewards/rejected": -2.5952760747168213,
|
| 2154 |
+
"rewards/accuracies": 0.80625,
|
| 2155 |
+
"rewards/margins": 1.4239867361262442,
|
| 2156 |
+
"logps/chosen": -63.035974252223966,
|
| 2157 |
+
"logps/rejected": -72.69420886039734,
|
| 2158 |
+
"epoch": 1.888,
|
| 2159 |
+
"step": 1180
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"loss": 0.40268754959106445,
|
| 2163 |
+
"grad_norm": 0.23110902309417725,
|
| 2164 |
+
"learning_rate": 4.88e-07,
|
| 2165 |
+
"entropy": 0.8643654328130651,
|
| 2166 |
+
"num_tokens": 15988701.0,
|
| 2167 |
+
"logits/chosen": -1.5981690852550166,
|
| 2168 |
+
"logits/rejected": -1.2750415268127462,
|
| 2169 |
+
"mean_token_accuracy": 0.701544170640409,
|
| 2170 |
+
"rewards/chosen": -1.112778129801154,
|
| 2171 |
+
"rewards/rejected": -2.7608449917752296,
|
| 2172 |
+
"rewards/accuracies": 0.825,
|
| 2173 |
+
"rewards/margins": 1.6480668615549803,
|
| 2174 |
+
"logps/chosen": -64.20587302446366,
|
| 2175 |
+
"logps/rejected": -80.89773898124695,
|
| 2176 |
+
"epoch": 1.904,
|
| 2177 |
+
"step": 1190
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"loss": 0.41562328338623045,
|
| 2181 |
+
"grad_norm": 0.15650777518749237,
|
| 2182 |
+
"learning_rate": 4.0800000000000005e-07,
|
| 2183 |
+
"entropy": 0.8853779669007054,
|
| 2184 |
+
"num_tokens": 16123227.0,
|
| 2185 |
+
"logits/chosen": -1.569114286797127,
|
| 2186 |
+
"logits/rejected": -1.2448332484858597,
|
| 2187 |
+
"mean_token_accuracy": 0.690029120258987,
|
| 2188 |
+
"rewards/chosen": -1.034638920883299,
|
| 2189 |
+
"rewards/rejected": -2.6027355837635695,
|
| 2190 |
+
"rewards/accuracies": 0.80625,
|
| 2191 |
+
"rewards/margins": 1.5680966638028622,
|
| 2192 |
+
"logps/chosen": -66.08311747312545,
|
| 2193 |
+
"logps/rejected": -78.85131504535676,
|
| 2194 |
+
"epoch": 1.92,
|
| 2195 |
+
"step": 1200
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"loss": 0.4205287456512451,
|
| 2199 |
+
"grad_norm": 0.491484671831131,
|
| 2200 |
+
"learning_rate": 3.280000000000001e-07,
|
| 2201 |
+
"entropy": 0.8714240607339889,
|
| 2202 |
+
"num_tokens": 16257799.0,
|
| 2203 |
+
"logits/chosen": -1.4655088099357518,
|
| 2204 |
+
"logits/rejected": -1.3042592207065205,
|
| 2205 |
+
"mean_token_accuracy": 0.6825227704830468,
|
| 2206 |
+
"rewards/chosen": -1.1030471502803265,
|
| 2207 |
+
"rewards/rejected": -2.9069085040129723,
|
| 2208 |
+
"rewards/accuracies": 0.84375,
|
| 2209 |
+
"rewards/margins": 1.8038613547571003,
|
| 2210 |
+
"logps/chosen": -68.32293397188187,
|
| 2211 |
+
"logps/rejected": -81.13640749454498,
|
| 2212 |
+
"epoch": 1.936,
|
| 2213 |
+
"step": 1210
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"loss": 0.3603882074356079,
|
| 2217 |
+
"grad_norm": 0.2688979208469391,
|
| 2218 |
+
"learning_rate": 2.48e-07,
|
| 2219 |
+
"entropy": 0.8230870452229283,
|
| 2220 |
+
"num_tokens": 16392755.0,
|
| 2221 |
+
"logits/chosen": -1.6268576701981179,
|
| 2222 |
+
"logits/rejected": -1.2249365496555098,
|
| 2223 |
+
"mean_token_accuracy": 0.7257544286549091,
|
| 2224 |
+
"rewards/chosen": -1.1070520436100195,
|
| 2225 |
+
"rewards/rejected": -3.0178756157412865,
|
| 2226 |
+
"rewards/accuracies": 0.84375,
|
| 2227 |
+
"rewards/margins": 1.9108235760591925,
|
| 2228 |
+
"logps/chosen": -67.37908220291138,
|
| 2229 |
+
"logps/rejected": -85.2555167913437,
|
| 2230 |
+
"epoch": 1.952,
|
| 2231 |
+
"step": 1220
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"loss": 0.3265702247619629,
|
| 2235 |
+
"grad_norm": 0.23845519125461578,
|
| 2236 |
+
"learning_rate": 1.68e-07,
|
| 2237 |
+
"entropy": 0.803886600601254,
|
| 2238 |
+
"num_tokens": 16526347.0,
|
| 2239 |
+
"logits/chosen": -1.6841453024142887,
|
| 2240 |
+
"logits/rejected": -1.339354665520107,
|
| 2241 |
+
"mean_token_accuracy": 0.7182136586867273,
|
| 2242 |
+
"rewards/chosen": -1.0071500184945763,
|
| 2243 |
+
"rewards/rejected": -2.8995601823553443,
|
| 2244 |
+
"rewards/accuracies": 0.85625,
|
| 2245 |
+
"rewards/margins": 1.8924101632088424,
|
| 2246 |
+
"logps/chosen": -58.51592025756836,
|
| 2247 |
+
"logps/rejected": -79.28181321620941,
|
| 2248 |
+
"epoch": 1.968,
|
| 2249 |
+
"step": 1230
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"loss": 0.3484647274017334,
|
| 2253 |
+
"grad_norm": 0.15120193362236023,
|
| 2254 |
+
"learning_rate": 8.800000000000001e-08,
|
| 2255 |
+
"entropy": 0.9405306902343,
|
| 2256 |
+
"num_tokens": 16660816.0,
|
| 2257 |
+
"logits/chosen": -1.4743311841723874,
|
| 2258 |
+
"logits/rejected": -1.1229403092737846,
|
| 2259 |
+
"mean_token_accuracy": 0.6844532806426287,
|
| 2260 |
+
"rewards/chosen": -0.8398088376416126,
|
| 2261 |
+
"rewards/rejected": -2.532188464165665,
|
| 2262 |
+
"rewards/accuracies": 0.85,
|
| 2263 |
+
"rewards/margins": 1.6923796351999045,
|
| 2264 |
+
"logps/chosen": -63.42561637163162,
|
| 2265 |
+
"logps/rejected": -80.4124233365059,
|
| 2266 |
+
"epoch": 1.984,
|
| 2267 |
+
"step": 1240
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"loss": 0.4248667240142822,
|
| 2271 |
+
"grad_norm": 0.25758814811706543,
|
| 2272 |
+
"learning_rate": 8e-09,
|
| 2273 |
+
"entropy": 0.7425221406621858,
|
| 2274 |
+
"num_tokens": 16794078.0,
|
| 2275 |
+
"logits/chosen": -1.8793875294754805,
|
| 2276 |
+
"logits/rejected": -1.5526757533579585,
|
| 2277 |
+
"mean_token_accuracy": 0.6997581550851464,
|
| 2278 |
+
"rewards/chosen": -0.9529893849976361,
|
| 2279 |
+
"rewards/rejected": -2.631736672006082,
|
| 2280 |
+
"rewards/accuracies": 0.79375,
|
| 2281 |
+
"rewards/margins": 1.678747284412384,
|
| 2282 |
+
"logps/chosen": -57.37704545259476,
|
| 2283 |
+
"logps/rejected": -75.33999993801118,
|
| 2284 |
+
"epoch": 2.0,
|
| 2285 |
+
"step": 1250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"train_runtime": 5988.37,
|
| 2289 |
+
"train_samples_per_second": 3.34,
|
| 2290 |
+
"train_steps_per_second": 0.209,
|
| 2291 |
+
"total_flos": 1.3612284112597402e+17,
|
| 2292 |
+
"train_loss": 0.41622718410491943,
|
| 2293 |
+
"epoch": 2.0,
|
| 2294 |
+
"step": 1250
|
| 2295 |
+
}
|
| 2296 |
+
],
|
| 2297 |
+
"validation_monitor_size": 0,
|
| 2298 |
+
"validation_monitor_selection": "fixed_seed_random_without_replacement",
|
| 2299 |
+
"validation_monitor_seed": 22,
|
| 2300 |
+
"validation_monitor_indices": [],
|
| 2301 |
+
"early_stopping_patience": null
|
| 2302 |
+
}
|