Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- D_full_emf1_seed23/adapter_config.json +43 -0
- D_full_emf1_seed23/adapter_model.safetensors +3 -0
- D_full_emf1_seed23/chat_template.jinja +54 -0
- D_full_emf1_seed23/tokenizer.json +3 -0
- D_full_emf1_seed23/tokenizer_config.json +29 -0
- D_full_emf1_seed23/training_summary.json +2302 -0
.gitattributes
CHANGED
|
@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
D_full_emf1_seed22/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
D_full_emf1_seed23/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
D_full_emf1_seed23/adapter_config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"megatron_config": null,
|
| 23 |
+
"megatron_core": "megatron.core",
|
| 24 |
+
"modules_to_save": null,
|
| 25 |
+
"peft_type": "LORA",
|
| 26 |
+
"peft_version": "0.18.1",
|
| 27 |
+
"qalora_group_size": 16,
|
| 28 |
+
"r": 16,
|
| 29 |
+
"rank_pattern": {},
|
| 30 |
+
"revision": null,
|
| 31 |
+
"target_modules": [
|
| 32 |
+
"k_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"v_proj",
|
| 35 |
+
"q_proj"
|
| 36 |
+
],
|
| 37 |
+
"target_parameters": null,
|
| 38 |
+
"task_type": "CAUSAL_LM",
|
| 39 |
+
"trainable_token_indices": null,
|
| 40 |
+
"use_dora": false,
|
| 41 |
+
"use_qalora": false,
|
| 42 |
+
"use_rslora": false
|
| 43 |
+
}
|
D_full_emf1_seed23/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:97c635e919ed27f342872bd34b3097728e23ebc152c6db574e0f2f2ca869aad3
|
| 3 |
+
size 17462432
|
D_full_emf1_seed23/chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
D_full_emf1_seed23/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b4360dd6a184650ffc48056c2569bc603f896c5adfe94b10f1c79f809638aa5
|
| 3 |
+
size 11422166
|
D_full_emf1_seed23/tokenizer_config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": true,
|
| 24 |
+
"model_max_length": 131072,
|
| 25 |
+
"pad_token": "<|endoftext|>",
|
| 26 |
+
"split_special_tokens": false,
|
| 27 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 28 |
+
"unk_token": null
|
| 29 |
+
}
|
D_full_emf1_seed23/training_summary.json
ADDED
|
@@ -0,0 +1,2302 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_name": "Qwen/Qwen2.5-1.5B-Instruct",
|
| 3 |
+
"resolved_model": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
|
| 4 |
+
"num_train_records": 10000,
|
| 5 |
+
"num_validation_records_available": 0,
|
| 6 |
+
"training_config": {
|
| 7 |
+
"lora_rank": 16,
|
| 8 |
+
"lora_alpha": 32,
|
| 9 |
+
"learning_rate": 1e-05,
|
| 10 |
+
"num_epochs": 2,
|
| 11 |
+
"per_device_batch_size": 1,
|
| 12 |
+
"per_device_eval_batch_size": 1,
|
| 13 |
+
"gradient_accumulation_steps": 16,
|
| 14 |
+
"beta": 0.1,
|
| 15 |
+
"max_length": 512,
|
| 16 |
+
"max_prompt_length": 384,
|
| 17 |
+
"seed": 23,
|
| 18 |
+
"save_steps": 250,
|
| 19 |
+
"save_total_limit": 2,
|
| 20 |
+
"resume_from_checkpoint": null,
|
| 21 |
+
"precision_dtype": "torch.bfloat16",
|
| 22 |
+
"bf16": true,
|
| 23 |
+
"fp16": false,
|
| 24 |
+
"tokenizer_add_bos_token": false
|
| 25 |
+
},
|
| 26 |
+
"global_step": 1250,
|
| 27 |
+
"best_metric": null,
|
| 28 |
+
"best_model_checkpoint": null,
|
| 29 |
+
"train_metrics": {
|
| 30 |
+
"train_runtime": 6000.0528,
|
| 31 |
+
"train_samples_per_second": 3.333,
|
| 32 |
+
"train_steps_per_second": 0.208,
|
| 33 |
+
"total_flos": 1.3612284112597402e+17,
|
| 34 |
+
"train_loss": 0.452988232421875
|
| 35 |
+
},
|
| 36 |
+
"log_history": [
|
| 37 |
+
{
|
| 38 |
+
"loss": 0.6925445079803467,
|
| 39 |
+
"grad_norm": 0.13106924295425415,
|
| 40 |
+
"learning_rate": 9.928e-06,
|
| 41 |
+
"entropy": 0.8430028882343322,
|
| 42 |
+
"num_tokens": 133829.0,
|
| 43 |
+
"logits/chosen": -1.0539194697152947,
|
| 44 |
+
"logits/rejected": -1.0023873638157341,
|
| 45 |
+
"mean_token_accuracy": 0.7161535605788231,
|
| 46 |
+
"rewards/chosen": 0.0006855225969047751,
|
| 47 |
+
"rewards/rejected": -0.0007722664523498679,
|
| 48 |
+
"rewards/accuracies": 0.54375,
|
| 49 |
+
"rewards/margins": 0.001457789224514272,
|
| 50 |
+
"logps/chosen": -53.864716422557834,
|
| 51 |
+
"logps/rejected": -52.49959944486618,
|
| 52 |
+
"epoch": 0.016,
|
| 53 |
+
"step": 10
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"loss": 0.6846516132354736,
|
| 57 |
+
"grad_norm": 0.09445367753505707,
|
| 58 |
+
"learning_rate": 9.848000000000001e-06,
|
| 59 |
+
"entropy": 0.7865605641389266,
|
| 60 |
+
"num_tokens": 268096.0,
|
| 61 |
+
"logits/chosen": -1.1735790360523626,
|
| 62 |
+
"logits/rejected": -1.0035928764768882,
|
| 63 |
+
"mean_token_accuracy": 0.7369648787193,
|
| 64 |
+
"rewards/chosen": -0.013590605450008297,
|
| 65 |
+
"rewards/rejected": -0.031154271999957926,
|
| 66 |
+
"rewards/accuracies": 0.66875,
|
| 67 |
+
"rewards/margins": 0.017563666505157016,
|
| 68 |
+
"logps/chosen": -48.01739143133163,
|
| 69 |
+
"logps/rejected": -56.277624583244325,
|
| 70 |
+
"epoch": 0.032,
|
| 71 |
+
"step": 20
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"loss": 0.6588961601257324,
|
| 75 |
+
"grad_norm": 0.1251889169216156,
|
| 76 |
+
"learning_rate": 9.768e-06,
|
| 77 |
+
"entropy": 0.9058249982655979,
|
| 78 |
+
"num_tokens": 403100.0,
|
| 79 |
+
"logits/chosen": -0.8840507368526256,
|
| 80 |
+
"logits/rejected": -0.8496724761598647,
|
| 81 |
+
"mean_token_accuracy": 0.7215570319443941,
|
| 82 |
+
"rewards/chosen": -0.034349703128009425,
|
| 83 |
+
"rewards/rejected": -0.10719101511076587,
|
| 84 |
+
"rewards/accuracies": 0.78125,
|
| 85 |
+
"rewards/margins": 0.07284131128399167,
|
| 86 |
+
"logps/chosen": -61.04792865514755,
|
| 87 |
+
"logps/rejected": -61.20545055866241,
|
| 88 |
+
"epoch": 0.048,
|
| 89 |
+
"step": 30
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"loss": 0.6471754550933838,
|
| 93 |
+
"grad_norm": 0.10575459152460098,
|
| 94 |
+
"learning_rate": 9.688000000000001e-06,
|
| 95 |
+
"entropy": 0.8045565326465294,
|
| 96 |
+
"num_tokens": 536957.0,
|
| 97 |
+
"logits/chosen": -0.9277418629935623,
|
| 98 |
+
"logits/rejected": -0.8127428054126742,
|
| 99 |
+
"mean_token_accuracy": 0.7228790700435639,
|
| 100 |
+
"rewards/chosen": -0.0844828439738194,
|
| 101 |
+
"rewards/rejected": -0.18686079327890184,
|
| 102 |
+
"rewards/accuracies": 0.7875,
|
| 103 |
+
"rewards/margins": 0.10237794981803745,
|
| 104 |
+
"logps/chosen": -55.788045692443845,
|
| 105 |
+
"logps/rejected": -51.87638405561447,
|
| 106 |
+
"epoch": 0.064,
|
| 107 |
+
"step": 40
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"loss": 0.6388973236083985,
|
| 111 |
+
"grad_norm": 0.11447733640670776,
|
| 112 |
+
"learning_rate": 9.608e-06,
|
| 113 |
+
"entropy": 0.8310880722478032,
|
| 114 |
+
"num_tokens": 672162.0,
|
| 115 |
+
"logits/chosen": -0.932203738722083,
|
| 116 |
+
"logits/rejected": -0.8394863060549598,
|
| 117 |
+
"mean_token_accuracy": 0.7292114172130824,
|
| 118 |
+
"rewards/chosen": -0.17734425760645536,
|
| 119 |
+
"rewards/rejected": -0.3075421278263093,
|
| 120 |
+
"rewards/accuracies": 0.7,
|
| 121 |
+
"rewards/margins": 0.13019787039374933,
|
| 122 |
+
"logps/chosen": -57.186359715461734,
|
| 123 |
+
"logps/rejected": -59.51659116744995,
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"step": 50
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"loss": 0.584157657623291,
|
| 129 |
+
"grad_norm": 0.10827931016683578,
|
| 130 |
+
"learning_rate": 9.528000000000001e-06,
|
| 131 |
+
"entropy": 0.7849437956116162,
|
| 132 |
+
"num_tokens": 806028.0,
|
| 133 |
+
"logits/chosen": -1.0795894372571302,
|
| 134 |
+
"logits/rejected": -0.9225630624384638,
|
| 135 |
+
"mean_token_accuracy": 0.723660983517766,
|
| 136 |
+
"rewards/chosen": -0.20687476017265,
|
| 137 |
+
"rewards/rejected": -0.4755556690506637,
|
| 138 |
+
"rewards/accuracies": 0.7625,
|
| 139 |
+
"rewards/margins": 0.26868090690113605,
|
| 140 |
+
"logps/chosen": -56.57579958438873,
|
| 141 |
+
"logps/rejected": -55.61295554637909,
|
| 142 |
+
"epoch": 0.096,
|
| 143 |
+
"step": 60
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"loss": 0.5784035205841065,
|
| 147 |
+
"grad_norm": 0.1285804808139801,
|
| 148 |
+
"learning_rate": 9.448e-06,
|
| 149 |
+
"entropy": 0.7840334481210448,
|
| 150 |
+
"num_tokens": 940207.0,
|
| 151 |
+
"logits/chosen": -1.0125293609936816,
|
| 152 |
+
"logits/rejected": -0.8529993941692094,
|
| 153 |
+
"mean_token_accuracy": 0.7125357061624527,
|
| 154 |
+
"rewards/chosen": -0.36272725756798535,
|
| 155 |
+
"rewards/rejected": -0.6677537690615282,
|
| 156 |
+
"rewards/accuracies": 0.75,
|
| 157 |
+
"rewards/margins": 0.3050265107536688,
|
| 158 |
+
"logps/chosen": -56.21787704229355,
|
| 159 |
+
"logps/rejected": -60.29286493062973,
|
| 160 |
+
"epoch": 0.112,
|
| 161 |
+
"step": 70
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"loss": 0.5129581451416015,
|
| 165 |
+
"grad_norm": 0.12409700453281403,
|
| 166 |
+
"learning_rate": 9.368e-06,
|
| 167 |
+
"entropy": 0.8727153117069975,
|
| 168 |
+
"num_tokens": 1075324.0,
|
| 169 |
+
"logits/chosen": -0.9037921948846058,
|
| 170 |
+
"logits/rejected": -0.7711921778344156,
|
| 171 |
+
"mean_token_accuracy": 0.697669517993927,
|
| 172 |
+
"rewards/chosen": -0.45928073021859744,
|
| 173 |
+
"rewards/rejected": -0.9675684184629063,
|
| 174 |
+
"rewards/accuracies": 0.8125,
|
| 175 |
+
"rewards/margins": 0.5082876835484058,
|
| 176 |
+
"logps/chosen": -61.03637161254883,
|
| 177 |
+
"logps/rejected": -65.84102531671525,
|
| 178 |
+
"epoch": 0.128,
|
| 179 |
+
"step": 80
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"loss": 0.5763972282409668,
|
| 183 |
+
"grad_norm": 0.15729039907455444,
|
| 184 |
+
"learning_rate": 9.288e-06,
|
| 185 |
+
"entropy": 0.7766714802710339,
|
| 186 |
+
"num_tokens": 1210083.0,
|
| 187 |
+
"logits/chosen": -1.2244412665554385,
|
| 188 |
+
"logits/rejected": -0.9376890599800589,
|
| 189 |
+
"mean_token_accuracy": 0.6997604581527412,
|
| 190 |
+
"rewards/chosen": -0.7116548144666013,
|
| 191 |
+
"rewards/rejected": -1.167988013192371,
|
| 192 |
+
"rewards/accuracies": 0.7,
|
| 193 |
+
"rewards/margins": 0.4563331992831081,
|
| 194 |
+
"logps/chosen": -58.59838794469833,
|
| 195 |
+
"logps/rejected": -67.74620436429977,
|
| 196 |
+
"epoch": 0.144,
|
| 197 |
+
"step": 90
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"loss": 0.5035501956939697,
|
| 201 |
+
"grad_norm": 0.15428121387958527,
|
| 202 |
+
"learning_rate": 9.208e-06,
|
| 203 |
+
"entropy": 0.7132123028452042,
|
| 204 |
+
"num_tokens": 1343049.0,
|
| 205 |
+
"logits/chosen": -1.3961682586621582,
|
| 206 |
+
"logits/rejected": -1.0052104447720165,
|
| 207 |
+
"mean_token_accuracy": 0.7140212506055832,
|
| 208 |
+
"rewards/chosen": -0.7763332309667021,
|
| 209 |
+
"rewards/rejected": -1.3943074161908953,
|
| 210 |
+
"rewards/accuracies": 0.78125,
|
| 211 |
+
"rewards/margins": 0.6179741863161325,
|
| 212 |
+
"logps/chosen": -47.84958149194718,
|
| 213 |
+
"logps/rejected": -64.15119704008103,
|
| 214 |
+
"epoch": 0.16,
|
| 215 |
+
"step": 100
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"loss": 0.5138976573944092,
|
| 219 |
+
"grad_norm": 0.20266199111938477,
|
| 220 |
+
"learning_rate": 9.128e-06,
|
| 221 |
+
"entropy": 0.7525328360585263,
|
| 222 |
+
"num_tokens": 1477024.0,
|
| 223 |
+
"logits/chosen": -1.3525818556821039,
|
| 224 |
+
"logits/rejected": -1.1510224510108018,
|
| 225 |
+
"mean_token_accuracy": 0.7064479270949959,
|
| 226 |
+
"rewards/chosen": -0.9448475810582749,
|
| 227 |
+
"rewards/rejected": -1.6548427081666888,
|
| 228 |
+
"rewards/accuracies": 0.76875,
|
| 229 |
+
"rewards/margins": 0.7099951250478626,
|
| 230 |
+
"logps/chosen": -60.176875388622285,
|
| 231 |
+
"logps/rejected": -70.95503640174866,
|
| 232 |
+
"epoch": 0.176,
|
| 233 |
+
"step": 110
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"loss": 0.49973697662353517,
|
| 237 |
+
"grad_norm": 0.14444415271282196,
|
| 238 |
+
"learning_rate": 9.048e-06,
|
| 239 |
+
"entropy": 0.7536728045437485,
|
| 240 |
+
"num_tokens": 1610835.0,
|
| 241 |
+
"logits/chosen": -1.2999691318206685,
|
| 242 |
+
"logits/rejected": -1.0529130678090677,
|
| 243 |
+
"mean_token_accuracy": 0.7022320762276649,
|
| 244 |
+
"rewards/chosen": -1.0002616313366162,
|
| 245 |
+
"rewards/rejected": -1.7002884576795623,
|
| 246 |
+
"rewards/accuracies": 0.79375,
|
| 247 |
+
"rewards/margins": 0.7000268258154392,
|
| 248 |
+
"logps/chosen": -60.60152515172958,
|
| 249 |
+
"logps/rejected": -68.34930142164231,
|
| 250 |
+
"epoch": 0.192,
|
| 251 |
+
"step": 120
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"loss": 0.5298746585845947,
|
| 255 |
+
"grad_norm": 0.146174356341362,
|
| 256 |
+
"learning_rate": 8.968000000000001e-06,
|
| 257 |
+
"entropy": 0.8164717306150123,
|
| 258 |
+
"num_tokens": 1745826.0,
|
| 259 |
+
"logits/chosen": -1.23030573718855,
|
| 260 |
+
"logits/rejected": -1.0113300163986259,
|
| 261 |
+
"mean_token_accuracy": 0.6869832375086844,
|
| 262 |
+
"rewards/chosen": -1.121532903493062,
|
| 263 |
+
"rewards/rejected": -1.816200914932415,
|
| 264 |
+
"rewards/accuracies": 0.75,
|
| 265 |
+
"rewards/margins": 0.6946680108085275,
|
| 266 |
+
"logps/chosen": -65.62226895093917,
|
| 267 |
+
"logps/rejected": -73.85321601629258,
|
| 268 |
+
"epoch": 0.208,
|
| 269 |
+
"step": 130
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"loss": 0.45813837051391604,
|
| 273 |
+
"grad_norm": 0.1630212813615799,
|
| 274 |
+
"learning_rate": 8.888e-06,
|
| 275 |
+
"entropy": 0.8052778334313189,
|
| 276 |
+
"num_tokens": 1880204.0,
|
| 277 |
+
"logits/chosen": -1.3770508583620331,
|
| 278 |
+
"logits/rejected": -1.042430434119616,
|
| 279 |
+
"mean_token_accuracy": 0.6877053484320641,
|
| 280 |
+
"rewards/chosen": -1.0758028986107093,
|
| 281 |
+
"rewards/rejected": -1.998533202026738,
|
| 282 |
+
"rewards/accuracies": 0.8,
|
| 283 |
+
"rewards/margins": 0.9227303057909012,
|
| 284 |
+
"logps/chosen": -62.81753187179565,
|
| 285 |
+
"logps/rejected": -76.81555581092834,
|
| 286 |
+
"epoch": 0.224,
|
| 287 |
+
"step": 140
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"loss": 0.49399361610412595,
|
| 291 |
+
"grad_norm": 0.1177593395113945,
|
| 292 |
+
"learning_rate": 8.808000000000001e-06,
|
| 293 |
+
"entropy": 0.7507179609965533,
|
| 294 |
+
"num_tokens": 2014712.0,
|
| 295 |
+
"logits/chosen": -1.2794385478622627,
|
| 296 |
+
"logits/rejected": -0.9678105828288878,
|
| 297 |
+
"mean_token_accuracy": 0.6995128920301795,
|
| 298 |
+
"rewards/chosen": -1.1146712710498832,
|
| 299 |
+
"rewards/rejected": -1.9815675146179275,
|
| 300 |
+
"rewards/accuracies": 0.78125,
|
| 301 |
+
"rewards/margins": 0.8668962439522148,
|
| 302 |
+
"logps/chosen": -59.96316148042679,
|
| 303 |
+
"logps/rejected": -76.80156726837158,
|
| 304 |
+
"epoch": 0.24,
|
| 305 |
+
"step": 150
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"loss": 0.48389606475830077,
|
| 309 |
+
"grad_norm": 0.1614522784948349,
|
| 310 |
+
"learning_rate": 8.728e-06,
|
| 311 |
+
"entropy": 0.767244737688452,
|
| 312 |
+
"num_tokens": 2148822.0,
|
| 313 |
+
"logits/chosen": -1.4473581666816784,
|
| 314 |
+
"logits/rejected": -1.1119019244646073,
|
| 315 |
+
"mean_token_accuracy": 0.7071263090707361,
|
| 316 |
+
"rewards/chosen": -1.1861407140095253,
|
| 317 |
+
"rewards/rejected": -2.0622633112128823,
|
| 318 |
+
"rewards/accuracies": 0.7625,
|
| 319 |
+
"rewards/margins": 0.8761225962080061,
|
| 320 |
+
"logps/chosen": -63.51802418231964,
|
| 321 |
+
"logps/rejected": -73.83984756469727,
|
| 322 |
+
"epoch": 0.256,
|
| 323 |
+
"step": 160
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"loss": 0.538500165939331,
|
| 327 |
+
"grad_norm": 0.20714245736598969,
|
| 328 |
+
"learning_rate": 8.648000000000001e-06,
|
| 329 |
+
"entropy": 0.7619314239942468,
|
| 330 |
+
"num_tokens": 2282595.0,
|
| 331 |
+
"logits/chosen": -1.2495616339762585,
|
| 332 |
+
"logits/rejected": -1.0757967526889842,
|
| 333 |
+
"mean_token_accuracy": 0.7071390090510249,
|
| 334 |
+
"rewards/chosen": -1.1613323532452342,
|
| 335 |
+
"rewards/rejected": -1.9118764613522217,
|
| 336 |
+
"rewards/accuracies": 0.7375,
|
| 337 |
+
"rewards/margins": 0.7505441021174193,
|
| 338 |
+
"logps/chosen": -64.5843218922615,
|
| 339 |
+
"logps/rejected": -70.18264303207397,
|
| 340 |
+
"epoch": 0.272,
|
| 341 |
+
"step": 170
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"loss": 0.46224379539489746,
|
| 345 |
+
"grad_norm": 0.11571654677391052,
|
| 346 |
+
"learning_rate": 8.568e-06,
|
| 347 |
+
"entropy": 0.7852839057624805,
|
| 348 |
+
"num_tokens": 2418209.0,
|
| 349 |
+
"logits/chosen": -1.294616462231843,
|
| 350 |
+
"logits/rejected": -1.1091888609193803,
|
| 351 |
+
"mean_token_accuracy": 0.6972204014658928,
|
| 352 |
+
"rewards/chosen": -1.2717912456544582,
|
| 353 |
+
"rewards/rejected": -2.308060463386937,
|
| 354 |
+
"rewards/accuracies": 0.7875,
|
| 355 |
+
"rewards/margins": 1.036269218660891,
|
| 356 |
+
"logps/chosen": -70.35009993314743,
|
| 357 |
+
"logps/rejected": -85.53125276565552,
|
| 358 |
+
"epoch": 0.288,
|
| 359 |
+
"step": 180
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"loss": 0.5569064140319824,
|
| 363 |
+
"grad_norm": 0.14751504361629486,
|
| 364 |
+
"learning_rate": 8.488e-06,
|
| 365 |
+
"entropy": 0.7947878750099335,
|
| 366 |
+
"num_tokens": 2552678.0,
|
| 367 |
+
"logits/chosen": -1.2534910448087464,
|
| 368 |
+
"logits/rejected": -1.210045775692255,
|
| 369 |
+
"mean_token_accuracy": 0.6766968851909041,
|
| 370 |
+
"rewards/chosen": -1.2856982867000624,
|
| 371 |
+
"rewards/rejected": -2.0018560449709186,
|
| 372 |
+
"rewards/accuracies": 0.71875,
|
| 373 |
+
"rewards/margins": 0.7161577565595507,
|
| 374 |
+
"logps/chosen": -69.15452353954315,
|
| 375 |
+
"logps/rejected": -74.11077741384506,
|
| 376 |
+
"epoch": 0.304,
|
| 377 |
+
"step": 190
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"loss": 0.5386476516723633,
|
| 381 |
+
"grad_norm": 0.16784028708934784,
|
| 382 |
+
"learning_rate": 8.408e-06,
|
| 383 |
+
"entropy": 0.7158474079507868,
|
| 384 |
+
"num_tokens": 2686349.0,
|
| 385 |
+
"logits/chosen": -1.3480244280549025,
|
| 386 |
+
"logits/rejected": -1.1111837274633218,
|
| 387 |
+
"mean_token_accuracy": 0.7138814412057399,
|
| 388 |
+
"rewards/chosen": -1.1463188135036035,
|
| 389 |
+
"rewards/rejected": -1.8588492201874032,
|
| 390 |
+
"rewards/accuracies": 0.75,
|
| 391 |
+
"rewards/margins": 0.7125304029788821,
|
| 392 |
+
"logps/chosen": -64.51462678909301,
|
| 393 |
+
"logps/rejected": -68.44324979782104,
|
| 394 |
+
"epoch": 0.32,
|
| 395 |
+
"step": 200
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"loss": 0.41974539756774903,
|
| 399 |
+
"grad_norm": 0.14496931433677673,
|
| 400 |
+
"learning_rate": 8.328e-06,
|
| 401 |
+
"entropy": 0.7375683297024807,
|
| 402 |
+
"num_tokens": 2821059.0,
|
| 403 |
+
"logits/chosen": -1.5740966382454753,
|
| 404 |
+
"logits/rejected": -1.1473181951261093,
|
| 405 |
+
"mean_token_accuracy": 0.7172042991966009,
|
| 406 |
+
"rewards/chosen": -0.9963440826628357,
|
| 407 |
+
"rewards/rejected": -2.130572400346864,
|
| 408 |
+
"rewards/accuracies": 0.8375,
|
| 409 |
+
"rewards/margins": 1.1342283125966788,
|
| 410 |
+
"logps/chosen": -57.70991215705872,
|
| 411 |
+
"logps/rejected": -79.44850307703018,
|
| 412 |
+
"epoch": 0.336,
|
| 413 |
+
"step": 210
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"loss": 0.419812536239624,
|
| 417 |
+
"grad_norm": 0.16907528042793274,
|
| 418 |
+
"learning_rate": 8.248e-06,
|
| 419 |
+
"entropy": 0.6834075984719675,
|
| 420 |
+
"num_tokens": 2954598.0,
|
| 421 |
+
"logits/chosen": -1.5877836993661458,
|
| 422 |
+
"logits/rejected": -1.2788736724764815,
|
| 423 |
+
"mean_token_accuracy": 0.7134232673794031,
|
| 424 |
+
"rewards/chosen": -1.0092000641292542,
|
| 425 |
+
"rewards/rejected": -2.1286219318164514,
|
| 426 |
+
"rewards/accuracies": 0.81875,
|
| 427 |
+
"rewards/margins": 1.1194218643940985,
|
| 428 |
+
"logps/chosen": -57.48038564920425,
|
| 429 |
+
"logps/rejected": -71.7005465388298,
|
| 430 |
+
"epoch": 0.352,
|
| 431 |
+
"step": 220
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"loss": 0.5000779628753662,
|
| 435 |
+
"grad_norm": 0.21664294600486755,
|
| 436 |
+
"learning_rate": 8.168e-06,
|
| 437 |
+
"entropy": 0.6480607748584589,
|
| 438 |
+
"num_tokens": 3087946.0,
|
| 439 |
+
"logits/chosen": -1.6257516518788577,
|
| 440 |
+
"logits/rejected": -1.3811952023449765,
|
| 441 |
+
"mean_token_accuracy": 0.7170056706294418,
|
| 442 |
+
"rewards/chosen": -1.1005187036818824,
|
| 443 |
+
"rewards/rejected": -2.0874655603431167,
|
| 444 |
+
"rewards/accuracies": 0.75,
|
| 445 |
+
"rewards/margins": 0.9869468596763908,
|
| 446 |
+
"logps/chosen": -53.01660460233688,
|
| 447 |
+
"logps/rejected": -72.73401219844818,
|
| 448 |
+
"epoch": 0.368,
|
| 449 |
+
"step": 230
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"loss": 0.41640191078186034,
|
| 453 |
+
"grad_norm": 0.13618744909763336,
|
| 454 |
+
"learning_rate": 8.088e-06,
|
| 455 |
+
"entropy": 0.735815642774105,
|
| 456 |
+
"num_tokens": 3221706.0,
|
| 457 |
+
"logits/chosen": -1.4518982412503076,
|
| 458 |
+
"logits/rejected": -1.2350048695000762,
|
| 459 |
+
"mean_token_accuracy": 0.7111645181663334,
|
| 460 |
+
"rewards/chosen": -1.09686146506574,
|
| 461 |
+
"rewards/rejected": -2.2318367584608496,
|
| 462 |
+
"rewards/accuracies": 0.8375,
|
| 463 |
+
"rewards/margins": 1.1349752927199006,
|
| 464 |
+
"logps/chosen": -58.26607400178909,
|
| 465 |
+
"logps/rejected": -74.5952651977539,
|
| 466 |
+
"epoch": 0.384,
|
| 467 |
+
"step": 240
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"loss": 0.4902942180633545,
|
| 471 |
+
"grad_norm": 0.21052443981170654,
|
| 472 |
+
"learning_rate": 8.008e-06,
|
| 473 |
+
"entropy": 0.8077279043849558,
|
| 474 |
+
"num_tokens": 3356185.0,
|
| 475 |
+
"logits/chosen": -1.3508111761120858,
|
| 476 |
+
"logits/rejected": -1.1624117580002962,
|
| 477 |
+
"mean_token_accuracy": 0.6951078314334154,
|
| 478 |
+
"rewards/chosen": -1.1994671203377947,
|
| 479 |
+
"rewards/rejected": -2.173786994814873,
|
| 480 |
+
"rewards/accuracies": 0.7625,
|
| 481 |
+
"rewards/margins": 0.9743198692798615,
|
| 482 |
+
"logps/chosen": -70.02799752950668,
|
| 483 |
+
"logps/rejected": -73.7796833872795,
|
| 484 |
+
"epoch": 0.4,
|
| 485 |
+
"step": 250
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"loss": 0.429164457321167,
|
| 489 |
+
"grad_norm": 0.16093610227108002,
|
| 490 |
+
"learning_rate": 7.928e-06,
|
| 491 |
+
"entropy": 0.7886518793180585,
|
| 492 |
+
"num_tokens": 3490468.0,
|
| 493 |
+
"logits/chosen": -1.3261231065347217,
|
| 494 |
+
"logits/rejected": -1.098070382730149,
|
| 495 |
+
"mean_token_accuracy": 0.6995166232809424,
|
| 496 |
+
"rewards/chosen": -1.0775358187383972,
|
| 497 |
+
"rewards/rejected": -2.2392933295108377,
|
| 498 |
+
"rewards/accuracies": 0.86875,
|
| 499 |
+
"rewards/margins": 1.1617575039621442,
|
| 500 |
+
"logps/chosen": -63.11482725143433,
|
| 501 |
+
"logps/rejected": -80.99848538637161,
|
| 502 |
+
"epoch": 0.416,
|
| 503 |
+
"step": 260
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"loss": 0.4255237102508545,
|
| 507 |
+
"grad_norm": 0.194253608584404,
|
| 508 |
+
"learning_rate": 7.848000000000002e-06,
|
| 509 |
+
"entropy": 0.7764837886788882,
|
| 510 |
+
"num_tokens": 3625248.0,
|
| 511 |
+
"logits/chosen": -1.3838655430252698,
|
| 512 |
+
"logits/rejected": -1.1227413434837863,
|
| 513 |
+
"mean_token_accuracy": 0.6960799667984248,
|
| 514 |
+
"rewards/chosen": -1.1910365224350243,
|
| 515 |
+
"rewards/rejected": -2.2339234950137326,
|
| 516 |
+
"rewards/accuracies": 0.80625,
|
| 517 |
+
"rewards/margins": 1.0428869728930295,
|
| 518 |
+
"logps/chosen": -67.41155977249146,
|
| 519 |
+
"logps/rejected": -79.18053728342056,
|
| 520 |
+
"epoch": 0.432,
|
| 521 |
+
"step": 270
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"loss": 0.42293658256530764,
|
| 525 |
+
"grad_norm": 0.2024957239627838,
|
| 526 |
+
"learning_rate": 7.768e-06,
|
| 527 |
+
"entropy": 0.7524384143820498,
|
| 528 |
+
"num_tokens": 3759703.0,
|
| 529 |
+
"logits/chosen": -1.5700863775036777,
|
| 530 |
+
"logits/rejected": -1.2974879702996405,
|
| 531 |
+
"mean_token_accuracy": 0.7092206623405218,
|
| 532 |
+
"rewards/chosen": -1.181411183514865,
|
| 533 |
+
"rewards/rejected": -2.3785619896138086,
|
| 534 |
+
"rewards/accuracies": 0.8375,
|
| 535 |
+
"rewards/margins": 1.1971508040558547,
|
| 536 |
+
"logps/chosen": -64.79122279882431,
|
| 537 |
+
"logps/rejected": -78.0537760257721,
|
| 538 |
+
"epoch": 0.448,
|
| 539 |
+
"step": 280
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"loss": 0.49022955894470216,
|
| 543 |
+
"grad_norm": 0.23429764807224274,
|
| 544 |
+
"learning_rate": 7.688000000000002e-06,
|
| 545 |
+
"entropy": 0.7095387592096813,
|
| 546 |
+
"num_tokens": 3893200.0,
|
| 547 |
+
"logits/chosen": -1.644883176910696,
|
| 548 |
+
"logits/rejected": -1.3690703175511028,
|
| 549 |
+
"mean_token_accuracy": 0.7099431391805411,
|
| 550 |
+
"rewards/chosen": -1.208722565032076,
|
| 551 |
+
"rewards/rejected": -2.302569864381803,
|
| 552 |
+
"rewards/accuracies": 0.73125,
|
| 553 |
+
"rewards/margins": 1.0938472978305072,
|
| 554 |
+
"logps/chosen": -58.66111514568329,
|
| 555 |
+
"logps/rejected": -73.76804026365281,
|
| 556 |
+
"epoch": 0.464,
|
| 557 |
+
"step": 290
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"loss": 0.5151157855987549,
|
| 561 |
+
"grad_norm": 0.18875610828399658,
|
| 562 |
+
"learning_rate": 7.608000000000001e-06,
|
| 563 |
+
"entropy": 0.7552115411614068,
|
| 564 |
+
"num_tokens": 4027530.0,
|
| 565 |
+
"logits/chosen": -1.651739941035044,
|
| 566 |
+
"logits/rejected": -1.2722007430325624,
|
| 567 |
+
"mean_token_accuracy": 0.7032206024974584,
|
| 568 |
+
"rewards/chosen": -1.2901599130476824,
|
| 569 |
+
"rewards/rejected": -2.343320058938116,
|
| 570 |
+
"rewards/accuracies": 0.7625,
|
| 571 |
+
"rewards/margins": 1.0531601494178175,
|
| 572 |
+
"logps/chosen": -65.92262428998947,
|
| 573 |
+
"logps/rejected": -78.2461446762085,
|
| 574 |
+
"epoch": 0.48,
|
| 575 |
+
"step": 300
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"loss": 0.47307710647583007,
|
| 579 |
+
"grad_norm": 0.10863387584686279,
|
| 580 |
+
"learning_rate": 7.528000000000001e-06,
|
| 581 |
+
"entropy": 0.7621748403180391,
|
| 582 |
+
"num_tokens": 4162061.0,
|
| 583 |
+
"logits/chosen": -1.3771063650983688,
|
| 584 |
+
"logits/rejected": -1.2045002001825487,
|
| 585 |
+
"mean_token_accuracy": 0.6764604442752897,
|
| 586 |
+
"rewards/chosen": -1.2184180594878853,
|
| 587 |
+
"rewards/rejected": -2.2775028764503076,
|
| 588 |
+
"rewards/accuracies": 0.7875,
|
| 589 |
+
"rewards/margins": 1.0590848120860756,
|
| 590 |
+
"logps/chosen": -66.88217611312866,
|
| 591 |
+
"logps/rejected": -76.58725030422211,
|
| 592 |
+
"epoch": 0.496,
|
| 593 |
+
"step": 310
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"loss": 0.435822057723999,
|
| 597 |
+
"grad_norm": 0.2315858155488968,
|
| 598 |
+
"learning_rate": 7.4480000000000005e-06,
|
| 599 |
+
"entropy": 0.7862872060039081,
|
| 600 |
+
"num_tokens": 4295767.0,
|
| 601 |
+
"logits/chosen": -1.4690605622435848,
|
| 602 |
+
"logits/rejected": -1.270512638216285,
|
| 603 |
+
"mean_token_accuracy": 0.7008196750655771,
|
| 604 |
+
"rewards/chosen": -1.013645149738295,
|
| 605 |
+
"rewards/rejected": -2.058542217174545,
|
| 606 |
+
"rewards/accuracies": 0.81875,
|
| 607 |
+
"rewards/margins": 1.0448970628902317,
|
| 608 |
+
"logps/chosen": -62.458728992938994,
|
| 609 |
+
"logps/rejected": -71.00610809326172,
|
| 610 |
+
"epoch": 0.512,
|
| 611 |
+
"step": 320
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"loss": 0.5000080585479736,
|
| 615 |
+
"grad_norm": 0.18929436802864075,
|
| 616 |
+
"learning_rate": 7.3680000000000004e-06,
|
| 617 |
+
"entropy": 0.7674915350973606,
|
| 618 |
+
"num_tokens": 4430319.0,
|
| 619 |
+
"logits/chosen": -1.3667780571179007,
|
| 620 |
+
"logits/rejected": -1.1520406470273215,
|
| 621 |
+
"mean_token_accuracy": 0.7238027969375253,
|
| 622 |
+
"rewards/chosen": -1.1733891378389671,
|
| 623 |
+
"rewards/rejected": -2.109529243753059,
|
| 624 |
+
"rewards/accuracies": 0.7625,
|
| 625 |
+
"rewards/margins": 0.9361401099711657,
|
| 626 |
+
"logps/chosen": -65.92721303701401,
|
| 627 |
+
"logps/rejected": -78.01230471134186,
|
| 628 |
+
"epoch": 0.528,
|
| 629 |
+
"step": 330
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"loss": 0.46940155029296876,
|
| 633 |
+
"grad_norm": 0.14189189672470093,
|
| 634 |
+
"learning_rate": 7.288e-06,
|
| 635 |
+
"entropy": 0.7857535354618449,
|
| 636 |
+
"num_tokens": 4565294.0,
|
| 637 |
+
"logits/chosen": -1.4758047716980225,
|
| 638 |
+
"logits/rejected": -1.2134537743928386,
|
| 639 |
+
"mean_token_accuracy": 0.6944835038855672,
|
| 640 |
+
"rewards/chosen": -1.0471536613767967,
|
| 641 |
+
"rewards/rejected": -2.165165621833876,
|
| 642 |
+
"rewards/accuracies": 0.76875,
|
| 643 |
+
"rewards/margins": 1.1180119591765105,
|
| 644 |
+
"logps/chosen": -62.94718644618988,
|
| 645 |
+
"logps/rejected": -79.48942579030991,
|
| 646 |
+
"epoch": 0.544,
|
| 647 |
+
"step": 340
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"loss": 0.3645109415054321,
|
| 651 |
+
"grad_norm": 0.13742217421531677,
|
| 652 |
+
"learning_rate": 7.208e-06,
|
| 653 |
+
"entropy": 0.8591747929778648,
|
| 654 |
+
"num_tokens": 4700673.0,
|
| 655 |
+
"logits/chosen": -1.2994398314971372,
|
| 656 |
+
"logits/rejected": -0.9634321821730494,
|
| 657 |
+
"mean_token_accuracy": 0.6921095721423626,
|
| 658 |
+
"rewards/chosen": -0.9062414692642051,
|
| 659 |
+
"rewards/rejected": -2.211485195462592,
|
| 660 |
+
"rewards/accuracies": 0.875,
|
| 661 |
+
"rewards/margins": 1.3052437256323173,
|
| 662 |
+
"logps/chosen": -65.16439890861511,
|
| 663 |
+
"logps/rejected": -83.85232257843018,
|
| 664 |
+
"epoch": 0.56,
|
| 665 |
+
"step": 350
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"loss": 0.41549930572509763,
|
| 669 |
+
"grad_norm": 0.13630583882331848,
|
| 670 |
+
"learning_rate": 7.128e-06,
|
| 671 |
+
"entropy": 0.7697658954290091,
|
| 672 |
+
"num_tokens": 4834914.0,
|
| 673 |
+
"logits/chosen": -1.4746464063820786,
|
| 674 |
+
"logits/rejected": -1.2137255228371973,
|
| 675 |
+
"mean_token_accuracy": 0.6921661171130836,
|
| 676 |
+
"rewards/chosen": -1.057976609346224,
|
| 677 |
+
"rewards/rejected": -2.2644589847797763,
|
| 678 |
+
"rewards/accuracies": 0.81875,
|
| 679 |
+
"rewards/margins": 1.2064823711523787,
|
| 680 |
+
"logps/chosen": -61.45189489126206,
|
| 681 |
+
"logps/rejected": -74.3291876912117,
|
| 682 |
+
"epoch": 0.576,
|
| 683 |
+
"step": 360
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"loss": 0.5016202449798584,
|
| 687 |
+
"grad_norm": 0.20002736151218414,
|
| 688 |
+
"learning_rate": 7.048e-06,
|
| 689 |
+
"entropy": 0.8047739673376781,
|
| 690 |
+
"num_tokens": 4969481.0,
|
| 691 |
+
"logits/chosen": -1.3463584589105526,
|
| 692 |
+
"logits/rejected": -1.1913337460159867,
|
| 693 |
+
"mean_token_accuracy": 0.6995058806613088,
|
| 694 |
+
"rewards/chosen": -1.2831005254178307,
|
| 695 |
+
"rewards/rejected": -2.3363574695773424,
|
| 696 |
+
"rewards/accuracies": 0.76875,
|
| 697 |
+
"rewards/margins": 1.0532569397240876,
|
| 698 |
+
"logps/chosen": -70.19986630678177,
|
| 699 |
+
"logps/rejected": -73.64605617523193,
|
| 700 |
+
"epoch": 0.592,
|
| 701 |
+
"step": 370
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"loss": 0.4510155200958252,
|
| 705 |
+
"grad_norm": 0.2060878872871399,
|
| 706 |
+
"learning_rate": 6.968e-06,
|
| 707 |
+
"entropy": 0.7111540023062843,
|
| 708 |
+
"num_tokens": 5104134.0,
|
| 709 |
+
"logits/chosen": -1.608769429784676,
|
| 710 |
+
"logits/rejected": -1.205912504773936,
|
| 711 |
+
"mean_token_accuracy": 0.732072526961565,
|
| 712 |
+
"rewards/chosen": -1.1910309330371092,
|
| 713 |
+
"rewards/rejected": -2.485836934146937,
|
| 714 |
+
"rewards/accuracies": 0.7875,
|
| 715 |
+
"rewards/margins": 1.2948059952352196,
|
| 716 |
+
"logps/chosen": -63.09124077558518,
|
| 717 |
+
"logps/rejected": -78.16271492242814,
|
| 718 |
+
"epoch": 0.608,
|
| 719 |
+
"step": 380
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"loss": 0.3789437532424927,
|
| 723 |
+
"grad_norm": 0.16949090361595154,
|
| 724 |
+
"learning_rate": 6.888e-06,
|
| 725 |
+
"entropy": 0.8429804477724246,
|
| 726 |
+
"num_tokens": 5238990.0,
|
| 727 |
+
"logits/chosen": -1.344734235984508,
|
| 728 |
+
"logits/rejected": -1.0792878012791853,
|
| 729 |
+
"mean_token_accuracy": 0.7037996568717062,
|
| 730 |
+
"rewards/chosen": -0.9847486353828572,
|
| 731 |
+
"rewards/rejected": -2.385756150074303,
|
| 732 |
+
"rewards/accuracies": 0.85,
|
| 733 |
+
"rewards/margins": 1.4010075137019158,
|
| 734 |
+
"logps/chosen": -71.131576192379,
|
| 735 |
+
"logps/rejected": -78.69467395544052,
|
| 736 |
+
"epoch": 0.624,
|
| 737 |
+
"step": 390
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"loss": 0.4284682750701904,
|
| 741 |
+
"grad_norm": 0.12163913249969482,
|
| 742 |
+
"learning_rate": 6.808e-06,
|
| 743 |
+
"entropy": 0.8144261223496869,
|
| 744 |
+
"num_tokens": 5373007.0,
|
| 745 |
+
"logits/chosen": -1.5270186574762619,
|
| 746 |
+
"logits/rejected": -1.1437435740491604,
|
| 747 |
+
"mean_token_accuracy": 0.7220658838748932,
|
| 748 |
+
"rewards/chosen": -1.056326786777936,
|
| 749 |
+
"rewards/rejected": -2.289595467923209,
|
| 750 |
+
"rewards/accuracies": 0.81875,
|
| 751 |
+
"rewards/margins": 1.233268681820482,
|
| 752 |
+
"logps/chosen": -62.966822719573976,
|
| 753 |
+
"logps/rejected": -73.83658835887908,
|
| 754 |
+
"epoch": 0.64,
|
| 755 |
+
"step": 400
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"loss": 0.4095573902130127,
|
| 759 |
+
"grad_norm": 0.16153964400291443,
|
| 760 |
+
"learning_rate": 6.728e-06,
|
| 761 |
+
"entropy": 0.8055596031801542,
|
| 762 |
+
"num_tokens": 5507998.0,
|
| 763 |
+
"logits/chosen": -1.3580381465874098,
|
| 764 |
+
"logits/rejected": -1.0995244144140797,
|
| 765 |
+
"mean_token_accuracy": 0.7232280064374208,
|
| 766 |
+
"rewards/chosen": -1.1240590366447578,
|
| 767 |
+
"rewards/rejected": -2.432920588902198,
|
| 768 |
+
"rewards/accuracies": 0.8125,
|
| 769 |
+
"rewards/margins": 1.3088615476619452,
|
| 770 |
+
"logps/chosen": -67.67483884096146,
|
| 771 |
+
"logps/rejected": -79.64460562467575,
|
| 772 |
+
"epoch": 0.656,
|
| 773 |
+
"step": 410
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"loss": 0.4612462520599365,
|
| 777 |
+
"grad_norm": 0.18348106741905212,
|
| 778 |
+
"learning_rate": 6.648e-06,
|
| 779 |
+
"entropy": 0.838780041251448,
|
| 780 |
+
"num_tokens": 5643003.0,
|
| 781 |
+
"logits/chosen": -1.2733183602446794,
|
| 782 |
+
"logits/rejected": -1.0252975904138055,
|
| 783 |
+
"mean_token_accuracy": 0.7087390872649848,
|
| 784 |
+
"rewards/chosen": -1.0622361965593883,
|
| 785 |
+
"rewards/rejected": -2.2466039791703225,
|
| 786 |
+
"rewards/accuracies": 0.79375,
|
| 787 |
+
"rewards/margins": 1.1843677801080048,
|
| 788 |
+
"logps/chosen": -70.36416767835617,
|
| 789 |
+
"logps/rejected": -76.87691588401795,
|
| 790 |
+
"epoch": 0.672,
|
| 791 |
+
"step": 420
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"loss": 0.5008370876312256,
|
| 795 |
+
"grad_norm": 0.3213270604610443,
|
| 796 |
+
"learning_rate": 6.568000000000001e-06,
|
| 797 |
+
"entropy": 0.9088790851383237,
|
| 798 |
+
"num_tokens": 5778181.0,
|
| 799 |
+
"logits/chosen": -1.2890662002693065,
|
| 800 |
+
"logits/rejected": -0.992589689016879,
|
| 801 |
+
"mean_token_accuracy": 0.7122039219364524,
|
| 802 |
+
"rewards/chosen": -0.8109670748352074,
|
| 803 |
+
"rewards/rejected": -1.8928024921566249,
|
| 804 |
+
"rewards/accuracies": 0.79375,
|
| 805 |
+
"rewards/margins": 1.0818354201503098,
|
| 806 |
+
"logps/chosen": -67.84848840236664,
|
| 807 |
+
"logps/rejected": -74.7677659034729,
|
| 808 |
+
"epoch": 0.688,
|
| 809 |
+
"step": 430
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"loss": 0.48210883140563965,
|
| 813 |
+
"grad_norm": 0.17149955034255981,
|
| 814 |
+
"learning_rate": 6.488000000000001e-06,
|
| 815 |
+
"entropy": 0.7832793410867452,
|
| 816 |
+
"num_tokens": 5912330.0,
|
| 817 |
+
"logits/chosen": -1.5082117356293088,
|
| 818 |
+
"logits/rejected": -1.2583976809344715,
|
| 819 |
+
"mean_token_accuracy": 0.7213515549898147,
|
| 820 |
+
"rewards/chosen": -0.8076603990310105,
|
| 821 |
+
"rewards/rejected": -1.880569755542092,
|
| 822 |
+
"rewards/accuracies": 0.76875,
|
| 823 |
+
"rewards/margins": 1.0729093517176806,
|
| 824 |
+
"logps/chosen": -61.79361681938171,
|
| 825 |
+
"logps/rejected": -68.87128767967224,
|
| 826 |
+
"epoch": 0.704,
|
| 827 |
+
"step": 440
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"loss": 0.434552526473999,
|
| 831 |
+
"grad_norm": 0.17293645441532135,
|
| 832 |
+
"learning_rate": 6.408000000000001e-06,
|
| 833 |
+
"entropy": 0.85112169632921,
|
| 834 |
+
"num_tokens": 6046833.0,
|
| 835 |
+
"logits/chosen": -1.4646643974223927,
|
| 836 |
+
"logits/rejected": -1.236012645965301,
|
| 837 |
+
"mean_token_accuracy": 0.6955043694004417,
|
| 838 |
+
"rewards/chosen": -0.6805295475583989,
|
| 839 |
+
"rewards/rejected": -1.8859025583136826,
|
| 840 |
+
"rewards/accuracies": 0.80625,
|
| 841 |
+
"rewards/margins": 1.2053730081766845,
|
| 842 |
+
"logps/chosen": -66.01161818504333,
|
| 843 |
+
"logps/rejected": -72.59794241189957,
|
| 844 |
+
"epoch": 0.72,
|
| 845 |
+
"step": 450
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"loss": 0.4525154590606689,
|
| 849 |
+
"grad_norm": 0.13545100390911102,
|
| 850 |
+
"learning_rate": 6.328000000000001e-06,
|
| 851 |
+
"entropy": 0.7731699298979947,
|
| 852 |
+
"num_tokens": 6180736.0,
|
| 853 |
+
"logits/chosen": -1.5591463232715608,
|
| 854 |
+
"logits/rejected": -1.270020750548368,
|
| 855 |
+
"mean_token_accuracy": 0.7241024266928434,
|
| 856 |
+
"rewards/chosen": -0.6300986663118238,
|
| 857 |
+
"rewards/rejected": -1.714425936579937,
|
| 858 |
+
"rewards/accuracies": 0.775,
|
| 859 |
+
"rewards/margins": 1.0843272741883994,
|
| 860 |
+
"logps/chosen": -56.8658260345459,
|
| 861 |
+
"logps/rejected": -69.70775079727173,
|
| 862 |
+
"epoch": 0.736,
|
| 863 |
+
"step": 460
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"loss": 0.4177194118499756,
|
| 867 |
+
"grad_norm": 0.18396085500717163,
|
| 868 |
+
"learning_rate": 6.248000000000001e-06,
|
| 869 |
+
"entropy": 0.8316918499491294,
|
| 870 |
+
"num_tokens": 6314772.0,
|
| 871 |
+
"logits/chosen": -1.5249799779689304,
|
| 872 |
+
"logits/rejected": -1.1975052540111206,
|
| 873 |
+
"mean_token_accuracy": 0.7292953688651324,
|
| 874 |
+
"rewards/chosen": -0.6027689337788615,
|
| 875 |
+
"rewards/rejected": -1.7004036038677441,
|
| 876 |
+
"rewards/accuracies": 0.80625,
|
| 877 |
+
"rewards/margins": 1.097634667274542,
|
| 878 |
+
"logps/chosen": -56.09835275411606,
|
| 879 |
+
"logps/rejected": -71.92009768486022,
|
| 880 |
+
"epoch": 0.752,
|
| 881 |
+
"step": 470
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"loss": 0.46809401512146,
|
| 885 |
+
"grad_norm": 0.17706090211868286,
|
| 886 |
+
"learning_rate": 6.168000000000001e-06,
|
| 887 |
+
"entropy": 0.7581404208525783,
|
| 888 |
+
"num_tokens": 6449350.0,
|
| 889 |
+
"logits/chosen": -1.6381091898068014,
|
| 890 |
+
"logits/rejected": -1.2200889932680836,
|
| 891 |
+
"mean_token_accuracy": 0.7225047118030489,
|
| 892 |
+
"rewards/chosen": -0.7841349693480879,
|
| 893 |
+
"rewards/rejected": -1.9638352209556615,
|
| 894 |
+
"rewards/accuracies": 0.80625,
|
| 895 |
+
"rewards/margins": 1.1797002504579723,
|
| 896 |
+
"logps/chosen": -57.20818421840668,
|
| 897 |
+
"logps/rejected": -72.92842512130737,
|
| 898 |
+
"epoch": 0.768,
|
| 899 |
+
"step": 480
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"loss": 0.43768625259399413,
|
| 903 |
+
"grad_norm": 0.19130761921405792,
|
| 904 |
+
"learning_rate": 6.088000000000001e-06,
|
| 905 |
+
"entropy": 0.8540503058698959,
|
| 906 |
+
"num_tokens": 6584446.0,
|
| 907 |
+
"logits/chosen": -1.4665205629879048,
|
| 908 |
+
"logits/rejected": -1.209258112589755,
|
| 909 |
+
"mean_token_accuracy": 0.7200203452259302,
|
| 910 |
+
"rewards/chosen": -0.7167962765524862,
|
| 911 |
+
"rewards/rejected": -2.015049184113741,
|
| 912 |
+
"rewards/accuracies": 0.8,
|
| 913 |
+
"rewards/margins": 1.2982529026456178,
|
| 914 |
+
"logps/chosen": -61.549116575717925,
|
| 915 |
+
"logps/rejected": -78.22147595882416,
|
| 916 |
+
"epoch": 0.784,
|
| 917 |
+
"step": 490
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"loss": 0.5317215919494629,
|
| 921 |
+
"grad_norm": 0.23371382057666779,
|
| 922 |
+
"learning_rate": 6.008000000000001e-06,
|
| 923 |
+
"entropy": 0.8205239486735081,
|
| 924 |
+
"num_tokens": 6719638.0,
|
| 925 |
+
"logits/chosen": -1.371530065459643,
|
| 926 |
+
"logits/rejected": -1.2198672509346422,
|
| 927 |
+
"mean_token_accuracy": 0.6944150812923908,
|
| 928 |
+
"rewards/chosen": -0.9274588403815869,
|
| 929 |
+
"rewards/rejected": -1.8801139157498254,
|
| 930 |
+
"rewards/accuracies": 0.75625,
|
| 931 |
+
"rewards/margins": 0.9526550753042102,
|
| 932 |
+
"logps/chosen": -67.30557036399841,
|
| 933 |
+
"logps/rejected": -73.84395973682403,
|
| 934 |
+
"epoch": 0.8,
|
| 935 |
+
"step": 500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"loss": 0.45224776268005373,
|
| 939 |
+
"grad_norm": 0.182713121175766,
|
| 940 |
+
"learning_rate": 5.928000000000001e-06,
|
| 941 |
+
"entropy": 0.8279478559968994,
|
| 942 |
+
"num_tokens": 6853713.0,
|
| 943 |
+
"logits/chosen": -1.3519609814343059,
|
| 944 |
+
"logits/rejected": -1.1373096500825615,
|
| 945 |
+
"mean_token_accuracy": 0.708798126410693,
|
| 946 |
+
"rewards/chosen": -0.9130903312776354,
|
| 947 |
+
"rewards/rejected": -2.1072214540283314,
|
| 948 |
+
"rewards/accuracies": 0.775,
|
| 949 |
+
"rewards/margins": 1.1941311215050519,
|
| 950 |
+
"logps/chosen": -61.76497374773025,
|
| 951 |
+
"logps/rejected": -74.58511118888855,
|
| 952 |
+
"epoch": 0.816,
|
| 953 |
+
"step": 510
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"loss": 0.5070226192474365,
|
| 957 |
+
"grad_norm": 0.21901287138462067,
|
| 958 |
+
"learning_rate": 5.848000000000001e-06,
|
| 959 |
+
"entropy": 0.7715994188318291,
|
| 960 |
+
"num_tokens": 6988253.0,
|
| 961 |
+
"logits/chosen": -1.5578017508136286,
|
| 962 |
+
"logits/rejected": -1.3082164309090911,
|
| 963 |
+
"mean_token_accuracy": 0.7139061275869608,
|
| 964 |
+
"rewards/chosen": -0.9826734508067603,
|
| 965 |
+
"rewards/rejected": -2.0346623631892724,
|
| 966 |
+
"rewards/accuracies": 0.74375,
|
| 967 |
+
"rewards/margins": 1.051988916611299,
|
| 968 |
+
"logps/chosen": -62.744111156463624,
|
| 969 |
+
"logps/rejected": -71.96990052461624,
|
| 970 |
+
"epoch": 0.832,
|
| 971 |
+
"step": 520
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"loss": 0.45290522575378417,
|
| 975 |
+
"grad_norm": 0.3490127921104431,
|
| 976 |
+
"learning_rate": 5.7680000000000005e-06,
|
| 977 |
+
"entropy": 0.795071688198368,
|
| 978 |
+
"num_tokens": 7122655.0,
|
| 979 |
+
"logits/chosen": -1.4304161762086367,
|
| 980 |
+
"logits/rejected": -1.0907285266174562,
|
| 981 |
+
"mean_token_accuracy": 0.7224397015757859,
|
| 982 |
+
"rewards/chosen": -0.9329047386592719,
|
| 983 |
+
"rewards/rejected": -2.097123019141145,
|
| 984 |
+
"rewards/accuracies": 0.76875,
|
| 985 |
+
"rewards/margins": 1.164218282327056,
|
| 986 |
+
"logps/chosen": -55.81759611964226,
|
| 987 |
+
"logps/rejected": -77.36465810537338,
|
| 988 |
+
"epoch": 0.848,
|
| 989 |
+
"step": 530
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"loss": 0.45181903839111326,
|
| 993 |
+
"grad_norm": 0.27503907680511475,
|
| 994 |
+
"learning_rate": 5.6880000000000004e-06,
|
| 995 |
+
"entropy": 0.7538120721874293,
|
| 996 |
+
"num_tokens": 7256116.0,
|
| 997 |
+
"logits/chosen": -1.6767850959126522,
|
| 998 |
+
"logits/rejected": -1.3242235741540545,
|
| 999 |
+
"mean_token_accuracy": 0.7316061129793525,
|
| 1000 |
+
"rewards/chosen": -0.8911430915890378,
|
| 1001 |
+
"rewards/rejected": -2.0349854178726674,
|
| 1002 |
+
"rewards/accuracies": 0.775,
|
| 1003 |
+
"rewards/margins": 1.1438423283398151,
|
| 1004 |
+
"logps/chosen": -53.21874977350235,
|
| 1005 |
+
"logps/rejected": -76.5598596572876,
|
| 1006 |
+
"epoch": 0.864,
|
| 1007 |
+
"step": 540
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"loss": 0.5414995670318603,
|
| 1011 |
+
"grad_norm": 0.2073371410369873,
|
| 1012 |
+
"learning_rate": 5.608e-06,
|
| 1013 |
+
"entropy": 0.8148633663950022,
|
| 1014 |
+
"num_tokens": 7390370.0,
|
| 1015 |
+
"logits/chosen": -1.3765310828876476,
|
| 1016 |
+
"logits/rejected": -1.133697184223435,
|
| 1017 |
+
"mean_token_accuracy": 0.7053394909948111,
|
| 1018 |
+
"rewards/chosen": -0.90475755288644,
|
| 1019 |
+
"rewards/rejected": -1.7751331690698862,
|
| 1020 |
+
"rewards/accuracies": 0.74375,
|
| 1021 |
+
"rewards/margins": 0.8703756197355688,
|
| 1022 |
+
"logps/chosen": -61.96280990839004,
|
| 1023 |
+
"logps/rejected": -70.7764186501503,
|
| 1024 |
+
"epoch": 0.88,
|
| 1025 |
+
"step": 550
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"loss": 0.5176785945892334,
|
| 1029 |
+
"grad_norm": 0.27757805585861206,
|
| 1030 |
+
"learning_rate": 5.528e-06,
|
| 1031 |
+
"entropy": 0.7701463760924525,
|
| 1032 |
+
"num_tokens": 7524399.0,
|
| 1033 |
+
"logits/chosen": -1.4427345128285418,
|
| 1034 |
+
"logits/rejected": -1.2567077535753297,
|
| 1035 |
+
"mean_token_accuracy": 0.7136867575347423,
|
| 1036 |
+
"rewards/chosen": -0.9190105670393678,
|
| 1037 |
+
"rewards/rejected": -1.8803606960456818,
|
| 1038 |
+
"rewards/accuracies": 0.76875,
|
| 1039 |
+
"rewards/margins": 0.9613501313142478,
|
| 1040 |
+
"logps/chosen": -62.47981760501862,
|
| 1041 |
+
"logps/rejected": -65.90649355649948,
|
| 1042 |
+
"epoch": 0.896,
|
| 1043 |
+
"step": 560
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"loss": 0.4511863708496094,
|
| 1047 |
+
"grad_norm": 0.2078111171722412,
|
| 1048 |
+
"learning_rate": 5.448e-06,
|
| 1049 |
+
"entropy": 0.8631977764249313,
|
| 1050 |
+
"num_tokens": 7658604.0,
|
| 1051 |
+
"logits/chosen": -1.2926124538881012,
|
| 1052 |
+
"logits/rejected": -0.9838834319481562,
|
| 1053 |
+
"mean_token_accuracy": 0.7207308521494269,
|
| 1054 |
+
"rewards/chosen": -0.6676558028753788,
|
| 1055 |
+
"rewards/rejected": -1.7714358688157517,
|
| 1056 |
+
"rewards/accuracies": 0.83125,
|
| 1057 |
+
"rewards/margins": 1.1037800651043654,
|
| 1058 |
+
"logps/chosen": -56.60341747999191,
|
| 1059 |
+
"logps/rejected": -71.80331914424896,
|
| 1060 |
+
"epoch": 0.912,
|
| 1061 |
+
"step": 570
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"loss": 0.4576895713806152,
|
| 1065 |
+
"grad_norm": 0.13427318632602692,
|
| 1066 |
+
"learning_rate": 5.368000000000001e-06,
|
| 1067 |
+
"entropy": 0.8798671936965548,
|
| 1068 |
+
"num_tokens": 7793743.0,
|
| 1069 |
+
"logits/chosen": -1.2397059499933185,
|
| 1070 |
+
"logits/rejected": -0.9909034700496523,
|
| 1071 |
+
"mean_token_accuracy": 0.6908573735505342,
|
| 1072 |
+
"rewards/chosen": -0.7387171853566542,
|
| 1073 |
+
"rewards/rejected": -1.864225831045769,
|
| 1074 |
+
"rewards/accuracies": 0.80625,
|
| 1075 |
+
"rewards/margins": 1.1255086408928037,
|
| 1076 |
+
"logps/chosen": -64.580437374115,
|
| 1077 |
+
"logps/rejected": -73.18329973220825,
|
| 1078 |
+
"epoch": 0.928,
|
| 1079 |
+
"step": 580
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"loss": 0.42533297538757325,
|
| 1083 |
+
"grad_norm": 0.15502135455608368,
|
| 1084 |
+
"learning_rate": 5.288000000000001e-06,
|
| 1085 |
+
"entropy": 0.817069728297065,
|
| 1086 |
+
"num_tokens": 7928689.0,
|
| 1087 |
+
"logits/chosen": -1.2393895301783906,
|
| 1088 |
+
"logits/rejected": -1.1084123908454555,
|
| 1089 |
+
"mean_token_accuracy": 0.7128896466456354,
|
| 1090 |
+
"rewards/chosen": -0.7250971098641458,
|
| 1091 |
+
"rewards/rejected": -1.9982559287804178,
|
| 1092 |
+
"rewards/accuracies": 0.84375,
|
| 1093 |
+
"rewards/margins": 1.2731588228605688,
|
| 1094 |
+
"logps/chosen": -65.04422287940979,
|
| 1095 |
+
"logps/rejected": -75.35328090190887,
|
| 1096 |
+
"epoch": 0.944,
|
| 1097 |
+
"step": 590
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"loss": 0.3770539999008179,
|
| 1101 |
+
"grad_norm": 0.17613562941551208,
|
| 1102 |
+
"learning_rate": 5.208000000000001e-06,
|
| 1103 |
+
"entropy": 0.7778703475603834,
|
| 1104 |
+
"num_tokens": 8062710.0,
|
| 1105 |
+
"logits/chosen": -1.5389842542461873,
|
| 1106 |
+
"logits/rejected": -1.0342297198537742,
|
| 1107 |
+
"mean_token_accuracy": 0.7274640101008117,
|
| 1108 |
+
"rewards/chosen": -0.7972871424448386,
|
| 1109 |
+
"rewards/rejected": -2.1182173368986694,
|
| 1110 |
+
"rewards/accuracies": 0.84375,
|
| 1111 |
+
"rewards/margins": 1.320930197648704,
|
| 1112 |
+
"logps/chosen": -53.877902281284335,
|
| 1113 |
+
"logps/rejected": -77.50610139369965,
|
| 1114 |
+
"epoch": 0.96,
|
| 1115 |
+
"step": 600
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"loss": 0.4678995132446289,
|
| 1119 |
+
"grad_norm": 0.21209311485290527,
|
| 1120 |
+
"learning_rate": 5.128000000000001e-06,
|
| 1121 |
+
"entropy": 0.8264809994332609,
|
| 1122 |
+
"num_tokens": 8197009.0,
|
| 1123 |
+
"logits/chosen": -1.3161340033706996,
|
| 1124 |
+
"logits/rejected": -1.0665738771389415,
|
| 1125 |
+
"mean_token_accuracy": 0.6980931758880615,
|
| 1126 |
+
"rewards/chosen": -0.7594082839557814,
|
| 1127 |
+
"rewards/rejected": -1.9171317781874677,
|
| 1128 |
+
"rewards/accuracies": 0.8125,
|
| 1129 |
+
"rewards/margins": 1.157723485864699,
|
| 1130 |
+
"logps/chosen": -61.800520193576816,
|
| 1131 |
+
"logps/rejected": -72.15935192108154,
|
| 1132 |
+
"epoch": 0.976,
|
| 1133 |
+
"step": 610
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"loss": 0.4798553466796875,
|
| 1137 |
+
"grad_norm": 0.24422219395637512,
|
| 1138 |
+
"learning_rate": 5.048000000000001e-06,
|
| 1139 |
+
"entropy": 0.7734213294519577,
|
| 1140 |
+
"num_tokens": 8330156.0,
|
| 1141 |
+
"logits/chosen": -1.5627198204058765,
|
| 1142 |
+
"logits/rejected": -1.3033627091925726,
|
| 1143 |
+
"mean_token_accuracy": 0.7267525289207697,
|
| 1144 |
+
"rewards/chosen": -0.7573755404679104,
|
| 1145 |
+
"rewards/rejected": -1.7975832854863256,
|
| 1146 |
+
"rewards/accuracies": 0.79375,
|
| 1147 |
+
"rewards/margins": 1.040207749651745,
|
| 1148 |
+
"logps/chosen": -55.06023645401001,
|
| 1149 |
+
"logps/rejected": -68.08906935453415,
|
| 1150 |
+
"epoch": 0.992,
|
| 1151 |
+
"step": 620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"loss": 0.41657228469848634,
|
| 1155 |
+
"grad_norm": 0.2854296565055847,
|
| 1156 |
+
"learning_rate": 4.9680000000000005e-06,
|
| 1157 |
+
"entropy": 0.7377176167094148,
|
| 1158 |
+
"num_tokens": 8465041.0,
|
| 1159 |
+
"logits/chosen": -1.6654656300451234,
|
| 1160 |
+
"logits/rejected": -1.3485704232572624,
|
| 1161 |
+
"mean_token_accuracy": 0.6977982034906745,
|
| 1162 |
+
"rewards/chosen": -0.827795197776868,
|
| 1163 |
+
"rewards/rejected": -2.1947057927493008,
|
| 1164 |
+
"rewards/accuracies": 0.825,
|
| 1165 |
+
"rewards/margins": 1.3669105987995862,
|
| 1166 |
+
"logps/chosen": -59.07929306030273,
|
| 1167 |
+
"logps/rejected": -76.1623566031456,
|
| 1168 |
+
"epoch": 1.008,
|
| 1169 |
+
"step": 630
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"loss": 0.38203165531158445,
|
| 1173 |
+
"grad_norm": 0.18717509508132935,
|
| 1174 |
+
"learning_rate": 4.8880000000000005e-06,
|
| 1175 |
+
"entropy": 0.8211721338244388,
|
| 1176 |
+
"num_tokens": 8599500.0,
|
| 1177 |
+
"logits/chosen": -1.5114663958371213,
|
| 1178 |
+
"logits/rejected": -1.193993404852716,
|
| 1179 |
+
"mean_token_accuracy": 0.7073404714465141,
|
| 1180 |
+
"rewards/chosen": -0.823921048961347,
|
| 1181 |
+
"rewards/rejected": -2.257023681770079,
|
| 1182 |
+
"rewards/accuracies": 0.79375,
|
| 1183 |
+
"rewards/margins": 1.4331026304280385,
|
| 1184 |
+
"logps/chosen": -59.9299793958664,
|
| 1185 |
+
"logps/rejected": -78.36245132684708,
|
| 1186 |
+
"epoch": 1.024,
|
| 1187 |
+
"step": 640
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"loss": 0.4451735019683838,
|
| 1191 |
+
"grad_norm": 0.21657153964042664,
|
| 1192 |
+
"learning_rate": 4.808e-06,
|
| 1193 |
+
"entropy": 0.8258791111322352,
|
| 1194 |
+
"num_tokens": 8733676.0,
|
| 1195 |
+
"logits/chosen": -1.5098714195765397,
|
| 1196 |
+
"logits/rejected": -1.1212204464595934,
|
| 1197 |
+
"mean_token_accuracy": 0.7025207518599927,
|
| 1198 |
+
"rewards/chosen": -0.8813661304506241,
|
| 1199 |
+
"rewards/rejected": -2.205052966135554,
|
| 1200 |
+
"rewards/accuracies": 0.76875,
|
| 1201 |
+
"rewards/margins": 1.3236868323292583,
|
| 1202 |
+
"logps/chosen": -59.143254685401914,
|
| 1203 |
+
"logps/rejected": -79.33009437322616,
|
| 1204 |
+
"epoch": 1.04,
|
| 1205 |
+
"step": 650
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"loss": 0.44518437385559084,
|
| 1209 |
+
"grad_norm": 0.16692371666431427,
|
| 1210 |
+
"learning_rate": 4.728e-06,
|
| 1211 |
+
"entropy": 0.7720623685941973,
|
| 1212 |
+
"num_tokens": 8866920.0,
|
| 1213 |
+
"logits/chosen": -1.6099649834733871,
|
| 1214 |
+
"logits/rejected": -1.2753671202879713,
|
| 1215 |
+
"mean_token_accuracy": 0.7019124954938889,
|
| 1216 |
+
"rewards/chosen": -0.9054035557142924,
|
| 1217 |
+
"rewards/rejected": -2.102139440679457,
|
| 1218 |
+
"rewards/accuracies": 0.81875,
|
| 1219 |
+
"rewards/margins": 1.1967358850408345,
|
| 1220 |
+
"logps/chosen": -56.078457951545715,
|
| 1221 |
+
"logps/rejected": -74.86194944381714,
|
| 1222 |
+
"epoch": 1.056,
|
| 1223 |
+
"step": 660
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"loss": 0.46517186164855956,
|
| 1227 |
+
"grad_norm": 0.3806668817996979,
|
| 1228 |
+
"learning_rate": 4.648e-06,
|
| 1229 |
+
"entropy": 0.9248056679964065,
|
| 1230 |
+
"num_tokens": 9002809.0,
|
| 1231 |
+
"logits/chosen": -1.4305434844414238,
|
| 1232 |
+
"logits/rejected": -1.1489092484117014,
|
| 1233 |
+
"mean_token_accuracy": 0.6812765257433057,
|
| 1234 |
+
"rewards/chosen": -1.0725531240459532,
|
| 1235 |
+
"rewards/rejected": -2.4423089541844094,
|
| 1236 |
+
"rewards/accuracies": 0.8125,
|
| 1237 |
+
"rewards/margins": 1.3697558240965009,
|
| 1238 |
+
"logps/chosen": -74.38465806245804,
|
| 1239 |
+
"logps/rejected": -88.47626183032989,
|
| 1240 |
+
"epoch": 1.072,
|
| 1241 |
+
"step": 670
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"loss": 0.41553492546081544,
|
| 1245 |
+
"grad_norm": 0.22808755934238434,
|
| 1246 |
+
"learning_rate": 4.568e-06,
|
| 1247 |
+
"entropy": 0.8447681298654061,
|
| 1248 |
+
"num_tokens": 9138006.0,
|
| 1249 |
+
"logits/chosen": -1.591791772365951,
|
| 1250 |
+
"logits/rejected": -1.146695606775978,
|
| 1251 |
+
"mean_token_accuracy": 0.7053618848323822,
|
| 1252 |
+
"rewards/chosen": -1.0293371758598369,
|
| 1253 |
+
"rewards/rejected": -2.4234205832937734,
|
| 1254 |
+
"rewards/accuracies": 0.7875,
|
| 1255 |
+
"rewards/margins": 1.394083405379206,
|
| 1256 |
+
"logps/chosen": -60.523981761932376,
|
| 1257 |
+
"logps/rejected": -88.66759954690933,
|
| 1258 |
+
"epoch": 1.088,
|
| 1259 |
+
"step": 680
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"loss": 0.3357177972793579,
|
| 1263 |
+
"grad_norm": 0.2115124762058258,
|
| 1264 |
+
"learning_rate": 4.488e-06,
|
| 1265 |
+
"entropy": 0.77818886981986,
|
| 1266 |
+
"num_tokens": 9272926.0,
|
| 1267 |
+
"logits/chosen": -1.5370966699642798,
|
| 1268 |
+
"logits/rejected": -1.1325565169583864,
|
| 1269 |
+
"mean_token_accuracy": 0.7299308383837342,
|
| 1270 |
+
"rewards/chosen": -0.9643074480060022,
|
| 1271 |
+
"rewards/rejected": -2.687541061779484,
|
| 1272 |
+
"rewards/accuracies": 0.86875,
|
| 1273 |
+
"rewards/margins": 1.7232336099725216,
|
| 1274 |
+
"logps/chosen": -59.027079343795776,
|
| 1275 |
+
"logps/rejected": -84.05350723266602,
|
| 1276 |
+
"epoch": 1.104,
|
| 1277 |
+
"step": 690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"loss": 0.4211433410644531,
|
| 1281 |
+
"grad_norm": 0.23426513373851776,
|
| 1282 |
+
"learning_rate": 4.408000000000001e-06,
|
| 1283 |
+
"entropy": 0.7870391614065738,
|
| 1284 |
+
"num_tokens": 9407243.0,
|
| 1285 |
+
"logits/chosen": -1.6077830379249172,
|
| 1286 |
+
"logits/rejected": -1.2221489033385453,
|
| 1287 |
+
"mean_token_accuracy": 0.7154217397794127,
|
| 1288 |
+
"rewards/chosen": -1.0969871749985032,
|
| 1289 |
+
"rewards/rejected": -2.416306406632066,
|
| 1290 |
+
"rewards/accuracies": 0.825,
|
| 1291 |
+
"rewards/margins": 1.3193192333448678,
|
| 1292 |
+
"logps/chosen": -66.31653872728347,
|
| 1293 |
+
"logps/rejected": -74.53792465925217,
|
| 1294 |
+
"epoch": 1.12,
|
| 1295 |
+
"step": 700
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"loss": 0.35266427993774413,
|
| 1299 |
+
"grad_norm": 0.16782480478286743,
|
| 1300 |
+
"learning_rate": 4.328000000000001e-06,
|
| 1301 |
+
"entropy": 0.7525284075178206,
|
| 1302 |
+
"num_tokens": 9541881.0,
|
| 1303 |
+
"logits/chosen": -1.7815538050430466,
|
| 1304 |
+
"logits/rejected": -1.2432088022449368,
|
| 1305 |
+
"mean_token_accuracy": 0.7324884554371238,
|
| 1306 |
+
"rewards/chosen": -1.170151641382836,
|
| 1307 |
+
"rewards/rejected": -2.8821156861260535,
|
| 1308 |
+
"rewards/accuracies": 0.85625,
|
| 1309 |
+
"rewards/margins": 1.7119640467688442,
|
| 1310 |
+
"logps/chosen": -62.426369762420656,
|
| 1311 |
+
"logps/rejected": -82.97551012039185,
|
| 1312 |
+
"epoch": 1.1360000000000001,
|
| 1313 |
+
"step": 710
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"loss": 0.4414065361022949,
|
| 1317 |
+
"grad_norm": 0.17724663019180298,
|
| 1318 |
+
"learning_rate": 4.248000000000001e-06,
|
| 1319 |
+
"entropy": 0.7874836204689928,
|
| 1320 |
+
"num_tokens": 9676124.0,
|
| 1321 |
+
"logits/chosen": -1.5870644241142104,
|
| 1322 |
+
"logits/rejected": -1.2676512314420252,
|
| 1323 |
+
"mean_token_accuracy": 0.7085109381005168,
|
| 1324 |
+
"rewards/chosen": -1.1689644593454431,
|
| 1325 |
+
"rewards/rejected": -2.5695454872446133,
|
| 1326 |
+
"rewards/accuracies": 0.80625,
|
| 1327 |
+
"rewards/margins": 1.400581027008593,
|
| 1328 |
+
"logps/chosen": -64.36232153177261,
|
| 1329 |
+
"logps/rejected": -79.14827445745468,
|
| 1330 |
+
"epoch": 1.152,
|
| 1331 |
+
"step": 720
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"loss": 0.42041378021240233,
|
| 1335 |
+
"grad_norm": 0.15182077884674072,
|
| 1336 |
+
"learning_rate": 4.168000000000001e-06,
|
| 1337 |
+
"entropy": 0.7948870263673598,
|
| 1338 |
+
"num_tokens": 9809272.0,
|
| 1339 |
+
"logits/chosen": -1.5996510929231569,
|
| 1340 |
+
"logits/rejected": -1.2198636359873567,
|
| 1341 |
+
"mean_token_accuracy": 0.6895976271480322,
|
| 1342 |
+
"rewards/chosen": -0.9217694426333765,
|
| 1343 |
+
"rewards/rejected": -2.2751815680006984,
|
| 1344 |
+
"rewards/accuracies": 0.8375,
|
| 1345 |
+
"rewards/margins": 1.3534121236763894,
|
| 1346 |
+
"logps/chosen": -63.298885464668274,
|
| 1347 |
+
"logps/rejected": -69.15098886489868,
|
| 1348 |
+
"epoch": 1.168,
|
| 1349 |
+
"step": 730
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"loss": 0.4149898052215576,
|
| 1353 |
+
"grad_norm": 0.4299209713935852,
|
| 1354 |
+
"learning_rate": 4.0880000000000005e-06,
|
| 1355 |
+
"entropy": 0.8794720381731167,
|
| 1356 |
+
"num_tokens": 9944472.0,
|
| 1357 |
+
"logits/chosen": -1.3574370869924466,
|
| 1358 |
+
"logits/rejected": -1.0639106517819426,
|
| 1359 |
+
"mean_token_accuracy": 0.6969729635864497,
|
| 1360 |
+
"rewards/chosen": -0.867475303565152,
|
| 1361 |
+
"rewards/rejected": -2.344749124685768,
|
| 1362 |
+
"rewards/accuracies": 0.85,
|
| 1363 |
+
"rewards/margins": 1.4772738242521881,
|
| 1364 |
+
"logps/chosen": -65.34813396930694,
|
| 1365 |
+
"logps/rejected": -80.18645774126053,
|
| 1366 |
+
"epoch": 1.184,
|
| 1367 |
+
"step": 740
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"loss": 0.3969083309173584,
|
| 1371 |
+
"grad_norm": 0.16381904482841492,
|
| 1372 |
+
"learning_rate": 4.008e-06,
|
| 1373 |
+
"entropy": 0.8126660047855694,
|
| 1374 |
+
"num_tokens": 10078733.0,
|
| 1375 |
+
"logits/chosen": -1.3498129883483427,
|
| 1376 |
+
"logits/rejected": -1.1121957037881862,
|
| 1377 |
+
"mean_token_accuracy": 0.7128386048600077,
|
| 1378 |
+
"rewards/chosen": -0.8270440307072932,
|
| 1379 |
+
"rewards/rejected": -2.320264756004326,
|
| 1380 |
+
"rewards/accuracies": 0.81875,
|
| 1381 |
+
"rewards/margins": 1.4932207340840251,
|
| 1382 |
+
"logps/chosen": -58.0639621257782,
|
| 1383 |
+
"logps/rejected": -74.64224022626877,
|
| 1384 |
+
"epoch": 1.2,
|
| 1385 |
+
"step": 750
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"loss": 0.3923245668411255,
|
| 1389 |
+
"grad_norm": 0.09231153130531311,
|
| 1390 |
+
"learning_rate": 3.928e-06,
|
| 1391 |
+
"entropy": 0.7907555149227846,
|
| 1392 |
+
"num_tokens": 10212826.0,
|
| 1393 |
+
"logits/chosen": -1.5573132545464914,
|
| 1394 |
+
"logits/rejected": -1.2641684294564186,
|
| 1395 |
+
"mean_token_accuracy": 0.7216938810423017,
|
| 1396 |
+
"rewards/chosen": -0.9051268995186547,
|
| 1397 |
+
"rewards/rejected": -2.3114789802231828,
|
| 1398 |
+
"rewards/accuracies": 0.85625,
|
| 1399 |
+
"rewards/margins": 1.406352074991446,
|
| 1400 |
+
"logps/chosen": -58.32571488022804,
|
| 1401 |
+
"logps/rejected": -83.86995618343353,
|
| 1402 |
+
"epoch": 1.216,
|
| 1403 |
+
"step": 760
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"loss": 0.35955214500427246,
|
| 1407 |
+
"grad_norm": 0.16152027249336243,
|
| 1408 |
+
"learning_rate": 3.848e-06,
|
| 1409 |
+
"entropy": 0.7906852045387496,
|
| 1410 |
+
"num_tokens": 10346556.0,
|
| 1411 |
+
"logits/chosen": -1.5494209585029366,
|
| 1412 |
+
"logits/rejected": -1.2729574881882504,
|
| 1413 |
+
"mean_token_accuracy": 0.7100083495490253,
|
| 1414 |
+
"rewards/chosen": -0.8209864257834851,
|
| 1415 |
+
"rewards/rejected": -2.317021907842718,
|
| 1416 |
+
"rewards/accuracies": 0.88125,
|
| 1417 |
+
"rewards/margins": 1.4960354787297547,
|
| 1418 |
+
"logps/chosen": -58.37631057500839,
|
| 1419 |
+
"logps/rejected": -75.28673508167267,
|
| 1420 |
+
"epoch": 1.232,
|
| 1421 |
+
"step": 770
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"loss": 0.4129960060119629,
|
| 1425 |
+
"grad_norm": 0.24933257699012756,
|
| 1426 |
+
"learning_rate": 3.7680000000000006e-06,
|
| 1427 |
+
"entropy": 0.75614075922349,
|
| 1428 |
+
"num_tokens": 10480829.0,
|
| 1429 |
+
"logits/chosen": -1.6691249020660657,
|
| 1430 |
+
"logits/rejected": -1.218453002782342,
|
| 1431 |
+
"mean_token_accuracy": 0.725784600712359,
|
| 1432 |
+
"rewards/chosen": -1.1460461351904088,
|
| 1433 |
+
"rewards/rejected": -2.7181562868412583,
|
| 1434 |
+
"rewards/accuracies": 0.81875,
|
| 1435 |
+
"rewards/margins": 1.5721101415809244,
|
| 1436 |
+
"logps/chosen": -62.6519087433815,
|
| 1437 |
+
"logps/rejected": -81.97410740852357,
|
| 1438 |
+
"epoch": 1.248,
|
| 1439 |
+
"step": 780
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"loss": 0.4455737113952637,
|
| 1443 |
+
"grad_norm": 0.3030645549297333,
|
| 1444 |
+
"learning_rate": 3.6880000000000005e-06,
|
| 1445 |
+
"entropy": 0.7621263024513609,
|
| 1446 |
+
"num_tokens": 10614414.0,
|
| 1447 |
+
"logits/chosen": -1.700109034199165,
|
| 1448 |
+
"logits/rejected": -1.376628724655141,
|
| 1449 |
+
"mean_token_accuracy": 0.7104939077049494,
|
| 1450 |
+
"rewards/chosen": -1.0354830527561716,
|
| 1451 |
+
"rewards/rejected": -2.3573520673438906,
|
| 1452 |
+
"rewards/accuracies": 0.81875,
|
| 1453 |
+
"rewards/margins": 1.3218690163921565,
|
| 1454 |
+
"logps/chosen": -59.53394418954849,
|
| 1455 |
+
"logps/rejected": -75.5838432431221,
|
| 1456 |
+
"epoch": 1.264,
|
| 1457 |
+
"step": 790
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"loss": 0.4144607067108154,
|
| 1461 |
+
"grad_norm": 0.3101625144481659,
|
| 1462 |
+
"learning_rate": 3.6080000000000004e-06,
|
| 1463 |
+
"entropy": 0.7888491688296199,
|
| 1464 |
+
"num_tokens": 10748424.0,
|
| 1465 |
+
"logits/chosen": -1.744755668223603,
|
| 1466 |
+
"logits/rejected": -1.2454132069381836,
|
| 1467 |
+
"mean_token_accuracy": 0.7081136297434568,
|
| 1468 |
+
"rewards/chosen": -1.042457212501904,
|
| 1469 |
+
"rewards/rejected": -2.5617473446996883,
|
| 1470 |
+
"rewards/accuracies": 0.8125,
|
| 1471 |
+
"rewards/margins": 1.519290132354945,
|
| 1472 |
+
"logps/chosen": -61.47326543331146,
|
| 1473 |
+
"logps/rejected": -79.36946972608567,
|
| 1474 |
+
"epoch": 1.28,
|
| 1475 |
+
"step": 800
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"loss": 0.36916818618774416,
|
| 1479 |
+
"grad_norm": 0.18668942153453827,
|
| 1480 |
+
"learning_rate": 3.5280000000000004e-06,
|
| 1481 |
+
"entropy": 0.8457717320125084,
|
| 1482 |
+
"num_tokens": 10884117.0,
|
| 1483 |
+
"logits/chosen": -1.4954581386967238,
|
| 1484 |
+
"logits/rejected": -1.1664500756033909,
|
| 1485 |
+
"mean_token_accuracy": 0.7208706248551607,
|
| 1486 |
+
"rewards/chosen": -0.9436249577498529,
|
| 1487 |
+
"rewards/rejected": -2.5245022784452886,
|
| 1488 |
+
"rewards/accuracies": 0.85,
|
| 1489 |
+
"rewards/margins": 1.5808773178141564,
|
| 1490 |
+
"logps/chosen": -70.7584779381752,
|
| 1491 |
+
"logps/rejected": -82.38645179271698,
|
| 1492 |
+
"epoch": 1.296,
|
| 1493 |
+
"step": 810
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"loss": 0.41336798667907715,
|
| 1497 |
+
"grad_norm": 0.24846971035003662,
|
| 1498 |
+
"learning_rate": 3.4480000000000003e-06,
|
| 1499 |
+
"entropy": 0.7393746774076135,
|
| 1500 |
+
"num_tokens": 11018693.0,
|
| 1501 |
+
"logits/chosen": -1.5728018848198846,
|
| 1502 |
+
"logits/rejected": -1.3047334613617754,
|
| 1503 |
+
"mean_token_accuracy": 0.7239607466384769,
|
| 1504 |
+
"rewards/chosen": -0.9437924515921623,
|
| 1505 |
+
"rewards/rejected": -2.4474497922346927,
|
| 1506 |
+
"rewards/accuracies": 0.80625,
|
| 1507 |
+
"rewards/margins": 1.5036573376040905,
|
| 1508 |
+
"logps/chosen": -58.04731862545013,
|
| 1509 |
+
"logps/rejected": -79.88022357225418,
|
| 1510 |
+
"epoch": 1.312,
|
| 1511 |
+
"step": 820
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"loss": 0.3941138982772827,
|
| 1515 |
+
"grad_norm": 0.17962393164634705,
|
| 1516 |
+
"learning_rate": 3.368e-06,
|
| 1517 |
+
"entropy": 0.8191202863235958,
|
| 1518 |
+
"num_tokens": 11153738.0,
|
| 1519 |
+
"logits/chosen": -1.569502977982598,
|
| 1520 |
+
"logits/rejected": -1.3446180097251133,
|
| 1521 |
+
"mean_token_accuracy": 0.7013885753229261,
|
| 1522 |
+
"rewards/chosen": -1.0316214625607245,
|
| 1523 |
+
"rewards/rejected": -2.4835243557114155,
|
| 1524 |
+
"rewards/accuracies": 0.83125,
|
| 1525 |
+
"rewards/margins": 1.4519028885290026,
|
| 1526 |
+
"logps/chosen": -66.87270923852921,
|
| 1527 |
+
"logps/rejected": -82.45893884897232,
|
| 1528 |
+
"epoch": 1.328,
|
| 1529 |
+
"step": 830
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"loss": 0.40454444885253904,
|
| 1533 |
+
"grad_norm": 0.26399481296539307,
|
| 1534 |
+
"learning_rate": 3.288e-06,
|
| 1535 |
+
"entropy": 0.8281162801082246,
|
| 1536 |
+
"num_tokens": 11288148.0,
|
| 1537 |
+
"logits/chosen": -1.393732705892854,
|
| 1538 |
+
"logits/rejected": -1.134675861717077,
|
| 1539 |
+
"mean_token_accuracy": 0.711259607039392,
|
| 1540 |
+
"rewards/chosen": -0.8508172139419912,
|
| 1541 |
+
"rewards/rejected": -2.2421938594488893,
|
| 1542 |
+
"rewards/accuracies": 0.7875,
|
| 1543 |
+
"rewards/margins": 1.391376649821177,
|
| 1544 |
+
"logps/chosen": -61.118980848789214,
|
| 1545 |
+
"logps/rejected": -75.8330344080925,
|
| 1546 |
+
"epoch": 1.3439999999999999,
|
| 1547 |
+
"step": 840
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"loss": 0.3927537202835083,
|
| 1551 |
+
"grad_norm": 0.2384471893310547,
|
| 1552 |
+
"learning_rate": 3.208e-06,
|
| 1553 |
+
"entropy": 0.7859910909202881,
|
| 1554 |
+
"num_tokens": 11422122.0,
|
| 1555 |
+
"logits/chosen": -1.559968605653432,
|
| 1556 |
+
"logits/rejected": -1.2226467421665006,
|
| 1557 |
+
"mean_token_accuracy": 0.705523016769439,
|
| 1558 |
+
"rewards/chosen": -0.8587907724882825,
|
| 1559 |
+
"rewards/rejected": -2.3526173725025727,
|
| 1560 |
+
"rewards/accuracies": 0.86875,
|
| 1561 |
+
"rewards/margins": 1.4938265983946621,
|
| 1562 |
+
"logps/chosen": -58.47013804912567,
|
| 1563 |
+
"logps/rejected": -76.32856943607331,
|
| 1564 |
+
"epoch": 1.3599999999999999,
|
| 1565 |
+
"step": 850
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"loss": 0.4346799850463867,
|
| 1569 |
+
"grad_norm": 0.2608790099620819,
|
| 1570 |
+
"learning_rate": 3.1280000000000004e-06,
|
| 1571 |
+
"entropy": 0.7820992226828821,
|
| 1572 |
+
"num_tokens": 11555439.0,
|
| 1573 |
+
"logits/chosen": -1.5948384898114043,
|
| 1574 |
+
"logits/rejected": -1.2975419307108174,
|
| 1575 |
+
"mean_token_accuracy": 0.694941917154938,
|
| 1576 |
+
"rewards/chosen": -0.9103366011011531,
|
| 1577 |
+
"rewards/rejected": -2.206452490389347,
|
| 1578 |
+
"rewards/accuracies": 0.81875,
|
| 1579 |
+
"rewards/margins": 1.2961158849298955,
|
| 1580 |
+
"logps/chosen": -60.45212324857712,
|
| 1581 |
+
"logps/rejected": -70.58573877811432,
|
| 1582 |
+
"epoch": 1.376,
|
| 1583 |
+
"step": 860
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"loss": 0.47821383476257323,
|
| 1587 |
+
"grad_norm": 0.33640041947364807,
|
| 1588 |
+
"learning_rate": 3.0480000000000003e-06,
|
| 1589 |
+
"entropy": 0.7687381034600549,
|
| 1590 |
+
"num_tokens": 11689826.0,
|
| 1591 |
+
"logits/chosen": -1.7085993267610973,
|
| 1592 |
+
"logits/rejected": -1.386772186435975,
|
| 1593 |
+
"mean_token_accuracy": 0.7053665170446038,
|
| 1594 |
+
"rewards/chosen": -0.9841044571541715,
|
| 1595 |
+
"rewards/rejected": -2.290806566970423,
|
| 1596 |
+
"rewards/accuracies": 0.74375,
|
| 1597 |
+
"rewards/margins": 1.306702113803476,
|
| 1598 |
+
"logps/chosen": -61.73239058256149,
|
| 1599 |
+
"logps/rejected": -75.85606197118759,
|
| 1600 |
+
"epoch": 1.392,
|
| 1601 |
+
"step": 870
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"loss": 0.44809565544128416,
|
| 1605 |
+
"grad_norm": 0.19289684295654297,
|
| 1606 |
+
"learning_rate": 2.9680000000000002e-06,
|
| 1607 |
+
"entropy": 0.838317794475006,
|
| 1608 |
+
"num_tokens": 11824262.0,
|
| 1609 |
+
"logits/chosen": -1.415884176472813,
|
| 1610 |
+
"logits/rejected": -1.2410310266034903,
|
| 1611 |
+
"mean_token_accuracy": 0.6912003190256655,
|
| 1612 |
+
"rewards/chosen": -0.9521958501776681,
|
| 1613 |
+
"rewards/rejected": -2.2994155996944756,
|
| 1614 |
+
"rewards/accuracies": 0.8125,
|
| 1615 |
+
"rewards/margins": 1.3472197427414359,
|
| 1616 |
+
"logps/chosen": -67.10248304605484,
|
| 1617 |
+
"logps/rejected": -76.43788787126542,
|
| 1618 |
+
"epoch": 1.408,
|
| 1619 |
+
"step": 880
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"loss": 0.3983212947845459,
|
| 1623 |
+
"grad_norm": 0.137314572930336,
|
| 1624 |
+
"learning_rate": 2.888e-06,
|
| 1625 |
+
"entropy": 0.7771084957581479,
|
| 1626 |
+
"num_tokens": 11958211.0,
|
| 1627 |
+
"logits/chosen": -1.4873145300830262,
|
| 1628 |
+
"logits/rejected": -1.1852639710072062,
|
| 1629 |
+
"mean_token_accuracy": 0.7251299729570746,
|
| 1630 |
+
"rewards/chosen": -0.8975313696777448,
|
| 1631 |
+
"rewards/rejected": -2.268895266018808,
|
| 1632 |
+
"rewards/accuracies": 0.83125,
|
| 1633 |
+
"rewards/margins": 1.3713638976216316,
|
| 1634 |
+
"logps/chosen": -60.558188426494596,
|
| 1635 |
+
"logps/rejected": -73.10897732973099,
|
| 1636 |
+
"epoch": 1.424,
|
| 1637 |
+
"step": 890
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"loss": 0.3957575082778931,
|
| 1641 |
+
"grad_norm": 0.19129854440689087,
|
| 1642 |
+
"learning_rate": 2.808e-06,
|
| 1643 |
+
"entropy": 0.8165824526702636,
|
| 1644 |
+
"num_tokens": 12092461.0,
|
| 1645 |
+
"logits/chosen": -1.5874914451128626,
|
| 1646 |
+
"logits/rejected": -1.329515978271273,
|
| 1647 |
+
"mean_token_accuracy": 0.7155879216268659,
|
| 1648 |
+
"rewards/chosen": -0.7802266632163082,
|
| 1649 |
+
"rewards/rejected": -2.118549132300541,
|
| 1650 |
+
"rewards/accuracies": 0.85,
|
| 1651 |
+
"rewards/margins": 1.3383224747609348,
|
| 1652 |
+
"logps/chosen": -61.53330332040787,
|
| 1653 |
+
"logps/rejected": -73.7201205611229,
|
| 1654 |
+
"epoch": 1.44,
|
| 1655 |
+
"step": 900
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"loss": 0.45516347885131836,
|
| 1659 |
+
"grad_norm": 0.47428420186042786,
|
| 1660 |
+
"learning_rate": 2.728e-06,
|
| 1661 |
+
"entropy": 0.7915405740670394,
|
| 1662 |
+
"num_tokens": 12227051.0,
|
| 1663 |
+
"logits/chosen": -1.6732329517329005,
|
| 1664 |
+
"logits/rejected": -1.3390952331955357,
|
| 1665 |
+
"mean_token_accuracy": 0.7120846627280116,
|
| 1666 |
+
"rewards/chosen": -0.9510568362500635,
|
| 1667 |
+
"rewards/rejected": -2.322500329572358,
|
| 1668 |
+
"rewards/accuracies": 0.83125,
|
| 1669 |
+
"rewards/margins": 1.3714434984140098,
|
| 1670 |
+
"logps/chosen": -61.362674832344055,
|
| 1671 |
+
"logps/rejected": -76.65392490625382,
|
| 1672 |
+
"epoch": 1.456,
|
| 1673 |
+
"step": 910
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"loss": 0.44686598777770997,
|
| 1677 |
+
"grad_norm": 0.17026080191135406,
|
| 1678 |
+
"learning_rate": 2.648e-06,
|
| 1679 |
+
"entropy": 0.7746687031816691,
|
| 1680 |
+
"num_tokens": 12360999.0,
|
| 1681 |
+
"logits/chosen": -1.6152128870414397,
|
| 1682 |
+
"logits/rejected": -1.2774718241622642,
|
| 1683 |
+
"mean_token_accuracy": 0.6930691134184599,
|
| 1684 |
+
"rewards/chosen": -1.0418641724740154,
|
| 1685 |
+
"rewards/rejected": -2.4828481852709956,
|
| 1686 |
+
"rewards/accuracies": 0.8,
|
| 1687 |
+
"rewards/margins": 1.4409840082284062,
|
| 1688 |
+
"logps/chosen": -61.202474081516264,
|
| 1689 |
+
"logps/rejected": -78.0609146118164,
|
| 1690 |
+
"epoch": 1.472,
|
| 1691 |
+
"step": 920
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"loss": 0.41604127883911135,
|
| 1695 |
+
"grad_norm": 0.23857389390468597,
|
| 1696 |
+
"learning_rate": 2.568e-06,
|
| 1697 |
+
"entropy": 0.8160373901191633,
|
| 1698 |
+
"num_tokens": 12495017.0,
|
| 1699 |
+
"logits/chosen": -1.5685694604309866,
|
| 1700 |
+
"logits/rejected": -1.288015625585243,
|
| 1701 |
+
"mean_token_accuracy": 0.6899296241812408,
|
| 1702 |
+
"rewards/chosen": -0.9305556651204825,
|
| 1703 |
+
"rewards/rejected": -2.4140758615511,
|
| 1704 |
+
"rewards/accuracies": 0.81875,
|
| 1705 |
+
"rewards/margins": 1.483520200336352,
|
| 1706 |
+
"logps/chosen": -64.11480586528778,
|
| 1707 |
+
"logps/rejected": -76.87684850692749,
|
| 1708 |
+
"epoch": 1.488,
|
| 1709 |
+
"step": 930
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"loss": 0.41258530616760253,
|
| 1713 |
+
"grad_norm": 0.19719047844409943,
|
| 1714 |
+
"learning_rate": 2.488e-06,
|
| 1715 |
+
"entropy": 0.8179522630613064,
|
| 1716 |
+
"num_tokens": 12628825.0,
|
| 1717 |
+
"logits/chosen": -1.6273103299993834,
|
| 1718 |
+
"logits/rejected": -1.3035281686517532,
|
| 1719 |
+
"mean_token_accuracy": 0.7133962934836745,
|
| 1720 |
+
"rewards/chosen": -0.8954242441719543,
|
| 1721 |
+
"rewards/rejected": -2.4310544121120072,
|
| 1722 |
+
"rewards/accuracies": 0.8125,
|
| 1723 |
+
"rewards/margins": 1.5356301709543914,
|
| 1724 |
+
"logps/chosen": -66.09550179243088,
|
| 1725 |
+
"logps/rejected": -76.01929485797882,
|
| 1726 |
+
"epoch": 1.504,
|
| 1727 |
+
"step": 940
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"loss": 0.38037173748016356,
|
| 1731 |
+
"grad_norm": 0.13135337829589844,
|
| 1732 |
+
"learning_rate": 2.408e-06,
|
| 1733 |
+
"entropy": 0.7855818087118678,
|
| 1734 |
+
"num_tokens": 12763474.0,
|
| 1735 |
+
"logits/chosen": -1.64132315324206,
|
| 1736 |
+
"logits/rejected": -1.2714727687609813,
|
| 1737 |
+
"mean_token_accuracy": 0.715859504416585,
|
| 1738 |
+
"rewards/chosen": -0.9226883488332532,
|
| 1739 |
+
"rewards/rejected": -2.500108515098691,
|
| 1740 |
+
"rewards/accuracies": 0.8375,
|
| 1741 |
+
"rewards/margins": 1.577420162782073,
|
| 1742 |
+
"logps/chosen": -62.92022385001182,
|
| 1743 |
+
"logps/rejected": -80.72790479660034,
|
| 1744 |
+
"epoch": 1.52,
|
| 1745 |
+
"step": 950
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"loss": 0.4409198760986328,
|
| 1749 |
+
"grad_norm": 0.31303665041923523,
|
| 1750 |
+
"learning_rate": 2.3280000000000004e-06,
|
| 1751 |
+
"entropy": 0.8127735982066951,
|
| 1752 |
+
"num_tokens": 12897107.0,
|
| 1753 |
+
"logits/chosen": -1.4399391210346026,
|
| 1754 |
+
"logits/rejected": -1.230631261634329,
|
| 1755 |
+
"mean_token_accuracy": 0.7026735378429294,
|
| 1756 |
+
"rewards/chosen": -0.934967651846091,
|
| 1757 |
+
"rewards/rejected": -2.3322817377047613,
|
| 1758 |
+
"rewards/accuracies": 0.8125,
|
| 1759 |
+
"rewards/margins": 1.397314086277038,
|
| 1760 |
+
"logps/chosen": -62.542164742946625,
|
| 1761 |
+
"logps/rejected": -73.16161739826202,
|
| 1762 |
+
"epoch": 1.536,
|
| 1763 |
+
"step": 960
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"loss": 0.4280536651611328,
|
| 1767 |
+
"grad_norm": 0.23763145506381989,
|
| 1768 |
+
"learning_rate": 2.2480000000000003e-06,
|
| 1769 |
+
"entropy": 0.7570945325132925,
|
| 1770 |
+
"num_tokens": 13030353.0,
|
| 1771 |
+
"logits/chosen": -1.7085164768196859,
|
| 1772 |
+
"logits/rejected": -1.3653086009027349,
|
| 1773 |
+
"mean_token_accuracy": 0.7043427483178675,
|
| 1774 |
+
"rewards/chosen": -0.8528924703747179,
|
| 1775 |
+
"rewards/rejected": -2.3349083499619154,
|
| 1776 |
+
"rewards/accuracies": 0.84375,
|
| 1777 |
+
"rewards/margins": 1.4820158663205802,
|
| 1778 |
+
"logps/chosen": -54.941452419757844,
|
| 1779 |
+
"logps/rejected": -76.01085047721863,
|
| 1780 |
+
"epoch": 1.552,
|
| 1781 |
+
"step": 970
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"loss": 0.4093909740447998,
|
| 1785 |
+
"grad_norm": 0.15855129063129425,
|
| 1786 |
+
"learning_rate": 2.1680000000000002e-06,
|
| 1787 |
+
"entropy": 0.8031133261421928,
|
| 1788 |
+
"num_tokens": 13164375.0,
|
| 1789 |
+
"logits/chosen": -1.6047969528216535,
|
| 1790 |
+
"logits/rejected": -1.3126213199277794,
|
| 1791 |
+
"mean_token_accuracy": 0.7228384926915169,
|
| 1792 |
+
"rewards/chosen": -0.9882222296495456,
|
| 1793 |
+
"rewards/rejected": -2.488050222734455,
|
| 1794 |
+
"rewards/accuracies": 0.7875,
|
| 1795 |
+
"rewards/margins": 1.49982799179852,
|
| 1796 |
+
"logps/chosen": -55.37274434566498,
|
| 1797 |
+
"logps/rejected": -80.98553432226181,
|
| 1798 |
+
"epoch": 1.568,
|
| 1799 |
+
"step": 980
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"loss": 0.4008018016815186,
|
| 1803 |
+
"grad_norm": 0.2002083659172058,
|
| 1804 |
+
"learning_rate": 2.088e-06,
|
| 1805 |
+
"entropy": 0.7641958858119324,
|
| 1806 |
+
"num_tokens": 13298403.0,
|
| 1807 |
+
"logits/chosen": -1.6807618807671876,
|
| 1808 |
+
"logits/rejected": -1.4103444427587308,
|
| 1809 |
+
"mean_token_accuracy": 0.7030234351754189,
|
| 1810 |
+
"rewards/chosen": -1.01242817313032,
|
| 1811 |
+
"rewards/rejected": -2.5514646800002083,
|
| 1812 |
+
"rewards/accuracies": 0.85,
|
| 1813 |
+
"rewards/margins": 1.5390365039929748,
|
| 1814 |
+
"logps/chosen": -60.99553867578506,
|
| 1815 |
+
"logps/rejected": -78.78072011470795,
|
| 1816 |
+
"epoch": 1.584,
|
| 1817 |
+
"step": 990
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"loss": 0.36574759483337405,
|
| 1821 |
+
"grad_norm": 0.20195689797401428,
|
| 1822 |
+
"learning_rate": 2.008e-06,
|
| 1823 |
+
"entropy": 0.7913652153627482,
|
| 1824 |
+
"num_tokens": 13433900.0,
|
| 1825 |
+
"logits/chosen": -1.6205719198466766,
|
| 1826 |
+
"logits/rejected": -1.2734592295015583,
|
| 1827 |
+
"mean_token_accuracy": 0.7198475800454617,
|
| 1828 |
+
"rewards/chosen": -1.0852013303767307,
|
| 1829 |
+
"rewards/rejected": -2.7301569358445703,
|
| 1830 |
+
"rewards/accuracies": 0.83125,
|
| 1831 |
+
"rewards/margins": 1.6449556024745107,
|
| 1832 |
+
"logps/chosen": -61.82057131528855,
|
| 1833 |
+
"logps/rejected": -90.16037954092026,
|
| 1834 |
+
"epoch": 1.6,
|
| 1835 |
+
"step": 1000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"loss": 0.4564220905303955,
|
| 1839 |
+
"grad_norm": 0.23998261988162994,
|
| 1840 |
+
"learning_rate": 1.928e-06,
|
| 1841 |
+
"entropy": 0.7775021609733812,
|
| 1842 |
+
"num_tokens": 13567714.0,
|
| 1843 |
+
"logits/chosen": -1.6069885571949023,
|
| 1844 |
+
"logits/rejected": -1.2448455268540815,
|
| 1845 |
+
"mean_token_accuracy": 0.7207059916108847,
|
| 1846 |
+
"rewards/chosen": -0.9900020190340001,
|
| 1847 |
+
"rewards/rejected": -2.2949246817035602,
|
| 1848 |
+
"rewards/accuracies": 0.80625,
|
| 1849 |
+
"rewards/margins": 1.30492265666835,
|
| 1850 |
+
"logps/chosen": -56.62804977893829,
|
| 1851 |
+
"logps/rejected": -74.89186335802079,
|
| 1852 |
+
"epoch": 1.616,
|
| 1853 |
+
"step": 1010
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"loss": 0.42406411170959474,
|
| 1857 |
+
"grad_norm": 0.15388637781143188,
|
| 1858 |
+
"learning_rate": 1.8480000000000001e-06,
|
| 1859 |
+
"entropy": 0.8878116343752481,
|
| 1860 |
+
"num_tokens": 13702663.0,
|
| 1861 |
+
"logits/chosen": -1.4586615375446985,
|
| 1862 |
+
"logits/rejected": -1.202710052546826,
|
| 1863 |
+
"mean_token_accuracy": 0.697345650754869,
|
| 1864 |
+
"rewards/chosen": -0.8269692128349562,
|
| 1865 |
+
"rewards/rejected": -2.332517017016653,
|
| 1866 |
+
"rewards/accuracies": 0.8,
|
| 1867 |
+
"rewards/margins": 1.5055477982386947,
|
| 1868 |
+
"logps/chosen": -67.33907754421234,
|
| 1869 |
+
"logps/rejected": -78.72829058170319,
|
| 1870 |
+
"epoch": 1.6320000000000001,
|
| 1871 |
+
"step": 1020
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"loss": 0.5315371036529541,
|
| 1875 |
+
"grad_norm": 0.457735151052475,
|
| 1876 |
+
"learning_rate": 1.7680000000000003e-06,
|
| 1877 |
+
"entropy": 0.7946982815090451,
|
| 1878 |
+
"num_tokens": 13836837.0,
|
| 1879 |
+
"logits/chosen": -1.5394904545256396,
|
| 1880 |
+
"logits/rejected": -1.3461529774623355,
|
| 1881 |
+
"mean_token_accuracy": 0.7092686965130269,
|
| 1882 |
+
"rewards/chosen": -1.002942830277607,
|
| 1883 |
+
"rewards/rejected": -2.0356479603942716,
|
| 1884 |
+
"rewards/accuracies": 0.78125,
|
| 1885 |
+
"rewards/margins": 1.032705131266266,
|
| 1886 |
+
"logps/chosen": -59.03670599460602,
|
| 1887 |
+
"logps/rejected": -74.02852568626403,
|
| 1888 |
+
"epoch": 1.6480000000000001,
|
| 1889 |
+
"step": 1030
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"loss": 0.38461852073669434,
|
| 1893 |
+
"grad_norm": 0.39034000039100647,
|
| 1894 |
+
"learning_rate": 1.6880000000000002e-06,
|
| 1895 |
+
"entropy": 0.8253316642483697,
|
| 1896 |
+
"num_tokens": 13971505.0,
|
| 1897 |
+
"logits/chosen": -1.5883768642108542,
|
| 1898 |
+
"logits/rejected": -1.1797071518191893,
|
| 1899 |
+
"mean_token_accuracy": 0.7223691996186972,
|
| 1900 |
+
"rewards/chosen": -0.8903896576404804,
|
| 1901 |
+
"rewards/rejected": -2.397843297244981,
|
| 1902 |
+
"rewards/accuracies": 0.8625,
|
| 1903 |
+
"rewards/margins": 1.5074536396190523,
|
| 1904 |
+
"logps/chosen": -66.58087438344955,
|
| 1905 |
+
"logps/rejected": -80.14289541244507,
|
| 1906 |
+
"epoch": 1.6640000000000001,
|
| 1907 |
+
"step": 1040
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"loss": 0.382495641708374,
|
| 1911 |
+
"grad_norm": 0.1509310007095337,
|
| 1912 |
+
"learning_rate": 1.608e-06,
|
| 1913 |
+
"entropy": 0.7834752728114835,
|
| 1914 |
+
"num_tokens": 14105815.0,
|
| 1915 |
+
"logits/chosen": -1.7158877727658357,
|
| 1916 |
+
"logits/rejected": -1.4181691521558832,
|
| 1917 |
+
"mean_token_accuracy": 0.7118882562965154,
|
| 1918 |
+
"rewards/chosen": -0.8125111373268737,
|
| 1919 |
+
"rewards/rejected": -2.3814144938398387,
|
| 1920 |
+
"rewards/accuracies": 0.825,
|
| 1921 |
+
"rewards/margins": 1.5689033489674329,
|
| 1922 |
+
"logps/chosen": -59.85800029039383,
|
| 1923 |
+
"logps/rejected": -76.31336205005645,
|
| 1924 |
+
"epoch": 1.6800000000000002,
|
| 1925 |
+
"step": 1050
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"loss": 0.3997781276702881,
|
| 1929 |
+
"grad_norm": 0.16973033547401428,
|
| 1930 |
+
"learning_rate": 1.528e-06,
|
| 1931 |
+
"entropy": 0.7523306506569497,
|
| 1932 |
+
"num_tokens": 14238856.0,
|
| 1933 |
+
"logits/chosen": -1.6062346133682144,
|
| 1934 |
+
"logits/rejected": -1.2810464342593764,
|
| 1935 |
+
"mean_token_accuracy": 0.7258081361651421,
|
| 1936 |
+
"rewards/chosen": -0.6798284199699992,
|
| 1937 |
+
"rewards/rejected": -2.1458812093129382,
|
| 1938 |
+
"rewards/accuracies": 0.825,
|
| 1939 |
+
"rewards/margins": 1.4660527911037207,
|
| 1940 |
+
"logps/chosen": -52.1538228392601,
|
| 1941 |
+
"logps/rejected": -69.79548509120941,
|
| 1942 |
+
"epoch": 1.696,
|
| 1943 |
+
"step": 1060
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"loss": 0.4718148708343506,
|
| 1947 |
+
"grad_norm": 0.20180265605449677,
|
| 1948 |
+
"learning_rate": 1.4480000000000002e-06,
|
| 1949 |
+
"entropy": 0.8426862467269529,
|
| 1950 |
+
"num_tokens": 14373384.0,
|
| 1951 |
+
"logits/chosen": -1.414872733093532,
|
| 1952 |
+
"logits/rejected": -1.2033987674482276,
|
| 1953 |
+
"mean_token_accuracy": 0.7098671667277813,
|
| 1954 |
+
"rewards/chosen": -0.7758016889652936,
|
| 1955 |
+
"rewards/rejected": -1.9979332615272143,
|
| 1956 |
+
"rewards/accuracies": 0.80625,
|
| 1957 |
+
"rewards/margins": 1.2221315732225775,
|
| 1958 |
+
"logps/chosen": -61.47320030927658,
|
| 1959 |
+
"logps/rejected": -75.88288476467133,
|
| 1960 |
+
"epoch": 1.712,
|
| 1961 |
+
"step": 1070
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"loss": 0.40792551040649416,
|
| 1965 |
+
"grad_norm": 0.12882812321186066,
|
| 1966 |
+
"learning_rate": 1.368e-06,
|
| 1967 |
+
"entropy": 0.8100415779626928,
|
| 1968 |
+
"num_tokens": 14507731.0,
|
| 1969 |
+
"logits/chosen": -1.554711909013513,
|
| 1970 |
+
"logits/rejected": -1.286161994780994,
|
| 1971 |
+
"mean_token_accuracy": 0.7126545551232993,
|
| 1972 |
+
"rewards/chosen": -0.7657065154984594,
|
| 1973 |
+
"rewards/rejected": -2.140209031692939,
|
| 1974 |
+
"rewards/accuracies": 0.83125,
|
| 1975 |
+
"rewards/margins": 1.3745025174692274,
|
| 1976 |
+
"logps/chosen": -59.775043559074405,
|
| 1977 |
+
"logps/rejected": -73.95411350727082,
|
| 1978 |
+
"epoch": 1.728,
|
| 1979 |
+
"step": 1080
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"loss": 0.460341215133667,
|
| 1983 |
+
"grad_norm": 0.22980692982673645,
|
| 1984 |
+
"learning_rate": 1.288e-06,
|
| 1985 |
+
"entropy": 0.8181033694170765,
|
| 1986 |
+
"num_tokens": 14642108.0,
|
| 1987 |
+
"logits/chosen": -1.6803376592297787,
|
| 1988 |
+
"logits/rejected": -1.3255758714406825,
|
| 1989 |
+
"mean_token_accuracy": 0.6975156743079424,
|
| 1990 |
+
"rewards/chosen": -0.8895529725443339,
|
| 1991 |
+
"rewards/rejected": -2.2926501425448804,
|
| 1992 |
+
"rewards/accuracies": 0.78125,
|
| 1993 |
+
"rewards/margins": 1.4030971705913544,
|
| 1994 |
+
"logps/chosen": -61.22287553548813,
|
| 1995 |
+
"logps/rejected": -76.53585437536239,
|
| 1996 |
+
"epoch": 1.744,
|
| 1997 |
+
"step": 1090
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"loss": 0.38245701789855957,
|
| 2001 |
+
"grad_norm": 0.3598242700099945,
|
| 2002 |
+
"learning_rate": 1.2080000000000001e-06,
|
| 2003 |
+
"entropy": 0.8136682996293529,
|
| 2004 |
+
"num_tokens": 14775766.0,
|
| 2005 |
+
"logits/chosen": -1.748982279233705,
|
| 2006 |
+
"logits/rejected": -1.3789075393037096,
|
| 2007 |
+
"mean_token_accuracy": 0.704171646013856,
|
| 2008 |
+
"rewards/chosen": -0.8589008188049775,
|
| 2009 |
+
"rewards/rejected": -2.463993888767436,
|
| 2010 |
+
"rewards/accuracies": 0.80625,
|
| 2011 |
+
"rewards/margins": 1.6050930766854434,
|
| 2012 |
+
"logps/chosen": -59.82941732406616,
|
| 2013 |
+
"logps/rejected": -79.58926742076873,
|
| 2014 |
+
"epoch": 1.76,
|
| 2015 |
+
"step": 1100
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"loss": 0.3972444772720337,
|
| 2019 |
+
"grad_norm": 0.1654120236635208,
|
| 2020 |
+
"learning_rate": 1.128e-06,
|
| 2021 |
+
"entropy": 0.8605185098713264,
|
| 2022 |
+
"num_tokens": 14910427.0,
|
| 2023 |
+
"logits/chosen": -1.5821960481799344,
|
| 2024 |
+
"logits/rejected": -1.3003056640036224,
|
| 2025 |
+
"mean_token_accuracy": 0.7196848660707473,
|
| 2026 |
+
"rewards/chosen": -0.6700921900264802,
|
| 2027 |
+
"rewards/rejected": -2.134859440010041,
|
| 2028 |
+
"rewards/accuracies": 0.8375,
|
| 2029 |
+
"rewards/margins": 1.464767256891355,
|
| 2030 |
+
"logps/chosen": -62.29144715070724,
|
| 2031 |
+
"logps/rejected": -75.43621026277542,
|
| 2032 |
+
"epoch": 1.776,
|
| 2033 |
+
"step": 1110
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"loss": 0.3881003141403198,
|
| 2037 |
+
"grad_norm": 0.31780576705932617,
|
| 2038 |
+
"learning_rate": 1.0480000000000002e-06,
|
| 2039 |
+
"entropy": 0.7536086188949411,
|
| 2040 |
+
"num_tokens": 15044039.0,
|
| 2041 |
+
"logits/chosen": -1.841065192100882,
|
| 2042 |
+
"logits/rejected": -1.4748827260004975,
|
| 2043 |
+
"mean_token_accuracy": 0.7127156775444746,
|
| 2044 |
+
"rewards/chosen": -0.8550866460078396,
|
| 2045 |
+
"rewards/rejected": -2.4443933775830375,
|
| 2046 |
+
"rewards/accuracies": 0.8125,
|
| 2047 |
+
"rewards/margins": 1.589306729659438,
|
| 2048 |
+
"logps/chosen": -58.62155553102493,
|
| 2049 |
+
"logps/rejected": -75.7669577240944,
|
| 2050 |
+
"epoch": 1.792,
|
| 2051 |
+
"step": 1120
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"loss": 0.4013789653778076,
|
| 2055 |
+
"grad_norm": 0.2001297026872635,
|
| 2056 |
+
"learning_rate": 9.68e-07,
|
| 2057 |
+
"entropy": 0.7950713085068856,
|
| 2058 |
+
"num_tokens": 15178463.0,
|
| 2059 |
+
"logits/chosen": -1.5685507002654315,
|
| 2060 |
+
"logits/rejected": -1.2775390358253036,
|
| 2061 |
+
"mean_token_accuracy": 0.7291223388165236,
|
| 2062 |
+
"rewards/chosen": -0.9785112287499942,
|
| 2063 |
+
"rewards/rejected": -2.3648490178980865,
|
| 2064 |
+
"rewards/accuracies": 0.80625,
|
| 2065 |
+
"rewards/margins": 1.3863377826288343,
|
| 2066 |
+
"logps/chosen": -57.15689754486084,
|
| 2067 |
+
"logps/rejected": -79.94988844394683,
|
| 2068 |
+
"epoch": 1.808,
|
| 2069 |
+
"step": 1130
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"loss": 0.49084672927856443,
|
| 2073 |
+
"grad_norm": 0.282905250787735,
|
| 2074 |
+
"learning_rate": 8.880000000000001e-07,
|
| 2075 |
+
"entropy": 0.8204654254193884,
|
| 2076 |
+
"num_tokens": 15312634.0,
|
| 2077 |
+
"logits/chosen": -1.7517282709034656,
|
| 2078 |
+
"logits/rejected": -1.3190052625389146,
|
| 2079 |
+
"mean_token_accuracy": 0.702907775528729,
|
| 2080 |
+
"rewards/chosen": -1.0180173698186992,
|
| 2081 |
+
"rewards/rejected": -2.349904873233754,
|
| 2082 |
+
"rewards/accuracies": 0.74375,
|
| 2083 |
+
"rewards/margins": 1.3318874984048308,
|
| 2084 |
+
"logps/chosen": -60.69031958580017,
|
| 2085 |
+
"logps/rejected": -82.0468565583229,
|
| 2086 |
+
"epoch": 1.8239999999999998,
|
| 2087 |
+
"step": 1140
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"loss": 0.41397271156311033,
|
| 2091 |
+
"grad_norm": 0.24510224163532257,
|
| 2092 |
+
"learning_rate": 8.08e-07,
|
| 2093 |
+
"entropy": 0.7518876982045185,
|
| 2094 |
+
"num_tokens": 15447212.0,
|
| 2095 |
+
"logits/chosen": -1.7534279282280079,
|
| 2096 |
+
"logits/rejected": -1.4486691671151501,
|
| 2097 |
+
"mean_token_accuracy": 0.7159834343940019,
|
| 2098 |
+
"rewards/chosen": -0.8752170269319322,
|
| 2099 |
+
"rewards/rejected": -2.376165826232318,
|
| 2100 |
+
"rewards/accuracies": 0.8125,
|
| 2101 |
+
"rewards/margins": 1.5009488008916378,
|
| 2102 |
+
"logps/chosen": -62.758730709552765,
|
| 2103 |
+
"logps/rejected": -75.16731867790222,
|
| 2104 |
+
"epoch": 1.8399999999999999,
|
| 2105 |
+
"step": 1150
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"loss": 0.394866418838501,
|
| 2109 |
+
"grad_norm": 0.24172107875347137,
|
| 2110 |
+
"learning_rate": 7.280000000000001e-07,
|
| 2111 |
+
"entropy": 0.8392532223559102,
|
| 2112 |
+
"num_tokens": 15582198.0,
|
| 2113 |
+
"logits/chosen": -1.611792128266145,
|
| 2114 |
+
"logits/rejected": -1.3081702823225343,
|
| 2115 |
+
"mean_token_accuracy": 0.6921530704945326,
|
| 2116 |
+
"rewards/chosen": -0.891454957460519,
|
| 2117 |
+
"rewards/rejected": -2.299446607741993,
|
| 2118 |
+
"rewards/accuracies": 0.825,
|
| 2119 |
+
"rewards/margins": 1.4079916514456272,
|
| 2120 |
+
"logps/chosen": -66.89741308689118,
|
| 2121 |
+
"logps/rejected": -77.80138423442841,
|
| 2122 |
+
"epoch": 1.8559999999999999,
|
| 2123 |
+
"step": 1160
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"loss": 0.36963462829589844,
|
| 2127 |
+
"grad_norm": 0.10854224115610123,
|
| 2128 |
+
"learning_rate": 6.48e-07,
|
| 2129 |
+
"entropy": 0.7845139768323861,
|
| 2130 |
+
"num_tokens": 15716912.0,
|
| 2131 |
+
"logits/chosen": -1.669064124473416,
|
| 2132 |
+
"logits/rejected": -1.3501869835794094,
|
| 2133 |
+
"mean_token_accuracy": 0.6950992489233613,
|
| 2134 |
+
"rewards/chosen": -1.0043300430756061,
|
| 2135 |
+
"rewards/rejected": -2.6331260551698508,
|
| 2136 |
+
"rewards/accuracies": 0.85625,
|
| 2137 |
+
"rewards/margins": 1.6287960140965878,
|
| 2138 |
+
"logps/chosen": -63.72115622758865,
|
| 2139 |
+
"logps/rejected": -79.68026465177536,
|
| 2140 |
+
"epoch": 1.8719999999999999,
|
| 2141 |
+
"step": 1170
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"loss": 0.4263630390167236,
|
| 2145 |
+
"grad_norm": 0.5233930349349976,
|
| 2146 |
+
"learning_rate": 5.680000000000001e-07,
|
| 2147 |
+
"entropy": 0.784919643367175,
|
| 2148 |
+
"num_tokens": 15851192.0,
|
| 2149 |
+
"logits/chosen": -1.6582383122340094,
|
| 2150 |
+
"logits/rejected": -1.4570516865536298,
|
| 2151 |
+
"mean_token_accuracy": 0.7130840895697474,
|
| 2152 |
+
"rewards/chosen": -0.924873108274187,
|
| 2153 |
+
"rewards/rejected": -2.3207803161232734,
|
| 2154 |
+
"rewards/accuracies": 0.79375,
|
| 2155 |
+
"rewards/margins": 1.3959072068799288,
|
| 2156 |
+
"logps/chosen": -62.35584118366241,
|
| 2157 |
+
"logps/rejected": -71.84658821821213,
|
| 2158 |
+
"epoch": 1.888,
|
| 2159 |
+
"step": 1180
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"loss": 0.4169583797454834,
|
| 2163 |
+
"grad_norm": 0.2042553722858429,
|
| 2164 |
+
"learning_rate": 4.88e-07,
|
| 2165 |
+
"entropy": 0.7986534590294468,
|
| 2166 |
+
"num_tokens": 15986206.0,
|
| 2167 |
+
"logits/chosen": -1.625323136112677,
|
| 2168 |
+
"logits/rejected": -1.3687231216559412,
|
| 2169 |
+
"mean_token_accuracy": 0.709136931784451,
|
| 2170 |
+
"rewards/chosen": -0.9780737107619644,
|
| 2171 |
+
"rewards/rejected": -2.4485225888201967,
|
| 2172 |
+
"rewards/accuracies": 0.80625,
|
| 2173 |
+
"rewards/margins": 1.4704488803166895,
|
| 2174 |
+
"logps/chosen": -69.08510230779648,
|
| 2175 |
+
"logps/rejected": -75.75201008319854,
|
| 2176 |
+
"epoch": 1.904,
|
| 2177 |
+
"step": 1190
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"loss": 0.49296092987060547,
|
| 2181 |
+
"grad_norm": 0.15136049687862396,
|
| 2182 |
+
"learning_rate": 4.0800000000000005e-07,
|
| 2183 |
+
"entropy": 0.7618352607241832,
|
| 2184 |
+
"num_tokens": 16119991.0,
|
| 2185 |
+
"logits/chosen": -1.6908642001340586,
|
| 2186 |
+
"logits/rejected": -1.4333833514773324,
|
| 2187 |
+
"mean_token_accuracy": 0.7040262173861265,
|
| 2188 |
+
"rewards/chosen": -1.0545528033093432,
|
| 2189 |
+
"rewards/rejected": -2.3307074559561443,
|
| 2190 |
+
"rewards/accuracies": 0.78125,
|
| 2191 |
+
"rewards/margins": 1.276154650375247,
|
| 2192 |
+
"logps/chosen": -61.58516470193863,
|
| 2193 |
+
"logps/rejected": -72.52572029829025,
|
| 2194 |
+
"epoch": 1.92,
|
| 2195 |
+
"step": 1200
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"loss": 0.43592305183410646,
|
| 2199 |
+
"grad_norm": 0.13972248136997223,
|
| 2200 |
+
"learning_rate": 3.280000000000001e-07,
|
| 2201 |
+
"entropy": 0.8202263483690331,
|
| 2202 |
+
"num_tokens": 16254379.0,
|
| 2203 |
+
"logits/chosen": -1.6105296053254246,
|
| 2204 |
+
"logits/rejected": -1.2790083528502694,
|
| 2205 |
+
"mean_token_accuracy": 0.7206195367500186,
|
| 2206 |
+
"rewards/chosen": -0.9561634575518838,
|
| 2207 |
+
"rewards/rejected": -2.4314136517699807,
|
| 2208 |
+
"rewards/accuracies": 0.8125,
|
| 2209 |
+
"rewards/margins": 1.4752501933835447,
|
| 2210 |
+
"logps/chosen": -60.72013909816742,
|
| 2211 |
+
"logps/rejected": -81.33328273296357,
|
| 2212 |
+
"epoch": 1.936,
|
| 2213 |
+
"step": 1210
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"loss": 0.45383148193359374,
|
| 2217 |
+
"grad_norm": 0.3797275125980377,
|
| 2218 |
+
"learning_rate": 2.48e-07,
|
| 2219 |
+
"entropy": 0.8815733156981878,
|
| 2220 |
+
"num_tokens": 16389460.0,
|
| 2221 |
+
"logits/chosen": -1.29787982782845,
|
| 2222 |
+
"logits/rejected": -1.1528923226810408,
|
| 2223 |
+
"mean_token_accuracy": 0.6916974617168308,
|
| 2224 |
+
"rewards/chosen": -0.8694630793921533,
|
| 2225 |
+
"rewards/rejected": -2.179070246774063,
|
| 2226 |
+
"rewards/accuracies": 0.8,
|
| 2227 |
+
"rewards/margins": 1.3096071674488485,
|
| 2228 |
+
"logps/chosen": -68.63869899511337,
|
| 2229 |
+
"logps/rejected": -76.16187700033188,
|
| 2230 |
+
"epoch": 1.952,
|
| 2231 |
+
"step": 1220
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"loss": 0.4355313301086426,
|
| 2235 |
+
"grad_norm": 0.20707087218761444,
|
| 2236 |
+
"learning_rate": 1.68e-07,
|
| 2237 |
+
"entropy": 0.8257959268114063,
|
| 2238 |
+
"num_tokens": 16523707.0,
|
| 2239 |
+
"logits/chosen": -1.5610055057823797,
|
| 2240 |
+
"logits/rejected": -1.2919258374928781,
|
| 2241 |
+
"mean_token_accuracy": 0.7034160443581641,
|
| 2242 |
+
"rewards/chosen": -0.8847868947021198,
|
| 2243 |
+
"rewards/rejected": -2.1883990670132336,
|
| 2244 |
+
"rewards/accuracies": 0.81875,
|
| 2245 |
+
"rewards/margins": 1.3036121717654168,
|
| 2246 |
+
"logps/chosen": -63.40972650051117,
|
| 2247 |
+
"logps/rejected": -76.90884720087051,
|
| 2248 |
+
"epoch": 1.968,
|
| 2249 |
+
"step": 1230
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"loss": 0.4611621379852295,
|
| 2253 |
+
"grad_norm": 0.2545054852962494,
|
| 2254 |
+
"learning_rate": 8.800000000000001e-08,
|
| 2255 |
+
"entropy": 0.8352547256625258,
|
| 2256 |
+
"num_tokens": 16659265.0,
|
| 2257 |
+
"logits/chosen": -1.5704459906119035,
|
| 2258 |
+
"logits/rejected": -1.2433016542196085,
|
| 2259 |
+
"mean_token_accuracy": 0.7052778167650103,
|
| 2260 |
+
"rewards/chosen": -0.9597635358106344,
|
| 2261 |
+
"rewards/rejected": -2.3284011104144158,
|
| 2262 |
+
"rewards/accuracies": 0.80625,
|
| 2263 |
+
"rewards/margins": 1.3686375686433165,
|
| 2264 |
+
"logps/chosen": -68.85014188289642,
|
| 2265 |
+
"logps/rejected": -77.95877922773361,
|
| 2266 |
+
"epoch": 1.984,
|
| 2267 |
+
"step": 1240
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"loss": 0.4011569023132324,
|
| 2271 |
+
"grad_norm": 0.2973766326904297,
|
| 2272 |
+
"learning_rate": 8e-09,
|
| 2273 |
+
"entropy": 0.8292072300857398,
|
| 2274 |
+
"num_tokens": 16794078.0,
|
| 2275 |
+
"logits/chosen": -1.5225679573518773,
|
| 2276 |
+
"logits/rejected": -1.2299553825384233,
|
| 2277 |
+
"mean_token_accuracy": 0.727184671908617,
|
| 2278 |
+
"rewards/chosen": -0.8619254208082566,
|
| 2279 |
+
"rewards/rejected": -2.372782723305863,
|
| 2280 |
+
"rewards/accuracies": 0.80625,
|
| 2281 |
+
"rewards/margins": 1.5108572976663708,
|
| 2282 |
+
"logps/chosen": -63.519434094429016,
|
| 2283 |
+
"logps/rejected": -76.77453536987305,
|
| 2284 |
+
"epoch": 2.0,
|
| 2285 |
+
"step": 1250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"train_runtime": 6000.0528,
|
| 2289 |
+
"train_samples_per_second": 3.333,
|
| 2290 |
+
"train_steps_per_second": 0.208,
|
| 2291 |
+
"total_flos": 1.3612284112597402e+17,
|
| 2292 |
+
"train_loss": 0.452988232421875,
|
| 2293 |
+
"epoch": 2.0,
|
| 2294 |
+
"step": 1250
|
| 2295 |
+
}
|
| 2296 |
+
],
|
| 2297 |
+
"validation_monitor_size": 0,
|
| 2298 |
+
"validation_monitor_selection": "fixed_seed_random_without_replacement",
|
| 2299 |
+
"validation_monitor_seed": 23,
|
| 2300 |
+
"validation_monitor_indices": [],
|
| 2301 |
+
"early_stopping_patience": null
|
| 2302 |
+
}
|