Hugging Face's logo

cucl2
/
AnyAudio-Judge-7B

Audio-Text-to-Text
Transformers
Safetensors
English
Chinese
qwen2_5_omni
text-to-audio
audio
audio-language-model
instruction-following
rubric-based-evaluation
judge-model
Model card Files
xet
 Community
cucl2 commited on
Commit
e9d6389
·
verified ·
1 Parent(s): 83a725f

Add files using upload-large-folder tool

Browse files
Files changed (20) hide show
  1. .gitattributes +1 -0
  2. README.md +65 -0
  3. added_tokens.json +24 -0
  4. chat_template.jinja +7 -0
  5. config.json +634 -0
  6. generation_config.json +8 -0
  7. merges.txt +0 -0
  8. model-00001-of-00004.safetensors +3 -0
  9. model-00002-of-00004.safetensors +3 -0
  10. model-00003-of-00004.safetensors +3 -0
  11. model-00004-of-00004.safetensors +3 -0
  12. model.safetensors.index.json +0 -0
  13. preprocessor_config.json +31 -0
  14. special_tokens_map.json +38 -0
  15. spk_dict.pt +3 -0
  16. tokenizer.json +3 -0
  17. tokenizer_config.json +222 -0
  18. trainer_state.json +1354 -0
  19. video_preprocessor_config.json +55 -0
  20. vocab.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ - zh
6
+ library_name: transformers
7
+ base_model: Qwen/Qwen2.5-Omni-7B
8
+ tags:
9
+ - audio
10
+ - audio-language-model
11
+ - instruction-following
12
+ - rubric-based-evaluation
13
+ - judge-model
14
+ pipeline_tag: audio-text-to-text
15
+ ---
16
+
17
+ # AnyAudio-Judge-7B
18
+
19
+ `AnyAudio-Judge-7B` is a **dynamic rubric-based audio judge** built on top of [Qwen2.5-Omni-7B](https://huggingface.co/Qwen/Qwen2.5-Omni-7B). It predicts, for each yes/no rubric item describing one verifiable aspect of an audio caption, whether the audio satisfies that aspect — together with a short evidence string.
20
+
21
+ This is the smaller variant of the AnyAudio-Judge family. The larger `AnyAudio-Judge-30B` (initialized from Qwen3-Omni-30B-A3B-Captioner) is the variant reported in the paper. The 7B model is trained on the same SFT corpus and is intended for users who need a more efficient evaluator.
22
+
23
+ > Companion benchmark: [`cucl2/AnyAudio-Judge-Bench`](https://huggingface.co/datasets/cucl2/AnyAudio-Judge-Bench)
24
+ > Companion corpus: [`cucl2/AnyAudio-Judge-Corpus`](https://huggingface.co/datasets/cucl2/AnyAudio-Judge-Corpus)
25
+ > Companion 30B model: [`cucl2/AnyAudio-Judge-30B`](https://huggingface.co/cucl2/AnyAudio-Judge-30B)
26
+
27
+ ## Training
28
+
29
+ - **Base**: Qwen2.5-Omni-7B
30
+ - **Corpus**: 105K (audio, instruction, rubric, CoT) tuples (see `cucl2/AnyAudio-Judge-Corpus`)
31
+ - **Stage**: full-parameter SFT for 1 epoch
32
+ - 16 × H20 96GB
33
+ - per-device batch size 4, grad accumulation 1
34
+ - learning rate 1e-5
35
+
36
+ ## Usage
37
+
38
+ ```python
39
+ from anyaudio_judge import AnyAudioJudge, decompose_instruction
40
+
41
+ caption = "A gentle, delicate female voice, with soft and smooth pitch, calm and restrained throughout."
42
+ rubric = decompose_instruction(caption) # external LLM call
43
+
44
+ judge = AnyAudioJudge.from_pretrained("cucl2/AnyAudio-Judge-7B")
45
+ result = judge.judge("./demo.wav", rubric)
46
+ print("alignment_score:", result.score)
47
+ for item in result.items:
48
+ print(item.question, "->", item.answer)
49
+ ```
50
+
51
+ (See the [GitHub repo](https://github.com/) for the full pipeline including external rubric decomposition.)
52
+
53
+ ## License
54
+
55
+ Apache-2.0, inheriting the license of the base Qwen2.5-Omni-7B model.
56
+
57
+ ## Citation
58
+
59
+ ```bibtex
60
+ @inproceedings{anyaudiojudge2026,
61
+ title = {AnyAudio-Judge: A Dynamic Rubric-Based Benchmark and Evaluator for Audio Instruction Following},
62
+ booktitle = {Proceedings of ACL},
63
+ year = {2026}
64
+ }
65
+ ```
added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|AUDIO|>": 151646,
5
+ "<|IMAGE|>": 151655,
6
+ "<|VIDEO|>": 151656,
7
+ "<|audio_bos|>": 151647,
8
+ "<|audio_eos|>": 151648,
9
+ "<|box_end|>": 151649,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|quad_end|>": 151651,
19
+ "<|quad_start|>": 151650,
20
+ "<|repo_name|>": 151663,
21
+ "<|vision_bos|>": 151652,
22
+ "<|vision_eos|>": 151653,
23
+ "<|vision_pad|>": 151654
24
+ }
chat_template.jinja ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {% set audio_count = namespace(value=0) %}{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
2
+ You are a helpful assistant.<|im_end|>
3
+ {% endif %}<|im_start|>{{ message['role'] }}
4
+ {% if message['content'] is string %}{{ message['content'] }}<|im_end|>
5
+ {% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_bos|><|IMAGE|><|vision_eos|>{% elif content['type'] == 'audio' or 'audio' in content or 'audio_url' in content %}{% set audio_count.value = audio_count.value + 1 %}{% if add_audio_id %}Audio {{ audio_count.value }}: {% endif %}<|audio_bos|><|AUDIO|><|audio_eos|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_bos|><|VIDEO|><|vision_eos|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
6
+ {% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
7
+ {% endif %}
config.json ADDED
@@ -0,0 +1,634 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2_5OmniForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "enable_audio_output": false,
7
+ "enable_talker": true,
8
+ "eos_token_id": 151645,
9
+ "hidden_size": 3584,
10
+ "keys_to_ignore_at_inference": [
11
+ "past_key_values",
12
+ "hidden_states",
13
+ "attention_mask"
14
+ ],
15
+ "model_type": "qwen2_5_omni",
16
+ "pad_token_id": 151643,
17
+ "talker_config": {
18
+ "_attn_implementation_autoset": true,
19
+ "_name_or_path": "Qwen2.5-Omni-7B/talker",
20
+ "architectures": [
21
+ "Qwen2OmniTalkerForConditionalGeneration"
22
+ ],
23
+ "attention_dropout": 0.0,
24
+ "audio_end_token_id": 151648,
25
+ "audio_start_token_id": 151647,
26
+ "audio_token_index": 151646,
27
+ "dtype": "bfloat16",
28
+ "embedding_size": 3584,
29
+ "head_dim": 128,
30
+ "hidden_act": "silu",
31
+ "hidden_size": 896,
32
+ "image_token_index": 151655,
33
+ "init_std": 0.02,
34
+ "initializer_range": 0.02,
35
+ "intermediate_size": 18944,
36
+ "layer_types": [
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention",
53
+ "full_attention",
54
+ "full_attention",
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention"
61
+ ],
62
+ "max_position_embeddings": 32768,
63
+ "max_window_layers": 28,
64
+ "model_type": "qwen2_5_omni_talker",
65
+ "num_attention_heads": 12,
66
+ "num_hidden_layers": 24,
67
+ "num_key_value_heads": 4,
68
+ "position_id_per_seconds": 25,
69
+ "rms_norm_eps": 1e-06,
70
+ "rope_scaling": {
71
+ "mrope_section": [
72
+ 16,
73
+ 24,
74
+ 24
75
+ ],
76
+ "rope_type": "default",
77
+ "type": "default"
78
+ },
79
+ "rope_theta": 1000000.0,
80
+ "seconds_per_chunk": 2,
81
+ "sliding_window": null,
82
+ "spatial_merge_size": 2,
83
+ "tts_codec_end_token_id": 8294,
84
+ "tts_codec_mask_token_id": 8296,
85
+ "tts_codec_pad_token_id": 8292,
86
+ "tts_codec_start_token_id": 8293,
87
+ "tts_text_end_token_id": 151861,
88
+ "tts_text_pad_token_id": 151859,
89
+ "tts_text_start_token_id": 151860,
90
+ "use_cache": true,
91
+ "use_sliding_window": false,
92
+ "video_token_index": 151656,
93
+ "vision_end_token_id": 151653,
94
+ "vision_start_token_id": 151652,
95
+ "vocab_size": 8448
96
+ },
97
+ "thinker_config": {
98
+ "_attn_implementation_autoset": true,
99
+ "_name_or_path": "Qwen2.5-Omni-7B/thinker",
100
+ "architectures": [
101
+ "Qwen2OmniNaViTThinkerForConditionalGeneration"
102
+ ],
103
+ "audio_config": {
104
+ "_attn_implementation_autoset": true,
105
+ "_name_or_path": "",
106
+ "activation_dropout": 0.0,
107
+ "activation_function": "gelu",
108
+ "add_cross_attention": false,
109
+ "architectures": null,
110
+ "attention_dropout": 0.0,
111
+ "bad_words_ids": null,
112
+ "begin_suppress_tokens": null,
113
+ "bos_token_id": null,
114
+ "chunk_size_feed_forward": 0,
115
+ "cross_attention_hidden_size": null,
116
+ "d_model": 1280,
117
+ "decoder_start_token_id": null,
118
+ "diversity_penalty": 0.0,
119
+ "do_sample": false,
120
+ "dropout": 0.0,
121
+ "dtype": "bfloat16",
122
+ "early_stopping": false,
123
+ "encoder_attention_heads": 20,
124
+ "encoder_ffn_dim": 5120,
125
+ "encoder_layerdrop": 0.0,
126
+ "encoder_layers": 32,
127
+ "encoder_no_repeat_ngram_size": 0,
128
+ "eos_token_id": null,
129
+ "exponential_decay_length_penalty": null,
130
+ "finetuning_task": null,
131
+ "forced_bos_token_id": null,
132
+ "forced_eos_token_id": null,
133
+ "id2label": {
134
+ "0": "LABEL_0",
135
+ "1": "LABEL_1"
136
+ },
137
+ "init_std": 0.02,
138
+ "initializer_range": 0.02,
139
+ "is_decoder": false,
140
+ "is_encoder_decoder": false,
141
+ "label2id": {
142
+ "LABEL_0": 0,
143
+ "LABEL_1": 1
144
+ },
145
+ "length_penalty": 1.0,
146
+ "max_length": 20,
147
+ "max_source_positions": 1500,
148
+ "min_length": 0,
149
+ "model_type": "qwen2_5_omni_audio_encoder",
150
+ "n_window": 100,
151
+ "no_repeat_ngram_size": 0,
152
+ "num_beam_groups": 1,
153
+ "num_beams": 1,
154
+ "num_hidden_layers": 32,
155
+ "num_mel_bins": 128,
156
+ "num_return_sequences": 1,
157
+ "output_attentions": false,
158
+ "output_dim": 3584,
159
+ "output_hidden_states": false,
160
+ "output_scores": false,
161
+ "pad_token_id": null,
162
+ "prefix": null,
163
+ "problem_type": null,
164
+ "pruned_heads": {},
165
+ "remove_invalid_values": false,
166
+ "repetition_penalty": 1.0,
167
+ "return_dict": true,
168
+ "return_dict_in_generate": false,
169
+ "scale_embedding": false,
170
+ "sep_token_id": null,
171
+ "suppress_tokens": null,
172
+ "task_specific_params": null,
173
+ "temperature": 1.0,
174
+ "tf_legacy_loss": false,
175
+ "tie_encoder_decoder": false,
176
+ "tie_word_embeddings": true,
177
+ "tokenizer_class": null,
178
+ "top_k": 50,
179
+ "top_p": 1.0,
180
+ "torchscript": false,
181
+ "typical_p": 1.0,
182
+ "use_bfloat16": false
183
+ },
184
+ "audio_end_token_id": 151648,
185
+ "audio_start_token_id": 151647,
186
+ "bos_token_id": 151644,
187
+ "dtype": "bfloat16",
188
+ "eos_token_id": 151645,
189
+ "ignore_index": -100,
190
+ "init_std": 0.02,
191
+ "initializer_range": 0.02,
192
+ "model_type": "qwen2_5_omni_thinker",
193
+ "pad_token_id": 151643,
194
+ "position_id_per_seconds": 25,
195
+ "seconds_per_chunk": 2,
196
+ "text_config": {
197
+ "_name_or_path": "",
198
+ "add_cross_attention": false,
199
+ "architectures": null,
200
+ "attention_dropout": 0.0,
201
+ "bad_words_ids": null,
202
+ "begin_suppress_tokens": null,
203
+ "bos_token_id": null,
204
+ "chunk_size_feed_forward": 0,
205
+ "cross_attention_hidden_size": null,
206
+ "decoder_start_token_id": null,
207
+ "diversity_penalty": 0.0,
208
+ "do_sample": false,
209
+ "dtype": "bfloat16",
210
+ "early_stopping": false,
211
+ "encoder_no_repeat_ngram_size": 0,
212
+ "eos_token_id": null,
213
+ "exponential_decay_length_penalty": null,
214
+ "finetuning_task": null,
215
+ "forced_bos_token_id": null,
216
+ "forced_eos_token_id": null,
217
+ "hidden_act": "silu",
218
+ "hidden_size": 3584,
219
+ "id2label": {
220
+ "0": "LABEL_0",
221
+ "1": "LABEL_1"
222
+ },
223
+ "init_std": 0.02,
224
+ "initializer_range": 0.02,
225
+ "intermediate_size": 18944,
226
+ "is_decoder": false,
227
+ "is_encoder_decoder": false,
228
+ "label2id": {
229
+ "LABEL_0": 0,
230
+ "LABEL_1": 1
231
+ },
232
+ "layer_types": [
233
+ "full_attention",
234
+ "full_attention",
235
+ "full_attention",
236
+ "full_attention",
237
+ "full_attention",
238
+ "full_attention",
239
+ "full_attention",
240
+ "full_attention",
241
+ "full_attention",
242
+ "full_attention",
243
+ "full_attention",
244
+ "full_attention",
245
+ "full_attention",
246
+ "full_attention",
247
+ "full_attention",
248
+ "full_attention",
249
+ "full_attention",
250
+ "full_attention",
251
+ "full_attention",
252
+ "full_attention",
253
+ "full_attention",
254
+ "full_attention",
255
+ "full_attention",
256
+ "full_attention",
257
+ "full_attention",
258
+ "full_attention",
259
+ "full_attention",
260
+ "full_attention"
261
+ ],
262
+ "length_penalty": 1.0,
263
+ "max_length": 20,
264
+ "max_position_embeddings": 32768,
265
+ "max_window_layers": 28,
266
+ "min_length": 0,
267
+ "model_type": "qwen2_5_omni_text",
268
+ "no_repeat_ngram_size": 0,
269
+ "num_attention_heads": 28,
270
+ "num_beam_groups": 1,
271
+ "num_beams": 1,
272
+ "num_hidden_layers": 28,
273
+ "num_key_value_heads": 4,
274
+ "num_return_sequences": 1,
275
+ "output_attentions": false,
276
+ "output_hidden_states": false,
277
+ "output_scores": false,
278
+ "pad_token_id": 151643,
279
+ "prefix": null,
280
+ "problem_type": null,
281
+ "pruned_heads": {},
282
+ "remove_invalid_values": false,
283
+ "repetition_penalty": 1.0,
284
+ "return_dict": true,
285
+ "return_dict_in_generate": false,
286
+ "rms_norm_eps": 1e-06,
287
+ "rope_scaling": {
288
+ "mrope_section": [
289
+ 16,
290
+ 24,
291
+ 24
292
+ ],
293
+ "rope_type": "default",
294
+ "type": "default"
295
+ },
296
+ "rope_theta": 1000000.0,
297
+ "sep_token_id": null,
298
+ "sliding_window": null,
299
+ "suppress_tokens": null,
300
+ "task_specific_params": null,
301
+ "temperature": 1.0,
302
+ "tie_encoder_decoder": false,
303
+ "tie_word_embeddings": false,
304
+ "tokenizer_class": null,
305
+ "top_k": 50,
306
+ "top_p": 1.0,
307
+ "torchscript": false,
308
+ "typical_p": 1.0,
309
+ "use_cache": false,
310
+ "use_sliding_window": false,
311
+ "vocab_size": 152064
312
+ },
313
+ "user_token_id": 872,
314
+ "vision_config": {
315
+ "_attn_implementation_autoset": true,
316
+ "_name_or_path": "",
317
+ "add_cross_attention": false,
318
+ "architectures": null,
319
+ "bad_words_ids": null,
320
+ "begin_suppress_tokens": null,
321
+ "bos_token_id": null,
322
+ "chunk_size_feed_forward": 0,
323
+ "cross_attention_hidden_size": null,
324
+ "decoder_start_token_id": null,
325
+ "depth": 32,
326
+ "diversity_penalty": 0.0,
327
+ "do_sample": false,
328
+ "dtype": "bfloat16",
329
+ "early_stopping": false,
330
+ "embed_dim": 1280,
331
+ "encoder_no_repeat_ngram_size": 0,
332
+ "eos_token_id": null,
333
+ "exponential_decay_length_penalty": null,
334
+ "finetuning_task": null,
335
+ "forced_bos_token_id": null,
336
+ "forced_eos_token_id": null,
337
+ "fullatt_block_indexes": [
338
+ 7,
339
+ 15,
340
+ 23,
341
+ 31
342
+ ],
343
+ "hidden_act": "silu",
344
+ "hidden_size": 1280,
345
+ "id2label": {
346
+ "0": "LABEL_0",
347
+ "1": "LABEL_1"
348
+ },
349
+ "in_channels": 3,
350
+ "in_chans": 3,
351
+ "init_std": 0.02,
352
+ "initializer_range": 0.02,
353
+ "intermediate_size": 3420,
354
+ "is_decoder": false,
355
+ "is_encoder_decoder": false,
356
+ "label2id": {
357
+ "LABEL_0": 0,
358
+ "LABEL_1": 1
359
+ },
360
+ "length_penalty": 1.0,
361
+ "max_length": 20,
362
+ "min_length": 0,
363
+ "model_type": "qwen2_5_omni_vision_encoder",
364
+ "no_repeat_ngram_size": 0,
365
+ "num_beam_groups": 1,
366
+ "num_beams": 1,
367
+ "num_heads": 16,
368
+ "num_return_sequences": 1,
369
+ "out_hidden_size": 3584,
370
+ "output_attentions": false,
371
+ "output_hidden_states": false,
372
+ "output_scores": false,
373
+ "pad_token_id": null,
374
+ "patch_size": 14,
375
+ "prefix": null,
376
+ "problem_type": null,
377
+ "pruned_heads": {},
378
+ "remove_invalid_values": false,
379
+ "repetition_penalty": 1.0,
380
+ "return_dict": true,
381
+ "return_dict_in_generate": false,
382
+ "sep_token_id": null,
383
+ "spatial_merge_size": 2,
384
+ "spatial_patch_size": 14,
385
+ "suppress_tokens": null,
386
+ "task_specific_params": null,
387
+ "temperature": 1.0,
388
+ "temporal_patch_size": 2,
389
+ "tf_legacy_loss": false,
390
+ "tie_encoder_decoder": false,
391
+ "tie_word_embeddings": true,
392
+ "tokenizer_class": null,
393
+ "tokens_per_second": 25,
394
+ "top_k": 50,
395
+ "top_p": 1.0,
396
+ "torchscript": false,
397
+ "typical_p": 1.0,
398
+ "use_bfloat16": false,
399
+ "window_size": 112
400
+ },
401
+ "vision_end_token_id": 151653,
402
+ "vision_start_token_id": 151652,
403
+ "vision_token_id": 151654,
404
+ "audio_token_id": 151646,
405
+ "image_token_id": 151655,
406
+ "video_token_id": 151656
407
+ },
408
+ "token2wav_config": {
409
+ "_attn_implementation_autoset": true,
410
+ "bigvgan_config": {
411
+ "_attn_implementation_autoset": true,
412
+ "_name_or_path": "",
413
+ "add_cross_attention": false,
414
+ "architectures": null,
415
+ "bad_words_ids": null,
416
+ "begin_suppress_tokens": null,
417
+ "bos_token_id": null,
418
+ "chunk_size_feed_forward": 0,
419
+ "cross_attention_hidden_size": null,
420
+ "decoder_start_token_id": null,
421
+ "diversity_penalty": 0.0,
422
+ "do_sample": false,
423
+ "dtype": null,
424
+ "early_stopping": false,
425
+ "encoder_no_repeat_ngram_size": 0,
426
+ "eos_token_id": null,
427
+ "exponential_decay_length_penalty": null,
428
+ "finetuning_task": null,
429
+ "forced_bos_token_id": null,
430
+ "forced_eos_token_id": null,
431
+ "id2label": {
432
+ "0": "LABEL_0",
433
+ "1": "LABEL_1"
434
+ },
435
+ "is_decoder": false,
436
+ "is_encoder_decoder": false,
437
+ "label2id": {
438
+ "LABEL_0": 0,
439
+ "LABEL_1": 1
440
+ },
441
+ "length_penalty": 1.0,
442
+ "max_length": 20,
443
+ "mel_dim": 80,
444
+ "min_length": 0,
445
+ "model_type": "qwen2_5_omni_bigvgan",
446
+ "no_repeat_ngram_size": 0,
447
+ "num_beam_groups": 1,
448
+ "num_beams": 1,
449
+ "num_return_sequences": 1,
450
+ "output_attentions": false,
451
+ "output_hidden_states": false,
452
+ "output_scores": false,
453
+ "pad_token_id": null,
454
+ "prefix": null,
455
+ "problem_type": null,
456
+ "pruned_heads": {},
457
+ "remove_invalid_values": false,
458
+ "repetition_penalty": 1.0,
459
+ "resblock_dilation_sizes": [
460
+ [
461
+ 1,
462
+ 3,
463
+ 5
464
+ ],
465
+ [
466
+ 1,
467
+ 3,
468
+ 5
469
+ ],
470
+ [
471
+ 1,
472
+ 3,
473
+ 5
474
+ ]
475
+ ],
476
+ "resblock_kernel_sizes": [
477
+ 3,
478
+ 7,
479
+ 11
480
+ ],
481
+ "return_dict": true,
482
+ "return_dict_in_generate": false,
483
+ "sep_token_id": null,
484
+ "suppress_tokens": null,
485
+ "task_specific_params": null,
486
+ "temperature": 1.0,
487
+ "tf_legacy_loss": false,
488
+ "tie_encoder_decoder": false,
489
+ "tie_word_embeddings": true,
490
+ "tokenizer_class": null,
491
+ "top_k": 50,
492
+ "top_p": 1.0,
493
+ "torchscript": false,
494
+ "typical_p": 1.0,
495
+ "upsample_initial_channel": 1536,
496
+ "upsample_kernel_sizes": [
497
+ 11,
498
+ 7,
499
+ 4,
500
+ 4,
501
+ 4,
502
+ 4
503
+ ],
504
+ "upsample_rates": [
505
+ 5,
506
+ 3,
507
+ 2,
508
+ 2,
509
+ 2,
510
+ 2
511
+ ],
512
+ "use_bfloat16": false,
513
+ "use_bias_at_final": false
514
+ },
515
+ "dit_config": {
516
+ "_attn_implementation_autoset": true,
517
+ "_name_or_path": "",
518
+ "add_cross_attention": false,
519
+ "architectures": null,
520
+ "bad_words_ids": null,
521
+ "begin_suppress_tokens": null,
522
+ "block_size": 24,
523
+ "bos_token_id": null,
524
+ "chunk_size_feed_forward": 0,
525
+ "cross_attention_hidden_size": null,
526
+ "decoder_start_token_id": null,
527
+ "depth": 22,
528
+ "dim": 1024,
529
+ "diversity_penalty": 0.0,
530
+ "do_sample": false,
531
+ "dropout": 0.1,
532
+ "dtype": "float32",
533
+ "early_stopping": false,
534
+ "emb_dim": 512,
535
+ "enc_attention_channels": 64,
536
+ "enc_channels": [
537
+ 256,
538
+ 256,
539
+ 256,
540
+ 256,
541
+ 768
542
+ ],
543
+ "enc_dilations": [
544
+ 1,
545
+ 2,
546
+ 3,
547
+ 4,
548
+ 1
549
+ ],
550
+ "enc_dim": 128,
551
+ "enc_emb_dim": 192,
552
+ "enc_global_context": true,
553
+ "enc_kernel_sizes": [
554
+ 5,
555
+ 3,
556
+ 3,
557
+ 3,
558
+ 1
559
+ ],
560
+ "enc_lin_neurons": 192,
561
+ "enc_res2net_scale": 2,
562
+ "enc_se_channels": 64,
563
+ "encoder_no_repeat_ngram_size": 0,
564
+ "eos_token_id": null,
565
+ "exponential_decay_length_penalty": null,
566
+ "ff_mult": 2,
567
+ "finetuning_task": null,
568
+ "forced_bos_token_id": null,
569
+ "forced_eos_token_id": null,
570
+ "head_dim": 64,
571
+ "heads": 16,
572
+ "hidden_size": 1024,
573
+ "id2label": {
574
+ "0": "LABEL_0",
575
+ "1": "LABEL_1"
576
+ },
577
+ "is_decoder": false,
578
+ "is_encoder_decoder": false,
579
+ "label2id": {
580
+ "LABEL_0": 0,
581
+ "LABEL_1": 1
582
+ },
583
+ "length_penalty": 1.0,
584
+ "look_ahead_layers": [
585
+ 10
586
+ ],
587
+ "look_backward_layers": [
588
+ 0,
589
+ 20
590
+ ],
591
+ "max_length": 20,
592
+ "max_position_embeddings": 32768,
593
+ "mel_dim": 80,
594
+ "min_length": 0,
595
+ "model_type": "qwen2_5_omni_dit",
596
+ "no_repeat_ngram_size": 0,
597
+ "num_attention_heads": 16,
598
+ "num_beam_groups": 1,
599
+ "num_beams": 1,
600
+ "num_embeds": 8193,
601
+ "num_hidden_layers": 22,
602
+ "num_return_sequences": 1,
603
+ "output_attentions": false,
604
+ "output_hidden_states": false,
605
+ "output_scores": false,
606
+ "pad_token_id": null,
607
+ "prefix": null,
608
+ "problem_type": null,
609
+ "pruned_heads": {},
610
+ "remove_invalid_values": false,
611
+ "repeats": 2,
612
+ "repetition_penalty": 1.0,
613
+ "return_dict": true,
614
+ "return_dict_in_generate": false,
615
+ "rope_theta": 10000.0,
616
+ "sep_token_id": null,
617
+ "suppress_tokens": null,
618
+ "task_specific_params": null,
619
+ "temperature": 1.0,
620
+ "tf_legacy_loss": false,
621
+ "tie_encoder_decoder": false,
622
+ "tie_word_embeddings": true,
623
+ "tokenizer_class": null,
624
+ "top_k": 50,
625
+ "top_p": 1.0,
626
+ "torchscript": false,
627
+ "typical_p": 1.0,
628
+ "use_bfloat16": false
629
+ },
630
+ "dtype": "bfloat16",
631
+ "model_type": "qwen2_5_omni_token2wav"
632
+ },
633
+ "transformers_version": "4.57.0.dev0"
634
+ }
generation_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151643
6
+ ],
7
+ "transformers_version": "4.57.0.dev0"
8
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b33cc6724094d9c7a5f2a38e35d4865b0df177de57daf429c5516f0e517dc14
3
+ size 4985055536
model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95ab23397ebfa695f114d2b742643a81dbc935548c1bc33fdaaaf78c780d190e
3
+ size 4991496832
model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86a8abc329b01b2efb1466612bb02f1d47b670dabfb3172985f6d279d7c941bb
3
+ size 4991496936
model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f05899d906cbb2b39a303d812cc017210cf16579638041324026048ee93d0e9
3
+ size 2895740064
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 300,
3
+ "dither": 0.0,
4
+ "feature_extractor_type": "WhisperFeatureExtractor",
5
+ "feature_size": 128,
6
+ "hop_length": 160,
7
+ "image_mean": [
8
+ 0.48145466,
9
+ 0.4578275,
10
+ 0.40821073
11
+ ],
12
+ "image_processor_type": "Qwen2VLImageProcessor",
13
+ "image_std": [
14
+ 0.26862954,
15
+ 0.26130258,
16
+ 0.27577711
17
+ ],
18
+ "max_pixels": 12845056,
19
+ "merge_size": 2,
20
+ "min_pixels": 3136,
21
+ "n_fft": 400,
22
+ "n_samples": 4800000,
23
+ "nb_max_frames": 30000,
24
+ "padding_side": "right",
25
+ "padding_value": 0.0,
26
+ "patch_size": 14,
27
+ "processor_class": "Qwen2_5OmniProcessor",
28
+ "return_attention_mask": true,
29
+ "sampling_rate": 16000,
30
+ "temporal_patch_size": 2
31
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|AUDIO|>",
6
+ "<|audio_bos|>",
7
+ "<|audio_eos|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_bos|>",
12
+ "<|vision_eos|>",
13
+ "<|vision_pad|>",
14
+ "<|IMAGE|>",
15
+ "<|VIDEO|>"
16
+ ],
17
+ "audio_bos_token": "<|audio_bos|>",
18
+ "audio_eos_token": "<|audio_eos|>",
19
+ "audio_token": "<|AUDIO|>",
20
+ "eos_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "image_token": "<|IMAGE|>",
28
+ "pad_token": {
29
+ "content": "<|endoftext|>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false
34
+ },
35
+ "video_token": "<|VIDEO|>",
36
+ "vision_bos_token": "<|vision_bos|>",
37
+ "vision_eos_token": "<|vision_eos|>"
38
+ }
spk_dict.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a05609b28f5d42b7b748f0f07592545c8f1f6885b9ae8fff64baf56e86b2a18
3
+ size 259544
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8441917e39ae0244e06d704b95b3124795cec478e297f9afac39ba670d7e9d99
3
+ size 11421870
tokenizer_config.json ADDED
@@ -0,0 +1,222 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "151643": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "151644": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "151645": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "151646": {
29
+ "content": "<|AUDIO|>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "151647": {
37
+ "content": "<|audio_bos|>",
38
+ "lstrip": false,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "151648": {
45
+ "content": "<|audio_eos|>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "151649": {
53
+ "content": "<|box_end|>",
54
+ "lstrip": false,
55
+ "normalized": false,
56
+ "rstrip": false,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "151650": {
61
+ "content": "<|quad_start|>",
62
+ "lstrip": false,
63
+ "normalized": false,
64
+ "rstrip": false,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "151651": {
69
+ "content": "<|quad_end|>",
70
+ "lstrip": false,
71
+ "normalized": false,
72
+ "rstrip": false,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "151652": {
77
+ "content": "<|vision_bos|>",
78
+ "lstrip": false,
79
+ "normalized": false,
80
+ "rstrip": false,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "151653": {
85
+ "content": "<|vision_eos|>",
86
+ "lstrip": false,
87
+ "normalized": false,
88
+ "rstrip": false,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "151654": {
93
+ "content": "<|vision_pad|>",
94
+ "lstrip": false,
95
+ "normalized": false,
96
+ "rstrip": false,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "151655": {
101
+ "content": "<|IMAGE|>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "151656": {
109
+ "content": "<|VIDEO|>",
110
+ "lstrip": false,
111
+ "normalized": false,
112
+ "rstrip": false,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "151657": {
117
+ "content": "<tool_call>",
118
+ "lstrip": false,
119
+ "normalized": false,
120
+ "rstrip": false,
121
+ "single_word": false,
122
+ "special": false
123
+ },
124
+ "151658": {
125
+ "content": "</tool_call>",
126
+ "lstrip": false,
127
+ "normalized": false,
128
+ "rstrip": false,
129
+ "single_word": false,
130
+ "special": false
131
+ },
132
+ "151659": {
133
+ "content": "<|fim_prefix|>",
134
+ "lstrip": false,
135
+ "normalized": false,
136
+ "rstrip": false,
137
+ "single_word": false,
138
+ "special": false
139
+ },
140
+ "151660": {
141
+ "content": "<|fim_middle|>",
142
+ "lstrip": false,
143
+ "normalized": false,
144
+ "rstrip": false,
145
+ "single_word": false,
146
+ "special": false
147
+ },
148
+ "151661": {
149
+ "content": "<|fim_suffix|>",
150
+ "lstrip": false,
151
+ "normalized": false,
152
+ "rstrip": false,
153
+ "single_word": false,
154
+ "special": false
155
+ },
156
+ "151662": {
157
+ "content": "<|fim_pad|>",
158
+ "lstrip": false,
159
+ "normalized": false,
160
+ "rstrip": false,
161
+ "single_word": false,
162
+ "special": false
163
+ },
164
+ "151663": {
165
+ "content": "<|repo_name|>",
166
+ "lstrip": false,
167
+ "normalized": false,
168
+ "rstrip": false,
169
+ "single_word": false,
170
+ "special": false
171
+ },
172
+ "151664": {
173
+ "content": "<|file_sep|>",
174
+ "lstrip": false,
175
+ "normalized": false,
176
+ "rstrip": false,
177
+ "single_word": false,
178
+ "special": false
179
+ }
180
+ },
181
+ "additional_special_tokens": [
182
+ "<|im_start|>",
183
+ "<|im_end|>",
184
+ "<|AUDIO|>",
185
+ "<|audio_bos|>",
186
+ "<|audio_eos|>",
187
+ "<|box_end|>",
188
+ "<|quad_start|>",
189
+ "<|quad_end|>",
190
+ "<|vision_bos|>",
191
+ "<|vision_eos|>",
192
+ "<|vision_pad|>",
193
+ "<|IMAGE|>",
194
+ "<|VIDEO|>"
195
+ ],
196
+ "audio_bos_token": "<|audio_bos|>",
197
+ "audio_eos_token": "<|audio_eos|>",
198
+ "audio_token": "<|AUDIO|>",
199
+ "bos_token": null,
200
+ "clean_up_tokenization_spaces": false,
201
+ "eos_token": "<|im_end|>",
202
+ "errors": "replace",
203
+ "extra_special_tokens": {
204
+ "audio_bos_token": "<|audio_bos|>",
205
+ "audio_eos_token": "<|audio_eos|>",
206
+ "audio_token": "<|AUDIO|>",
207
+ "image_token": "<|IMAGE|>",
208
+ "video_token": "<|VIDEO|>",
209
+ "vision_bos_token": "<|vision_bos|>",
210
+ "vision_eos_token": "<|vision_eos|>"
211
+ },
212
+ "image_token": "<|IMAGE|>",
213
+ "model_max_length": 32768,
214
+ "pad_token": "<|endoftext|>",
215
+ "processor_class": "Qwen2_5OmniProcessor",
216
+ "split_special_tokens": false,
217
+ "tokenizer_class": "Qwen2Tokenizer",
218
+ "unk_token": null,
219
+ "video_token": "<|VIDEO|>",
220
+ "vision_bos_token": "<|vision_bos|>",
221
+ "vision_eos_token": "<|vision_eos|>"
222
+ }
trainer_state.json ADDED
@@ -0,0 +1,1354 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 200.0,
7
+ "global_step": 1641,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0006093845216331506,
14
+ "grad_norm": 5.139511599456721,
15
+ "learning_rate": 2.0000000000000002e-07,
16
+ "loss": 0.97658371925354,
17
+ "step": 1,
18
+ "token_acc": 0.752757254524039
19
+ },
20
+ {
21
+ "epoch": 0.006093845216331505,
22
+ "grad_norm": 4.02627916602829,
23
+ "learning_rate": 2.0000000000000003e-06,
24
+ "loss": 0.979477776421441,
25
+ "step": 10,
26
+ "token_acc": 0.7513394629500156
27
+ },
28
+ {
29
+ "epoch": 0.01218769043266301,
30
+ "grad_norm": 2.133243665962445,
31
+ "learning_rate": 4.000000000000001e-06,
32
+ "loss": 0.8085936546325684,
33
+ "step": 20,
34
+ "token_acc": 0.7841330046869243
35
+ },
36
+ {
37
+ "epoch": 0.018281535648994516,
38
+ "grad_norm": 1.6130698467539089,
39
+ "learning_rate": 6e-06,
40
+ "loss": 0.6815763473510742,
41
+ "step": 30,
42
+ "token_acc": 0.8088598674958689
43
+ },
44
+ {
45
+ "epoch": 0.02437538086532602,
46
+ "grad_norm": 1.5958993846193437,
47
+ "learning_rate": 8.000000000000001e-06,
48
+ "loss": 0.6228148937225342,
49
+ "step": 40,
50
+ "token_acc": 0.820709892041175
51
+ },
52
+ {
53
+ "epoch": 0.030469226081657527,
54
+ "grad_norm": 1.5835525047156958,
55
+ "learning_rate": 1e-05,
56
+ "loss": 0.5660243034362793,
57
+ "step": 50,
58
+ "token_acc": 0.8334044587872124
59
+ },
60
+ {
61
+ "epoch": 0.03656307129798903,
62
+ "grad_norm": 1.486912724454441,
63
+ "learning_rate": 9.999025267866269e-06,
64
+ "loss": 0.5425637722015381,
65
+ "step": 60,
66
+ "token_acc": 0.8377751665687463
67
+ },
68
+ {
69
+ "epoch": 0.042656916514320534,
70
+ "grad_norm": 1.5337888751937572,
71
+ "learning_rate": 9.996101451506166e-06,
72
+ "loss": 0.5277560710906982,
73
+ "step": 70,
74
+ "token_acc": 0.8436990187965481
75
+ },
76
+ {
77
+ "epoch": 0.04875076173065204,
78
+ "grad_norm": 1.3940886051095893,
79
+ "learning_rate": 9.991229690894796e-06,
80
+ "loss": 0.5164490699768066,
81
+ "step": 80,
82
+ "token_acc": 0.843745632858913
83
+ },
84
+ {
85
+ "epoch": 0.054844606946983544,
86
+ "grad_norm": 1.732808941800786,
87
+ "learning_rate": 9.984411885496807e-06,
88
+ "loss": 0.5111546516418457,
89
+ "step": 90,
90
+ "token_acc": 0.8453230842547292
91
+ },
92
+ {
93
+ "epoch": 0.06093845216331505,
94
+ "grad_norm": 1.502010177161109,
95
+ "learning_rate": 9.975650693525798e-06,
96
+ "loss": 0.5041120052337646,
97
+ "step": 100,
98
+ "token_acc": 0.8452364415692656
99
+ },
100
+ {
101
+ "epoch": 0.06703229737964655,
102
+ "grad_norm": 1.4601020474033113,
103
+ "learning_rate": 9.964949530907907e-06,
104
+ "loss": 0.5016684532165527,
105
+ "step": 110,
106
+ "token_acc": 0.8476517417815049
107
+ },
108
+ {
109
+ "epoch": 0.07312614259597806,
110
+ "grad_norm": 1.2952511202685608,
111
+ "learning_rate": 9.952312569949963e-06,
112
+ "loss": 0.481311559677124,
113
+ "step": 120,
114
+ "token_acc": 0.8533387907153323
115
+ },
116
+ {
117
+ "epoch": 0.07921998781230957,
118
+ "grad_norm": 1.4971652792627679,
119
+ "learning_rate": 9.937744737712734e-06,
120
+ "loss": 0.4774615287780762,
121
+ "step": 130,
122
+ "token_acc": 0.8534425009304056
123
+ },
124
+ {
125
+ "epoch": 0.08531383302864107,
126
+ "grad_norm": 1.3779208298780503,
127
+ "learning_rate": 9.921251714089898e-06,
128
+ "loss": 0.4781217575073242,
129
+ "step": 140,
130
+ "token_acc": 0.8540719832383131
131
+ },
132
+ {
133
+ "epoch": 0.09140767824497258,
134
+ "grad_norm": 1.2833076700573953,
135
+ "learning_rate": 9.9028399295935e-06,
136
+ "loss": 0.470335865020752,
137
+ "step": 150,
138
+ "token_acc": 0.855729364137813
139
+ },
140
+ {
141
+ "epoch": 0.09750152346130408,
142
+ "grad_norm": 1.275667403101274,
143
+ "learning_rate": 9.882516562846735e-06,
144
+ "loss": 0.4639917850494385,
145
+ "step": 160,
146
+ "token_acc": 0.856720295350119
147
+ },
148
+ {
149
+ "epoch": 0.1035953686776356,
150
+ "grad_norm": 1.3612747480512406,
151
+ "learning_rate": 9.860289537785058e-06,
152
+ "loss": 0.46750926971435547,
153
+ "step": 170,
154
+ "token_acc": 0.8558511446900368
155
+ },
156
+ {
157
+ "epoch": 0.10968921389396709,
158
+ "grad_norm": 1.415269189793894,
159
+ "learning_rate": 9.83616752056669e-06,
160
+ "loss": 0.4647522926330566,
161
+ "step": 180,
162
+ "token_acc": 0.8561545157582173
163
+ },
164
+ {
165
+ "epoch": 0.1157830591102986,
166
+ "grad_norm": 1.1892931159809241,
167
+ "learning_rate": 9.810159916193763e-06,
168
+ "loss": 0.45995321273803713,
169
+ "step": 190,
170
+ "token_acc": 0.8573448602405447
171
+ },
172
+ {
173
+ "epoch": 0.1218769043266301,
174
+ "grad_norm": 1.3064527572301536,
175
+ "learning_rate": 9.782276864845351e-06,
176
+ "loss": 0.4638189792633057,
177
+ "step": 200,
178
+ "token_acc": 0.8556243509610675
179
+ },
180
+ {
181
+ "epoch": 0.12797074954296161,
182
+ "grad_norm": 1.314500952457575,
183
+ "learning_rate": 9.752529237923914e-06,
184
+ "loss": 0.4490074634552002,
185
+ "step": 210,
186
+ "token_acc": 0.8597137978120026
187
+ },
188
+ {
189
+ "epoch": 0.1340645947592931,
190
+ "grad_norm": 1.1373343380715917,
191
+ "learning_rate": 9.720928633816596e-06,
192
+ "loss": 0.4501980781555176,
193
+ "step": 220,
194
+ "token_acc": 0.859648604947868
195
+ },
196
+ {
197
+ "epoch": 0.14015843997562463,
198
+ "grad_norm": 1.230125123804479,
199
+ "learning_rate": 9.687487373373103e-06,
200
+ "loss": 0.44935040473937987,
201
+ "step": 230,
202
+ "token_acc": 0.8588833095343918
203
+ },
204
+ {
205
+ "epoch": 0.14625228519195613,
206
+ "grad_norm": 1.6077603667743245,
207
+ "learning_rate": 9.652218495101894e-06,
208
+ "loss": 0.44729223251342776,
209
+ "step": 240,
210
+ "token_acc": 0.85995841942315
211
+ },
212
+ {
213
+ "epoch": 0.15234613040828762,
214
+ "grad_norm": 1.2688011877880512,
215
+ "learning_rate": 9.61513575008656e-06,
216
+ "loss": 0.43803844451904295,
217
+ "step": 250,
218
+ "token_acc": 0.8627407782309685
219
+ },
220
+ {
221
+ "epoch": 0.15843997562461914,
222
+ "grad_norm": 1.178112411608835,
223
+ "learning_rate": 9.576253596624367e-06,
224
+ "loss": 0.43675899505615234,
225
+ "step": 260,
226
+ "token_acc": 0.8637918234089942
227
+ },
228
+ {
229
+ "epoch": 0.16453382084095064,
230
+ "grad_norm": 1.294840110909924,
231
+ "learning_rate": 9.53558719458908e-06,
232
+ "loss": 0.4456604480743408,
233
+ "step": 270,
234
+ "token_acc": 0.8607995996575384
235
+ },
236
+ {
237
+ "epoch": 0.17062766605728213,
238
+ "grad_norm": 1.2275948762976965,
239
+ "learning_rate": 9.49315239952023e-06,
240
+ "loss": 0.44009056091308596,
241
+ "step": 280,
242
+ "token_acc": 0.8624380989923378
243
+ },
244
+ {
245
+ "epoch": 0.17672151127361366,
246
+ "grad_norm": 1.1835184130640346,
247
+ "learning_rate": 9.448965756441154e-06,
248
+ "loss": 0.43228535652160643,
249
+ "step": 290,
250
+ "token_acc": 0.8642425086011517
251
+ },
252
+ {
253
+ "epoch": 0.18281535648994515,
254
+ "grad_norm": 1.0506109771841785,
255
+ "learning_rate": 9.403044493408205e-06,
256
+ "loss": 0.4331789970397949,
257
+ "step": 300,
258
+ "token_acc": 0.8650773124725752
259
+ },
260
+ {
261
+ "epoch": 0.18890920170627665,
262
+ "grad_norm": 1.2538833747077607,
263
+ "learning_rate": 9.355406514793667e-06,
264
+ "loss": 0.44378862380981443,
265
+ "step": 310,
266
+ "token_acc": 0.8612388746191983
267
+ },
268
+ {
269
+ "epoch": 0.19500304692260817,
270
+ "grad_norm": 1.1649745576637627,
271
+ "learning_rate": 9.306070394304955e-06,
272
+ "loss": 0.4216612339019775,
273
+ "step": 320,
274
+ "token_acc": 0.8665872154728236
275
+ },
276
+ {
277
+ "epoch": 0.20109689213893966,
278
+ "grad_norm": 1.1640287810040342,
279
+ "learning_rate": 9.255055367742868e-06,
280
+ "loss": 0.43276224136352537,
281
+ "step": 330,
282
+ "token_acc": 0.864183550146075
283
+ },
284
+ {
285
+ "epoch": 0.2071907373552712,
286
+ "grad_norm": 1.1205258611684763,
287
+ "learning_rate": 9.202381325501683e-06,
288
+ "loss": 0.42910175323486327,
289
+ "step": 340,
290
+ "token_acc": 0.8651370039640893
291
+ },
292
+ {
293
+ "epoch": 0.21328458257160268,
294
+ "grad_norm": 1.1044569461174318,
295
+ "learning_rate": 9.148068804814032e-06,
296
+ "loss": 0.425107479095459,
297
+ "step": 350,
298
+ "token_acc": 0.8655852823220787
299
+ },
300
+ {
301
+ "epoch": 0.21937842778793418,
302
+ "grad_norm": 1.256804160100686,
303
+ "learning_rate": 9.092138981743588e-06,
304
+ "loss": 0.4197092533111572,
305
+ "step": 360,
306
+ "token_acc": 0.8678029564108461
307
+ },
308
+ {
309
+ "epoch": 0.2254722730042657,
310
+ "grad_norm": 1.219596805077906,
311
+ "learning_rate": 9.034613662928665e-06,
312
+ "loss": 0.4218160629272461,
313
+ "step": 370,
314
+ "token_acc": 0.8669598748703018
315
+ },
316
+ {
317
+ "epoch": 0.2315661182205972,
318
+ "grad_norm": 1.0651000148469036,
319
+ "learning_rate": 8.975515277079961e-06,
320
+ "loss": 0.4222999095916748,
321
+ "step": 380,
322
+ "token_acc": 0.8668562219942147
323
+ },
324
+ {
325
+ "epoch": 0.2376599634369287,
326
+ "grad_norm": 1.1478789808745513,
327
+ "learning_rate": 8.91486686623577e-06,
328
+ "loss": 0.41972966194152833,
329
+ "step": 390,
330
+ "token_acc": 0.8667315262188772
331
+ },
332
+ {
333
+ "epoch": 0.2437538086532602,
334
+ "grad_norm": 1.0060533858058822,
335
+ "learning_rate": 8.85269207677806e-06,
336
+ "loss": 0.4143358707427979,
337
+ "step": 400,
338
+ "token_acc": 0.8689943563130941
339
+ },
340
+ {
341
+ "epoch": 0.2498476538695917,
342
+ "grad_norm": 1.2219261737292129,
343
+ "learning_rate": 8.789015150212907e-06,
344
+ "loss": 0.41486186981201173,
345
+ "step": 410,
346
+ "token_acc": 0.867653374528066
347
+ },
348
+ {
349
+ "epoch": 0.25594149908592323,
350
+ "grad_norm": 1.2842286146778168,
351
+ "learning_rate": 8.72386091371891e-06,
352
+ "loss": 0.4264723777770996,
353
+ "step": 420,
354
+ "token_acc": 0.865345114787771
355
+ },
356
+ {
357
+ "epoch": 0.2620353443022547,
358
+ "grad_norm": 1.0550559155752623,
359
+ "learning_rate": 8.657254770467252e-06,
360
+ "loss": 0.40860881805419924,
361
+ "step": 430,
362
+ "token_acc": 0.8694270527928576
363
+ },
364
+ {
365
+ "epoch": 0.2681291895185862,
366
+ "grad_norm": 1.1246909396790437,
367
+ "learning_rate": 8.58922268971719e-06,
368
+ "loss": 0.4148720264434814,
369
+ "step": 440,
370
+ "token_acc": 0.86849521403236
371
+ },
372
+ {
373
+ "epoch": 0.2742230347349177,
374
+ "grad_norm": 1.1877909840033853,
375
+ "learning_rate": 8.51979119669081e-06,
376
+ "loss": 0.4155715465545654,
377
+ "step": 450,
378
+ "token_acc": 0.8686248236499153
379
+ },
380
+ {
381
+ "epoch": 0.28031687995124926,
382
+ "grad_norm": 1.03946599413896,
383
+ "learning_rate": 8.448987362231054e-06,
384
+ "loss": 0.4156056880950928,
385
+ "step": 460,
386
+ "token_acc": 0.8682606492506055
387
+ },
388
+ {
389
+ "epoch": 0.28641072516758076,
390
+ "grad_norm": 1.1045440790462375,
391
+ "learning_rate": 8.376838792246978e-06,
392
+ "loss": 0.41259098052978516,
393
+ "step": 470,
394
+ "token_acc": 0.868615067345492
395
+ },
396
+ {
397
+ "epoch": 0.29250457038391225,
398
+ "grad_norm": 1.1044055109636997,
399
+ "learning_rate": 8.303373616950408e-06,
400
+ "loss": 0.41626744270324706,
401
+ "step": 480,
402
+ "token_acc": 0.867445116993405
403
+ },
404
+ {
405
+ "epoch": 0.29859841560024375,
406
+ "grad_norm": 1.0612884186160958,
407
+ "learning_rate": 8.228620479888172e-06,
408
+ "loss": 0.4087618350982666,
409
+ "step": 490,
410
+ "token_acc": 0.869433255622514
411
+ },
412
+ {
413
+ "epoch": 0.30469226081657524,
414
+ "grad_norm": 1.079879116921211,
415
+ "learning_rate": 8.152608526774188e-06,
416
+ "loss": 0.40863656997680664,
417
+ "step": 500,
418
+ "token_acc": 0.8705444341829626
419
+ },
420
+ {
421
+ "epoch": 0.31078610603290674,
422
+ "grad_norm": 1.0470334273877924,
423
+ "learning_rate": 8.075367394125755e-06,
424
+ "loss": 0.41130657196044923,
425
+ "step": 510,
426
+ "token_acc": 0.8699947913802195
427
+ },
428
+ {
429
+ "epoch": 0.3168799512492383,
430
+ "grad_norm": 1.2778777056879977,
431
+ "learning_rate": 7.996927197708486e-06,
432
+ "loss": 0.4074504852294922,
433
+ "step": 520,
434
+ "token_acc": 0.8711178129454153
435
+ },
436
+ {
437
+ "epoch": 0.3229737964655698,
438
+ "grad_norm": 1.133795250933889,
439
+ "learning_rate": 7.917318520794395e-06,
440
+ "loss": 0.4040180206298828,
441
+ "step": 530,
442
+ "token_acc": 0.8719991647774729
443
+ },
444
+ {
445
+ "epoch": 0.3290676416819013,
446
+ "grad_norm": 1.1320221274981666,
447
+ "learning_rate": 7.836572402237683e-06,
448
+ "loss": 0.4074112892150879,
449
+ "step": 540,
450
+ "token_acc": 0.8696679374619692
451
+ },
452
+ {
453
+ "epoch": 0.3351614868982328,
454
+ "grad_norm": 1.0153565229717176,
455
+ "learning_rate": 7.754720324372924e-06,
456
+ "loss": 0.4030743598937988,
457
+ "step": 550,
458
+ "token_acc": 0.8720831783254012
459
+ },
460
+ {
461
+ "epoch": 0.34125533211456427,
462
+ "grad_norm": 1.0985579621580885,
463
+ "learning_rate": 7.67179420074032e-06,
464
+ "loss": 0.3988363742828369,
465
+ "step": 560,
466
+ "token_acc": 0.8726780258889484
467
+ },
468
+ {
469
+ "epoch": 0.3473491773308958,
470
+ "grad_norm": 1.0584699143582574,
471
+ "learning_rate": 7.587826363642845e-06,
472
+ "loss": 0.4028042793273926,
473
+ "step": 570,
474
+ "token_acc": 0.8709437860238254
475
+ },
476
+ {
477
+ "epoch": 0.3534430225472273,
478
+ "grad_norm": 1.1632651891282637,
479
+ "learning_rate": 7.502849551540106e-06,
480
+ "loss": 0.3974143028259277,
481
+ "step": 580,
482
+ "token_acc": 0.8732772418431721
483
+ },
484
+ {
485
+ "epoch": 0.3595368677635588,
486
+ "grad_norm": 0.9585380945132779,
487
+ "learning_rate": 7.4168968962838524e-06,
488
+ "loss": 0.40021185874938964,
489
+ "step": 590,
490
+ "token_acc": 0.8715715660830257
491
+ },
492
+ {
493
+ "epoch": 0.3656307129798903,
494
+ "grad_norm": 0.939779800665415,
495
+ "learning_rate": 7.330001910200111e-06,
496
+ "loss": 0.39843976497650146,
497
+ "step": 600,
498
+ "token_acc": 0.8733910783350537
499
+ },
500
+ {
501
+ "epoch": 0.3717245581962218,
502
+ "grad_norm": 0.9815164073943617,
503
+ "learning_rate": 7.242198473022958e-06,
504
+ "loss": 0.3972899913787842,
505
+ "step": 610,
506
+ "token_acc": 0.8731910420095998
507
+ },
508
+ {
509
+ "epoch": 0.3778184034125533,
510
+ "grad_norm": 1.0569386302509218,
511
+ "learning_rate": 7.15352081868506e-06,
512
+ "loss": 0.4026960372924805,
513
+ "step": 620,
514
+ "token_acc": 0.8716591305210795
515
+ },
516
+ {
517
+ "epoch": 0.38391224862888484,
518
+ "grad_norm": 1.0897077358900225,
519
+ "learning_rate": 7.0640035219701085e-06,
520
+ "loss": 0.39238433837890624,
521
+ "step": 630,
522
+ "token_acc": 0.8741110700683207
523
+ },
524
+ {
525
+ "epoch": 0.39000609384521634,
526
+ "grad_norm": 1.0094259905078886,
527
+ "learning_rate": 6.973681485032359e-06,
528
+ "loss": 0.3934662342071533,
529
+ "step": 640,
530
+ "token_acc": 0.874180305698641
531
+ },
532
+ {
533
+ "epoch": 0.39609993906154783,
534
+ "grad_norm": 0.9880095870102604,
535
+ "learning_rate": 6.8825899237885215e-06,
536
+ "loss": 0.3929059743881226,
537
+ "step": 650,
538
+ "token_acc": 0.873847849697677
539
+ },
540
+ {
541
+ "epoch": 0.40219378427787933,
542
+ "grad_norm": 0.9583618057687778,
543
+ "learning_rate": 6.7907643541873446e-06,
544
+ "loss": 0.38638834953308104,
545
+ "step": 660,
546
+ "token_acc": 0.8764517709444076
547
+ },
548
+ {
549
+ "epoch": 0.4082876294942108,
550
+ "grad_norm": 1.1091462631909463,
551
+ "learning_rate": 6.698240578362179e-06,
552
+ "loss": 0.3935162782669067,
553
+ "step": 670,
554
+ "token_acc": 0.8743182876186542
555
+ },
556
+ {
557
+ "epoch": 0.4143814747105424,
558
+ "grad_norm": 0.959273015275344,
559
+ "learning_rate": 6.6050546706719984e-06,
560
+ "loss": 0.38172011375427245,
561
+ "step": 680,
562
+ "token_acc": 0.8772576395099669
563
+ },
564
+ {
565
+ "epoch": 0.42047531992687387,
566
+ "grad_norm": 1.0010757728338364,
567
+ "learning_rate": 6.511242963636257e-06,
568
+ "loss": 0.3927836179733276,
569
+ "step": 690,
570
+ "token_acc": 0.8740263817041508
571
+ },
572
+ {
573
+ "epoch": 0.42656916514320536,
574
+ "grad_norm": 1.045230237684538,
575
+ "learning_rate": 6.416842033769106e-06,
576
+ "loss": 0.38949809074401853,
577
+ "step": 700,
578
+ "token_acc": 0.8748742675586352
579
+ },
580
+ {
581
+ "epoch": 0.43266301035953686,
582
+ "grad_norm": 0.9849032327305663,
583
+ "learning_rate": 6.321888687318457e-06,
584
+ "loss": 0.39299988746643066,
585
+ "step": 710,
586
+ "token_acc": 0.8744398373706392
587
+ },
588
+ {
589
+ "epoch": 0.43875685557586835,
590
+ "grad_norm": 0.9773426657855283,
591
+ "learning_rate": 6.2264199459155105e-06,
592
+ "loss": 0.38987624645233154,
593
+ "step": 720,
594
+ "token_acc": 0.8749521585172907
595
+ },
596
+ {
597
+ "epoch": 0.4448507007921999,
598
+ "grad_norm": 1.037517468712357,
599
+ "learning_rate": 6.130473032140272e-06,
600
+ "loss": 0.38550682067871095,
601
+ "step": 730,
602
+ "token_acc": 0.8752092114104209
603
+ },
604
+ {
605
+ "epoch": 0.4509445460085314,
606
+ "grad_norm": 1.0310013780608072,
607
+ "learning_rate": 6.0340853550087345e-06,
608
+ "loss": 0.378936243057251,
609
+ "step": 740,
610
+ "token_acc": 0.878043851367452
611
+ },
612
+ {
613
+ "epoch": 0.4570383912248629,
614
+ "grad_norm": 0.8055934899750623,
615
+ "learning_rate": 5.937294495387377e-06,
616
+ "loss": 0.38777313232421873,
617
+ "step": 750,
618
+ "token_acc": 0.8762303990063655
619
+ },
620
+ {
621
+ "epoch": 0.4631322364411944,
622
+ "grad_norm": 1.0076731680308868,
623
+ "learning_rate": 5.840138191340651e-06,
624
+ "loss": 0.3867051601409912,
625
+ "step": 760,
626
+ "token_acc": 0.875447200037364
627
+ },
628
+ {
629
+ "epoch": 0.4692260816575259,
630
+ "grad_norm": 0.9392775195574543,
631
+ "learning_rate": 5.7426543234171736e-06,
632
+ "loss": 0.3799318552017212,
633
+ "step": 770,
634
+ "token_acc": 0.8780739671196323
635
+ },
636
+ {
637
+ "epoch": 0.4753199268738574,
638
+ "grad_norm": 0.9059297874010275,
639
+ "learning_rate": 5.644880899880382e-06,
640
+ "loss": 0.38845138549804686,
641
+ "step": 780,
642
+ "token_acc": 0.8756513846485855
643
+ },
644
+ {
645
+ "epoch": 0.48141377209018893,
646
+ "grad_norm": 1.0364591251718924,
647
+ "learning_rate": 5.546856041889374e-06,
648
+ "loss": 0.384658670425415,
649
+ "step": 790,
650
+ "token_acc": 0.8760285406658391
651
+ },
652
+ {
653
+ "epoch": 0.4875076173065204,
654
+ "grad_norm": 0.9573686942596932,
655
+ "learning_rate": 5.448617968635741e-06,
656
+ "loss": 0.3791942596435547,
657
+ "step": 800,
658
+ "token_acc": 0.8779162415307187
659
+ },
660
+ {
661
+ "epoch": 0.4936014625228519,
662
+ "grad_norm": 0.9636242802763855,
663
+ "learning_rate": 5.35020498244219e-06,
664
+ "loss": 0.37176291942596434,
665
+ "step": 810,
666
+ "token_acc": 0.8793090876456928
667
+ },
668
+ {
669
+ "epoch": 0.4996953077391834,
670
+ "grad_norm": 1.037660587481492,
671
+ "learning_rate": 5.251655453828728e-06,
672
+ "loss": 0.37394251823425295,
673
+ "step": 820,
674
+ "token_acc": 0.8786210190654307
675
+ },
676
+ {
677
+ "epoch": 0.505789152955515,
678
+ "grad_norm": 1.0719330406024963,
679
+ "learning_rate": 5.153007806552275e-06,
680
+ "loss": 0.3745760679244995,
681
+ "step": 830,
682
+ "token_acc": 0.8784241641412887
683
+ },
684
+ {
685
+ "epoch": 0.5118829981718465,
686
+ "grad_norm": 0.8899515496236061,
687
+ "learning_rate": 5.054300502625517e-06,
688
+ "loss": 0.3706503868103027,
689
+ "step": 840,
690
+ "token_acc": 0.8798184912767585
691
+ },
692
+ {
693
+ "epoch": 0.517976843388178,
694
+ "grad_norm": 0.9136772226114551,
695
+ "learning_rate": 4.9555720273208475e-06,
696
+ "loss": 0.3767611742019653,
697
+ "step": 850,
698
+ "token_acc": 0.8780427238279765
699
+ },
700
+ {
701
+ "epoch": 0.5240706886045094,
702
+ "grad_norm": 0.9760538746168989,
703
+ "learning_rate": 4.856860874165218e-06,
704
+ "loss": 0.37979438304901125,
705
+ "step": 860,
706
+ "token_acc": 0.8784071947906439
707
+ },
708
+ {
709
+ "epoch": 0.5301645338208409,
710
+ "grad_norm": 0.9424993647974058,
711
+ "learning_rate": 4.758205529931808e-06,
712
+ "loss": 0.3839302062988281,
713
+ "step": 870,
714
+ "token_acc": 0.8770481761661205
715
+ },
716
+ {
717
+ "epoch": 0.5362583790371724,
718
+ "grad_norm": 1.0293112779306877,
719
+ "learning_rate": 4.659644459634293e-06,
720
+ "loss": 0.3767723321914673,
721
+ "step": 880,
722
+ "token_acc": 0.8782181679486365
723
+ },
724
+ {
725
+ "epoch": 0.5423522242535039,
726
+ "grad_norm": 1.0743397927299763,
727
+ "learning_rate": 4.56121609152961e-06,
728
+ "loss": 0.3791919946670532,
729
+ "step": 890,
730
+ "token_acc": 0.8769342677312787
731
+ },
732
+ {
733
+ "epoch": 0.5484460694698354,
734
+ "grad_norm": 0.8651643017417293,
735
+ "learning_rate": 4.462958802135069e-06,
736
+ "loss": 0.36331801414489745,
737
+ "step": 900,
738
+ "token_acc": 0.8819762679763837
739
+ },
740
+ {
741
+ "epoch": 0.5545399146861669,
742
+ "grad_norm": 0.9197439306994798,
743
+ "learning_rate": 4.364910901265607e-06,
744
+ "loss": 0.3720353603363037,
745
+ "step": 910,
746
+ "token_acc": 0.8795370329732339
747
+ },
748
+ {
749
+ "epoch": 0.5606337599024985,
750
+ "grad_norm": 0.9973864478854872,
751
+ "learning_rate": 4.2671106170970734e-06,
752
+ "loss": 0.37818198204040526,
753
+ "step": 920,
754
+ "token_acc": 0.8787091854009224
755
+ },
756
+ {
757
+ "epoch": 0.56672760511883,
758
+ "grad_norm": 0.9979320322546561,
759
+ "learning_rate": 4.169596081261332e-06,
760
+ "loss": 0.368232798576355,
761
+ "step": 930,
762
+ "token_acc": 0.8808049967885766
763
+ },
764
+ {
765
+ "epoch": 0.5728214503351615,
766
+ "grad_norm": 0.9817455772913783,
767
+ "learning_rate": 4.072405313979021e-06,
768
+ "loss": 0.37091827392578125,
769
+ "step": 940,
770
+ "token_acc": 0.8796466097957818
771
+ },
772
+ {
773
+ "epoch": 0.578915295551493,
774
+ "grad_norm": 1.0935297334377472,
775
+ "learning_rate": 3.975576209235726e-06,
776
+ "loss": 0.3674028396606445,
777
+ "step": 950,
778
+ "token_acc": 0.8807917695163083
779
+ },
780
+ {
781
+ "epoch": 0.5850091407678245,
782
+ "grad_norm": 0.9835469765967159,
783
+ "learning_rate": 3.879146520007399e-06,
784
+ "loss": 0.3728478908538818,
785
+ "step": 960,
786
+ "token_acc": 0.8795413152600885
787
+ },
788
+ {
789
+ "epoch": 0.591102985984156,
790
+ "grad_norm": 0.9625183356689964,
791
+ "learning_rate": 3.7831538435407344e-06,
792
+ "loss": 0.37494525909423826,
793
+ "step": 970,
794
+ "token_acc": 0.8792245580635571
795
+ },
796
+ {
797
+ "epoch": 0.5971968312004875,
798
+ "grad_norm": 0.9012795424730173,
799
+ "learning_rate": 3.687635606694271e-06,
800
+ "loss": 0.3702352046966553,
801
+ "step": 980,
802
+ "token_acc": 0.8801223453080008
803
+ },
804
+ {
805
+ "epoch": 0.603290676416819,
806
+ "grad_norm": 0.9782757486531443,
807
+ "learning_rate": 3.592629051345936e-06,
808
+ "loss": 0.3673159837722778,
809
+ "step": 990,
810
+ "token_acc": 0.8810825035648933
811
+ },
812
+ {
813
+ "epoch": 0.6093845216331505,
814
+ "grad_norm": 1.0059100640922563,
815
+ "learning_rate": 3.4981712198726956e-06,
816
+ "loss": 0.3642214059829712,
817
+ "step": 1000,
818
+ "token_acc": 0.8818312088488447
819
+ },
820
+ {
821
+ "epoch": 0.615478366849482,
822
+ "grad_norm": 0.9395189399708234,
823
+ "learning_rate": 3.4042989407079986e-06,
824
+ "loss": 0.3784639358520508,
825
+ "step": 1010,
826
+ "token_acc": 0.8780194366406157
827
+ },
828
+ {
829
+ "epoch": 0.6215722120658135,
830
+ "grad_norm": 1.0425592930772825,
831
+ "learning_rate": 3.311048813982627e-06,
832
+ "loss": 0.36695384979248047,
833
+ "step": 1020,
834
+ "token_acc": 0.8809777292779815
835
+ },
836
+ {
837
+ "epoch": 0.6276660572821451,
838
+ "grad_norm": 0.9146308056797927,
839
+ "learning_rate": 3.218457197254583e-06,
840
+ "loss": 0.36698212623596194,
841
+ "step": 1030,
842
+ "token_acc": 0.8810339710207495
843
+ },
844
+ {
845
+ "epoch": 0.6337599024984766,
846
+ "grad_norm": 0.976263078958663,
847
+ "learning_rate": 3.1265601913335196e-06,
848
+ "loss": 0.365465784072876,
849
+ "step": 1040,
850
+ "token_acc": 0.8814162812670944
851
+ },
852
+ {
853
+ "epoch": 0.6398537477148081,
854
+ "grad_norm": 1.0567379406046713,
855
+ "learning_rate": 3.035393626205306e-06,
856
+ "loss": 0.3610874891281128,
857
+ "step": 1050,
858
+ "token_acc": 0.8824792140002385
859
+ },
860
+ {
861
+ "epoch": 0.6459475929311396,
862
+ "grad_norm": 1.0205537815943757,
863
+ "learning_rate": 2.944993047062161e-06,
864
+ "loss": 0.35759830474853516,
865
+ "step": 1060,
866
+ "token_acc": 0.8834624031976018
867
+ },
868
+ {
869
+ "epoch": 0.6520414381474711,
870
+ "grad_norm": 1.0280714401242652,
871
+ "learning_rate": 2.8553937004438425e-06,
872
+ "loss": 0.3574142217636108,
873
+ "step": 1070,
874
+ "token_acc": 0.884169503378651
875
+ },
876
+ {
877
+ "epoch": 0.6581352833638026,
878
+ "grad_norm": 1.0187298702407688,
879
+ "learning_rate": 2.766630520495277e-06,
880
+ "loss": 0.36029987335205077,
881
+ "step": 1080,
882
+ "token_acc": 0.8823869756562952
883
+ },
884
+ {
885
+ "epoch": 0.664229128580134,
886
+ "grad_norm": 0.9191494153561297,
887
+ "learning_rate": 2.67873811534598e-06,
888
+ "loss": 0.35897092819213866,
889
+ "step": 1090,
890
+ "token_acc": 0.8827260508533868
891
+ },
892
+ {
893
+ "epoch": 0.6703229737964655,
894
+ "grad_norm": 0.9492740813391064,
895
+ "learning_rate": 2.591750753616596e-06,
896
+ "loss": 0.36168532371520995,
897
+ "step": 1100,
898
+ "token_acc": 0.8825941425209475
899
+ },
900
+ {
901
+ "epoch": 0.676416819012797,
902
+ "grad_norm": 0.9644543574186545,
903
+ "learning_rate": 2.505702351057804e-06,
904
+ "loss": 0.3665107488632202,
905
+ "step": 1110,
906
+ "token_acc": 0.8816928952036972
907
+ },
908
+ {
909
+ "epoch": 0.6825106642291285,
910
+ "grad_norm": 0.9521683470371731,
911
+ "learning_rate": 2.4206264573268174e-06,
912
+ "loss": 0.35790448188781737,
913
+ "step": 1120,
914
+ "token_acc": 0.8832886728694526
915
+ },
916
+ {
917
+ "epoch": 0.68860450944546,
918
+ "grad_norm": 1.0783164983743936,
919
+ "learning_rate": 2.336556242906608e-06,
920
+ "loss": 0.3561516284942627,
921
+ "step": 1130,
922
+ "token_acc": 0.8839432945670233
923
+ },
924
+ {
925
+ "epoch": 0.6946983546617916,
926
+ "grad_norm": 0.9994299291097577,
927
+ "learning_rate": 2.2535244861729707e-06,
928
+ "loss": 0.3557067632675171,
929
+ "step": 1140,
930
+ "token_acc": 0.8837923958883728
931
+ },
932
+ {
933
+ "epoch": 0.7007921998781231,
934
+ "grad_norm": 1.039214819811771,
935
+ "learning_rate": 2.1715635606144653e-06,
936
+ "loss": 0.3563429832458496,
937
+ "step": 1150,
938
+ "token_acc": 0.8836427544336156
939
+ },
940
+ {
941
+ "epoch": 0.7068860450944546,
942
+ "grad_norm": 0.8549094000634878,
943
+ "learning_rate": 2.0907054222102367e-06,
944
+ "loss": 0.35337374210357664,
945
+ "step": 1160,
946
+ "token_acc": 0.8852147256677358
947
+ },
948
+ {
949
+ "epoch": 0.7129798903107861,
950
+ "grad_norm": 0.894156191232295,
951
+ "learning_rate": 2.0109815969705922e-06,
952
+ "loss": 0.359290337562561,
953
+ "step": 1170,
954
+ "token_acc": 0.8828725266946272
955
+ },
956
+ {
957
+ "epoch": 0.7190737355271176,
958
+ "grad_norm": 0.8673526133846996,
959
+ "learning_rate": 1.9324231686452478e-06,
960
+ "loss": 0.35991313457489016,
961
+ "step": 1180,
962
+ "token_acc": 0.8837700799671174
963
+ },
964
+ {
965
+ "epoch": 0.7251675807434491,
966
+ "grad_norm": 0.9356232121590031,
967
+ "learning_rate": 1.8550607666039877e-06,
968
+ "loss": 0.3538203716278076,
969
+ "step": 1190,
970
+ "token_acc": 0.8850202284200351
971
+ },
972
+ {
973
+ "epoch": 0.7312614259597806,
974
+ "grad_norm": 1.0163312252270116,
975
+ "learning_rate": 1.7789245538944971e-06,
976
+ "loss": 0.3607466459274292,
977
+ "step": 1200,
978
+ "token_acc": 0.8824661130842316
979
+ },
980
+ {
981
+ "epoch": 0.7373552711761121,
982
+ "grad_norm": 0.8390316456040804,
983
+ "learning_rate": 1.7040442154820036e-06,
984
+ "loss": 0.35505869388580324,
985
+ "step": 1210,
986
+ "token_acc": 0.8845901901507859
987
+ },
988
+ {
989
+ "epoch": 0.7434491163924436,
990
+ "grad_norm": 0.921086850463397,
991
+ "learning_rate": 1.6304489466753237e-06,
992
+ "loss": 0.35682291984558107,
993
+ "step": 1220,
994
+ "token_acc": 0.884017590582417
995
+ },
996
+ {
997
+ "epoch": 0.7495429616087751,
998
+ "grad_norm": 0.8352814372993298,
999
+ "learning_rate": 1.5581674417438143e-06,
1000
+ "loss": 0.3599454164505005,
1001
+ "step": 1230,
1002
+ "token_acc": 0.8830610223076613
1003
+ },
1004
+ {
1005
+ "epoch": 0.7556368068251066,
1006
+ "grad_norm": 0.9561368940432438,
1007
+ "learning_rate": 1.4872278827296855e-06,
1008
+ "loss": 0.3544511079788208,
1009
+ "step": 1240,
1010
+ "token_acc": 0.884971241183666
1011
+ },
1012
+ {
1013
+ "epoch": 0.7617306520414382,
1014
+ "grad_norm": 0.9963256225377098,
1015
+ "learning_rate": 1.417657928460029e-06,
1016
+ "loss": 0.35143122673034666,
1017
+ "step": 1250,
1018
+ "token_acc": 0.8854597977852672
1019
+ },
1020
+ {
1021
+ "epoch": 0.7678244972577697,
1022
+ "grad_norm": 1.0464860200353496,
1023
+ "learning_rate": 1.349484703762834e-06,
1024
+ "loss": 0.3545159101486206,
1025
+ "step": 1260,
1026
+ "token_acc": 0.8848001191868091
1027
+ },
1028
+ {
1029
+ "epoch": 0.7739183424741012,
1030
+ "grad_norm": 0.9553675018651967,
1031
+ "learning_rate": 1.2827347888912057e-06,
1032
+ "loss": 0.3540821552276611,
1033
+ "step": 1270,
1034
+ "token_acc": 0.8845431750704823
1035
+ },
1036
+ {
1037
+ "epoch": 0.7800121876904327,
1038
+ "grad_norm": 0.9171124221466627,
1039
+ "learning_rate": 1.2174342091599277e-06,
1040
+ "loss": 0.3459270477294922,
1041
+ "step": 1280,
1042
+ "token_acc": 0.8876378370255273
1043
+ },
1044
+ {
1045
+ "epoch": 0.7861060329067642,
1046
+ "grad_norm": 0.9897434740336704,
1047
+ "learning_rate": 1.1536084247983626e-06,
1048
+ "loss": 0.3577150821685791,
1049
+ "step": 1290,
1050
+ "token_acc": 0.8842498302783435
1051
+ },
1052
+ {
1053
+ "epoch": 0.7921998781230957,
1054
+ "grad_norm": 0.88979092902762,
1055
+ "learning_rate": 1.0912823210237033e-06,
1056
+ "loss": 0.350811505317688,
1057
+ "step": 1300,
1058
+ "token_acc": 0.8856008373344852
1059
+ },
1060
+ {
1061
+ "epoch": 0.7982937233394272,
1062
+ "grad_norm": 0.9287859784083828,
1063
+ "learning_rate": 1.0304801983383989e-06,
1064
+ "loss": 0.3551754951477051,
1065
+ "step": 1310,
1066
+ "token_acc": 0.8848410538592661
1067
+ },
1068
+ {
1069
+ "epoch": 0.8043875685557587,
1070
+ "grad_norm": 0.8802985747226686,
1071
+ "learning_rate": 9.712257630555589e-07,
1072
+ "loss": 0.35124433040618896,
1073
+ "step": 1320,
1074
+ "token_acc": 0.8857088187898194
1075
+ },
1076
+ {
1077
+ "epoch": 0.8104814137720902,
1078
+ "grad_norm": 0.9867993671885138,
1079
+ "learning_rate": 9.135421180560394e-07,
1080
+ "loss": 0.3533953666687012,
1081
+ "step": 1330,
1082
+ "token_acc": 0.8847630099080603
1083
+ },
1084
+ {
1085
+ "epoch": 0.8165752589884216,
1086
+ "grad_norm": 0.926446790364043,
1087
+ "learning_rate": 8.574517537807897e-07,
1088
+ "loss": 0.345960807800293,
1089
+ "step": 1340,
1090
+ "token_acc": 0.8876687663254338
1091
+ },
1092
+ {
1093
+ "epoch": 0.8226691042047533,
1094
+ "grad_norm": 0.9083144656784397,
1095
+ "learning_rate": 8.029765394619899e-07,
1096
+ "loss": 0.35233092308044434,
1097
+ "step": 1350,
1098
+ "token_acc": 0.8852270821778219
1099
+ },
1100
+ {
1101
+ "epoch": 0.8287629494210847,
1102
+ "grad_norm": 0.8865701179019319,
1103
+ "learning_rate": 7.501377145963939e-07,
1104
+ "loss": 0.35347394943237304,
1105
+ "step": 1360,
1106
+ "token_acc": 0.8848507491917527
1107
+ },
1108
+ {
1109
+ "epoch": 0.8348567946374162,
1110
+ "grad_norm": 0.8797844443235806,
1111
+ "learning_rate": 6.98955880664205e-07,
1112
+ "loss": 0.35233142375946047,
1113
+ "step": 1370,
1114
+ "token_acc": 0.8857494626572902
1115
+ },
1116
+ {
1117
+ "epoch": 0.8409506398537477,
1118
+ "grad_norm": 0.985930499180468,
1119
+ "learning_rate": 6.494509930967019e-07,
1120
+ "loss": 0.3484508991241455,
1121
+ "step": 1380,
1122
+ "token_acc": 0.8862226663569039
1123
+ },
1124
+ {
1125
+ "epoch": 0.8470444850700792,
1126
+ "grad_norm": 0.8385926015490823,
1127
+ "learning_rate": 6.016423534957616e-07,
1128
+ "loss": 0.34513344764709475,
1129
+ "step": 1390,
1130
+ "token_acc": 0.88766630420385
1131
+ },
1132
+ {
1133
+ "epoch": 0.8531383302864107,
1134
+ "grad_norm": 0.9469060182104153,
1135
+ "learning_rate": 5.555486021082979e-07,
1136
+ "loss": 0.3453853130340576,
1137
+ "step": 1400,
1138
+ "token_acc": 0.8872980190401473
1139
+ },
1140
+ {
1141
+ "epoch": 0.8592321755027422,
1142
+ "grad_norm": 1.0619116209691746,
1143
+ "learning_rate": 5.111877105585672e-07,
1144
+ "loss": 0.35715694427490235,
1145
+ "step": 1410,
1146
+ "token_acc": 0.8840584828365589
1147
+ },
1148
+ {
1149
+ "epoch": 0.8653260207190737,
1150
+ "grad_norm": 0.9990471419637711,
1151
+ "learning_rate": 4.6857697484116006e-07,
1152
+ "loss": 0.34844698905944826,
1153
+ "step": 1420,
1154
+ "token_acc": 0.8861595746957418
1155
+ },
1156
+ {
1157
+ "epoch": 0.8714198659354052,
1158
+ "grad_norm": 0.8653174829680845,
1159
+ "learning_rate": 4.277330085774156e-07,
1160
+ "loss": 0.34473817348480223,
1161
+ "step": 1430,
1162
+ "token_acc": 0.8869124712097335
1163
+ },
1164
+ {
1165
+ "epoch": 0.8775137111517367,
1166
+ "grad_norm": 0.978323586867761,
1167
+ "learning_rate": 3.886717365378867e-07,
1168
+ "loss": 0.3523882865905762,
1169
+ "step": 1440,
1170
+ "token_acc": 0.8849034480348013
1171
+ },
1172
+ {
1173
+ "epoch": 0.8836075563680682,
1174
+ "grad_norm": 1.0140389777919647,
1175
+ "learning_rate": 3.5140838843339073e-07,
1176
+ "loss": 0.3476292848587036,
1177
+ "step": 1450,
1178
+ "token_acc": 0.8866329934005767
1179
+ },
1180
+ {
1181
+ "epoch": 0.8897014015843998,
1182
+ "grad_norm": 1.0064657138214737,
1183
+ "learning_rate": 3.159574929770515e-07,
1184
+ "loss": 0.35365211963653564,
1185
+ "step": 1460,
1186
+ "token_acc": 0.8852465385385505
1187
+ },
1188
+ {
1189
+ "epoch": 0.8957952468007313,
1190
+ "grad_norm": 0.9324871915195588,
1191
+ "learning_rate": 2.8233287221965555e-07,
1192
+ "loss": 0.3441819190979004,
1193
+ "step": 1470,
1194
+ "token_acc": 0.8871095878318941
1195
+ },
1196
+ {
1197
+ "epoch": 0.9018890920170628,
1198
+ "grad_norm": 0.9055988245681192,
1199
+ "learning_rate": 2.5054763616053967e-07,
1200
+ "loss": 0.34738845825195314,
1201
+ "step": 1480,
1202
+ "token_acc": 0.8870410481583068
1203
+ },
1204
+ {
1205
+ "epoch": 0.9079829372333943,
1206
+ "grad_norm": 0.8845337059700371,
1207
+ "learning_rate": 2.2061417763608818e-07,
1208
+ "loss": 0.3496507167816162,
1209
+ "step": 1490,
1210
+ "token_acc": 0.8858089991712572
1211
+ },
1212
+ {
1213
+ "epoch": 0.9140767824497258,
1214
+ "grad_norm": 0.8884170981985747,
1215
+ "learning_rate": 1.9254416748786086e-07,
1216
+ "loss": 0.34417023658752444,
1217
+ "step": 1500,
1218
+ "token_acc": 0.8876897324425693
1219
+ },
1220
+ {
1221
+ "epoch": 0.9201706276660573,
1222
+ "grad_norm": 0.991921395955364,
1223
+ "learning_rate": 1.6634855001221195e-07,
1224
+ "loss": 0.3475677490234375,
1225
+ "step": 1510,
1226
+ "token_acc": 0.8866243585461391
1227
+ },
1228
+ {
1229
+ "epoch": 0.9262644728823888,
1230
+ "grad_norm": 0.8822494961130495,
1231
+ "learning_rate": 1.4203753869318882e-07,
1232
+ "loss": 0.35834810733795164,
1233
+ "step": 1520,
1234
+ "token_acc": 0.8836910930175179
1235
+ },
1236
+ {
1237
+ "epoch": 0.9323583180987203,
1238
+ "grad_norm": 1.0007870631310825,
1239
+ "learning_rate": 1.196206122203647e-07,
1240
+ "loss": 0.3498887777328491,
1241
+ "step": 1530,
1242
+ "token_acc": 0.8859136668935295
1243
+ },
1244
+ {
1245
+ "epoch": 0.9384521633150518,
1246
+ "grad_norm": 0.9115641715955437,
1247
+ "learning_rate": 9.910651079316824e-08,
1248
+ "loss": 0.3380606651306152,
1249
+ "step": 1540,
1250
+ "token_acc": 0.8888504997761748
1251
+ },
1252
+ {
1253
+ "epoch": 0.9445460085313833,
1254
+ "grad_norm": 0.9336474945041258,
1255
+ "learning_rate": 8.050323271314331e-08,
1256
+ "loss": 0.34683611392974856,
1257
+ "step": 1550,
1258
+ "token_acc": 0.8867626671565236
1259
+ },
1260
+ {
1261
+ "epoch": 0.9506398537477148,
1262
+ "grad_norm": 0.9153041920993996,
1263
+ "learning_rate": 6.381803126546405e-08,
1264
+ "loss": 0.3438985824584961,
1265
+ "step": 1560,
1266
+ "token_acc": 0.8876278171714178
1267
+ },
1268
+ {
1269
+ "epoch": 0.9567336989640464,
1270
+ "grad_norm": 0.8723491469657118,
1271
+ "learning_rate": 4.9057411890933714e-08,
1272
+ "loss": 0.35089046955108644,
1273
+ "step": 1570,
1274
+ "token_acc": 0.8854520115332541
1275
+ },
1276
+ {
1277
+ "epoch": 0.9628275441803779,
1278
+ "grad_norm": 0.8955372459045878,
1279
+ "learning_rate": 3.622712964956032e-08,
1280
+ "loss": 0.34657576084136965,
1281
+ "step": 1580,
1282
+ "token_acc": 0.8870879211520062
1283
+ },
1284
+ {
1285
+ "epoch": 0.9689213893967094,
1286
+ "grad_norm": 0.9416702515233623,
1287
+ "learning_rate": 2.5332186976697037e-08,
1288
+ "loss": 0.35133283138275145,
1289
+ "step": 1590,
1290
+ "token_acc": 0.8860156117328086
1291
+ },
1292
+ {
1293
+ "epoch": 0.9750152346130408,
1294
+ "grad_norm": 0.8711895076149625,
1295
+ "learning_rate": 1.637683173263238e-08,
1296
+ "loss": 0.35227146148681643,
1297
+ "step": 1600,
1298
+ "token_acc": 0.8855705009128142
1299
+ },
1300
+ {
1301
+ "epoch": 0.9811090798293723,
1302
+ "grad_norm": 0.983729311544991,
1303
+ "learning_rate": 9.364555546375054e-09,
1304
+ "loss": 0.34629082679748535,
1305
+ "step": 1610,
1306
+ "token_acc": 0.8869587094319709
1307
+ },
1308
+ {
1309
+ "epoch": 0.9872029250457038,
1310
+ "grad_norm": 0.94198406227018,
1311
+ "learning_rate": 4.2980924542984634e-09,
1312
+ "loss": 0.3403524875640869,
1313
+ "step": 1620,
1314
+ "token_acc": 0.8887918722020187
1315
+ },
1316
+ {
1317
+ "epoch": 0.9932967702620353,
1318
+ "grad_norm": 0.8500481071531769,
1319
+ "learning_rate": 1.179417834153429e-09,
1320
+ "loss": 0.3546321868896484,
1321
+ "step": 1630,
1322
+ "token_acc": 0.8848728077900511
1323
+ },
1324
+ {
1325
+ "epoch": 0.9993906154783668,
1326
+ "grad_norm": 0.9946340081463461,
1327
+ "learning_rate": 9.74763488759134e-12,
1328
+ "loss": 0.35070624351501467,
1329
+ "step": 1640,
1330
+ "token_acc": 0.8863686895606487
1331
+ }
1332
+ ],
1333
+ "logging_steps": 10,
1334
+ "max_steps": 1641,
1335
+ "num_input_tokens_seen": 0,
1336
+ "num_train_epochs": 1,
1337
+ "save_steps": 200,
1338
+ "stateful_callbacks": {
1339
+ "TrainerControl": {
1340
+ "args": {
1341
+ "should_epoch_stop": false,
1342
+ "should_evaluate": false,
1343
+ "should_log": false,
1344
+ "should_save": true,
1345
+ "should_training_stop": true
1346
+ },
1347
+ "attributes": {}
1348
+ }
1349
+ },
1350
+ "total_flos": 2214001985716224.0,
1351
+ "train_batch_size": 4,
1352
+ "trial_name": null,
1353
+ "trial_params": null
1354
+ }
video_preprocessor_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 300,
3
+ "crop_size": null,
4
+ "data_format": "channels_first",
5
+ "default_to_square": true,
6
+ "device": null,
7
+ "dither": 0.0,
8
+ "do_center_crop": null,
9
+ "do_convert_rgb": true,
10
+ "do_normalize": true,
11
+ "do_rescale": true,
12
+ "do_resize": true,
13
+ "do_sample_frames": false,
14
+ "feature_extractor_type": "WhisperFeatureExtractor",
15
+ "feature_size": 128,
16
+ "fps": null,
17
+ "hop_length": 160,
18
+ "image_mean": [
19
+ 0.48145466,
20
+ 0.4578275,
21
+ 0.40821073
22
+ ],
23
+ "image_std": [
24
+ 0.26862954,
25
+ 0.26130258,
26
+ 0.27577711
27
+ ],
28
+ "input_data_format": null,
29
+ "max_frames": 768,
30
+ "max_pixels": 12845056,
31
+ "merge_size": 2,
32
+ "min_frames": 4,
33
+ "min_pixels": 3136,
34
+ "n_fft": 400,
35
+ "n_samples": 4800000,
36
+ "nb_max_frames": 30000,
37
+ "num_frames": null,
38
+ "pad_size": null,
39
+ "padding_side": "right",
40
+ "padding_value": 0.0,
41
+ "patch_size": 14,
42
+ "processor_class": "Qwen2_5OmniProcessor",
43
+ "resample": 3,
44
+ "rescale_factor": 0.00392156862745098,
45
+ "return_attention_mask": true,
46
+ "return_metadata": false,
47
+ "sampling_rate": 16000,
48
+ "size": {
49
+ "longest_edge": 12845056,
50
+ "shortest_edge": 3136
51
+ },
52
+ "temporal_patch_size": 2,
53
+ "video_metadata": null,
54
+ "video_processor_type": "Qwen2VLVideoProcessor"
55
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff