owaski commited on
Commit
e2656cc
·
verified ·
1 Parent(s): 15dcd4c

Upload folder using huggingface_hub

Browse files
.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
added_tokens.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<tts_pad>": 151671,
9
+ "<tts_text_bos>": 151672,
10
+ "<tts_text_bos_single>": 151674,
11
+ "<tts_text_eod>": 151673,
12
+ "<|audio_end|>": 151670,
13
+ "<|audio_pad|>": 151675,
14
+ "<|audio_start|>": 151669,
15
+ "<|box_end|>": 151649,
16
+ "<|box_start|>": 151648,
17
+ "<|endoftext|>": 151643,
18
+ "<|file_sep|>": 151664,
19
+ "<|fim_middle|>": 151660,
20
+ "<|fim_pad|>": 151662,
21
+ "<|fim_prefix|>": 151659,
22
+ "<|fim_suffix|>": 151661,
23
+ "<|im_end|>": 151645,
24
+ "<|im_start|>": 151644,
25
+ "<|image_pad|>": 151655,
26
+ "<|object_ref_end|>": 151647,
27
+ "<|object_ref_start|>": 151646,
28
+ "<|quad_end|>": 151651,
29
+ "<|quad_start|>": 151650,
30
+ "<|repo_name|>": 151663,
31
+ "<|video_pad|>": 151656,
32
+ "<|vision_end|>": 151653,
33
+ "<|vision_pad|>": 151654,
34
+ "<|vision_start|>": 151652
35
+ }
args.json ADDED
@@ -0,0 +1,480 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct/",
3
+ "model_type": "qwen3_omni",
4
+ "model_revision": null,
5
+ "task_type": "causal_lm",
6
+ "torch_dtype": "bfloat16",
7
+ "attn_impl": null,
8
+ "new_special_tokens": [],
9
+ "num_labels": null,
10
+ "problem_type": null,
11
+ "rope_scaling": {
12
+ "interleaved": true,
13
+ "mrope_interleaved": true,
14
+ "mrope_section": [
15
+ 24,
16
+ 20,
17
+ 20
18
+ ],
19
+ "rope_type": "default",
20
+ "type": "default"
21
+ },
22
+ "device_map": null,
23
+ "max_memory": {},
24
+ "max_model_len": null,
25
+ "local_repo_path": null,
26
+ "init_strategy": null,
27
+ "template": "qwen3_omni",
28
+ "system": null,
29
+ "max_length": 2048,
30
+ "truncation_strategy": "delete",
31
+ "max_pixels": null,
32
+ "agent_template": null,
33
+ "norm_bbox": null,
34
+ "use_chat_template": true,
35
+ "padding_free": true,
36
+ "padding_side": "right",
37
+ "loss_scale": "default",
38
+ "sequence_parallel_size": 1,
39
+ "response_prefix": null,
40
+ "template_backend": "swift",
41
+ "dataset": [
42
+ "/data/group_data/li_lab/siqiouya/datasets/gigaspeech/manifests/train_s_de_v4_ner_baseline_aligned_rate1.0_k20_final.jsonl"
43
+ ],
44
+ "val_dataset": [],
45
+ "split_dataset_ratio": 0.01,
46
+ "data_seed": 42,
47
+ "dataset_num_proc": 8,
48
+ "load_from_cache_file": true,
49
+ "dataset_shuffle": true,
50
+ "val_dataset_shuffle": false,
51
+ "streaming": false,
52
+ "interleave_prob": null,
53
+ "stopping_strategy": "first_exhausted",
54
+ "shuffle_buffer_size": 1000,
55
+ "download_mode": "reuse_dataset_if_exists",
56
+ "columns": {},
57
+ "strict": false,
58
+ "remove_unused_columns": true,
59
+ "model_name": null,
60
+ "model_author": null,
61
+ "custom_dataset_info": [],
62
+ "quant_method": null,
63
+ "quant_bits": null,
64
+ "hqq_axis": null,
65
+ "bnb_4bit_compute_dtype": "bfloat16",
66
+ "bnb_4bit_quant_type": "nf4",
67
+ "bnb_4bit_use_double_quant": true,
68
+ "bnb_4bit_quant_storage": null,
69
+ "max_new_tokens": null,
70
+ "temperature": null,
71
+ "top_k": null,
72
+ "top_p": null,
73
+ "repetition_penalty": null,
74
+ "num_beams": 1,
75
+ "stream": false,
76
+ "stop_words": [],
77
+ "logprobs": false,
78
+ "top_logprobs": null,
79
+ "ckpt_dir": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct-mcore",
80
+ "lora_modules": [],
81
+ "tuner_backend": "peft",
82
+ "train_type": "lora",
83
+ "adapters": [],
84
+ "external_plugins": [],
85
+ "seed": 42,
86
+ "model_kwargs": {},
87
+ "load_args": false,
88
+ "load_data_args": false,
89
+ "packing": true,
90
+ "packing_length": 2048,
91
+ "lazy_tokenize": false,
92
+ "cached_dataset": [],
93
+ "custom_register_path": [],
94
+ "use_hf": false,
95
+ "hub_token": null,
96
+ "ddp_timeout": 18000000,
97
+ "ddp_backend": null,
98
+ "ignore_args_error": false,
99
+ "use_swift_lora": false,
100
+ "freeze_llm": false,
101
+ "freeze_vit": true,
102
+ "freeze_aligner": true,
103
+ "freeze_parameters": [],
104
+ "freeze_parameters_regex": null,
105
+ "freeze_parameters_ratio": 0.0,
106
+ "trainable_parameters": [],
107
+ "trainable_parameters_regex": null,
108
+ "adapter_load": null,
109
+ "target_modules": [
110
+ "all-linear"
111
+ ],
112
+ "target_regex": null,
113
+ "modules_to_save": [],
114
+ "lora_rank": 8,
115
+ "lora_alpha": 32,
116
+ "lora_dropout": 0.05,
117
+ "lora_bias": "none",
118
+ "lora_dtype": null,
119
+ "use_rslora": false,
120
+ "rlhf_type": null,
121
+ "ref_load": null,
122
+ "ref_adapter_load": null,
123
+ "beta": 0.1,
124
+ "rpo_alpha": null,
125
+ "reference_free": false,
126
+ "label_smoothing": 0.0,
127
+ "f_divergence_type": "reverse_kl",
128
+ "loss_type": null,
129
+ "desirable_weight": 1.0,
130
+ "undesirable_weight": 1.0,
131
+ "calculate_KL": null,
132
+ "center_rewards_coefficient": null,
133
+ "padded_vocab_size": 152064,
134
+ "initialize_embedding": false,
135
+ "mlp_padding_free": false,
136
+ "dataloader_persistent_workers": true,
137
+ "dataloader_prefetch_factor": 10,
138
+ "architectures": "Qwen3OmniMoeForConditionalGeneration",
139
+ "llm_architectures": null,
140
+ "max_epochs": 1,
141
+ "enable_dft_loss": false,
142
+ "enable_channel_loss": false,
143
+ "original_max_position_embeddings": null,
144
+ "partial_rotary_factor": null,
145
+ "use_shared_expert_gate": false,
146
+ "vit_gradient_checkpointing": false,
147
+ "gradient_checkpointing_kwargs": null,
148
+ "linear_num_value_heads": null,
149
+ "linear_num_key_heads": null,
150
+ "linear_key_head_dim": null,
151
+ "linear_value_head_dim": null,
152
+ "linear_conv_kernel_dim": null,
153
+ "layer_types": null,
154
+ "mrope_interleaved": true,
155
+ "micro_batch_size": 1,
156
+ "global_batch_size": 4,
157
+ "recompute_granularity": "full",
158
+ "recompute_method": "uniform",
159
+ "recompute_num_layers": 1,
160
+ "recompute_modules": [
161
+ "core_attn"
162
+ ],
163
+ "use_cpu_initialization": false,
164
+ "deterministic_mode": false,
165
+ "train_iters": null,
166
+ "log_interval": 10,
167
+ "tensorboard_dir": "/data/user_data/siqiouya/ckpts/infinisst-omni/gigaspeech-de-s_v4_ner_baseline_aligned_rate1.0_k20_final/v0-20260106-025939/runs",
168
+ "no_masked_softmax_fusion": false,
169
+ "no_bias_dropout_fusion": false,
170
+ "no_bias_swiglu_fusion": false,
171
+ "no_rope_fusion": false,
172
+ "no_gradient_accumulation_fusion": false,
173
+ "cross_entropy_loss_fusion": true,
174
+ "cross_entropy_fusion_impl": "native",
175
+ "calculate_per_token_loss": true,
176
+ "use_flash_attn": false,
177
+ "attention_backend": "flash",
178
+ "optimizer": "adam",
179
+ "optimizer_cpu_offload": false,
180
+ "optimizer_offload_fraction": 1.0,
181
+ "use_precision_aware_optimizer": false,
182
+ "main_grads_dtype": "fp32",
183
+ "main_params_dtype": "fp32",
184
+ "exp_avg_dtype": "fp32",
185
+ "exp_avg_sq_dtype": "fp32",
186
+ "dataloader_type": "cyclic",
187
+ "manual_gc": false,
188
+ "manual_gc_interval": 0,
189
+ "lr": 0.0001,
190
+ "lr_decay_style": "cosine",
191
+ "lr_decay_iters": null,
192
+ "lr_warmup_iters": 0,
193
+ "lr_warmup_fraction": 0.05,
194
+ "min_lr": 1e-05,
195
+ "weight_decay": 0.01,
196
+ "clip_grad": 1.0,
197
+ "adam_beta1": 0.9,
198
+ "adam_beta2": 0.95,
199
+ "adam_eps": 1e-08,
200
+ "sgd_momentum": 0.9,
201
+ "save": "/data/user_data/siqiouya/ckpts/infinisst-omni/gigaspeech-de-s_v4_ner_baseline_aligned_rate1.0_k20_final/v0-20260106-025939",
202
+ "save_interval": 200,
203
+ "no_save_optim": true,
204
+ "no_save_rng": true,
205
+ "load": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct-mcore",
206
+ "no_load_optim": false,
207
+ "no_load_rng": false,
208
+ "finetune": true,
209
+ "ckpt_format": "torch_dist",
210
+ "no_initialization": true,
211
+ "auto_detect_ckpt_format": true,
212
+ "exit_on_missing_checkpoint": true,
213
+ "distributed_backend": "nccl",
214
+ "local_rank": 0,
215
+ "use_distributed_optimizer": true,
216
+ "tensor_model_parallel_size": 1,
217
+ "pipeline_model_parallel_size": 1,
218
+ "decoder_first_pipeline_num_layers": null,
219
+ "decoder_last_pipeline_num_layers": null,
220
+ "sequence_parallel": false,
221
+ "context_parallel_size": 1,
222
+ "tp_comm_overlap": false,
223
+ "overlap_grad_reduce": false,
224
+ "overlap_param_gather": false,
225
+ "distributed_timeout_minutes": 300000,
226
+ "num_layers_per_virtual_pipeline_stage": null,
227
+ "num_virtual_stages_per_pipeline_rank": null,
228
+ "microbatch_group_size_per_virtual_pipeline_stage": null,
229
+ "pipeline_model_parallel_layout": null,
230
+ "num_layers": 48,
231
+ "hidden_size": 2048,
232
+ "ffn_hidden_size": null,
233
+ "num_attention_heads": 32,
234
+ "group_query_attention": true,
235
+ "num_query_groups": 4,
236
+ "max_position_embeddings": 65536,
237
+ "position_embedding_type": "mrope",
238
+ "mrope_section": [
239
+ 24,
240
+ 20,
241
+ 20
242
+ ],
243
+ "rotary_base": 1000000,
244
+ "rotary_percent": 1.0,
245
+ "rotary_interleaved": false,
246
+ "normalization": "RMSNorm",
247
+ "norm_epsilon": 1e-06,
248
+ "swiglu": true,
249
+ "untie_embeddings_and_output_weights": true,
250
+ "disable_bias_linear": true,
251
+ "add_qkv_bias": false,
252
+ "attention_dropout": 0.0,
253
+ "hidden_dropout": 0.0,
254
+ "kv_channels": 128,
255
+ "qk_layernorm": true,
256
+ "transformer_impl": "transformer_engine",
257
+ "num_experts": 128,
258
+ "moe_layer_freq": "1",
259
+ "moe_ffn_hidden_size": 768,
260
+ "moe_shared_expert_intermediate_size": null,
261
+ "moe_router_topk": 8,
262
+ "moe_router_pre_softmax": false,
263
+ "moe_router_dtype": "fp32",
264
+ "moe_router_score_function": "softmax",
265
+ "moe_router_bias_update_rate": 0.001,
266
+ "moe_router_enable_expert_bias": false,
267
+ "moe_router_topk_scaling_factor": null,
268
+ "moe_router_load_balancing_type": "aux_loss",
269
+ "expert_model_parallel_size": 2,
270
+ "expert_tensor_parallel_size": 1,
271
+ "moe_token_dispatcher_type": null,
272
+ "moe_enable_deepep": false,
273
+ "moe_grouped_gemm": true,
274
+ "moe_permute_fusion": true,
275
+ "moe_aux_loss_coeff": 0.001,
276
+ "moe_z_loss_coeff": null,
277
+ "moe_shared_expert_overlap": true,
278
+ "moe_layer_recompute": false,
279
+ "moe_expert_capacity_factor": null,
280
+ "moe_pad_expert_input_to_capacity": false,
281
+ "moe_token_drop_policy": null,
282
+ "multi_latent_attention": false,
283
+ "q_lora_rank": null,
284
+ "kv_lora_rank": 32,
285
+ "qk_head_dim": 128,
286
+ "qk_pos_emb_head_dim": 64,
287
+ "fp8_format": null,
288
+ "fp8_recipe": "delayed",
289
+ "fp8_amax_history_len": 1024,
290
+ "fp8_amax_compute_algo": "max",
291
+ "fp8_param_gather": false,
292
+ "fp16": false,
293
+ "bf16": true,
294
+ "apply_query_key_layer_scaling": false,
295
+ "attention_softmax_in_fp32": true,
296
+ "log_params_norm": false,
297
+ "log_throughput": false,
298
+ "tensorboard_log_interval": 1,
299
+ "tensorboard_queue_size": 50,
300
+ "log_timers_to_tensorboard": true,
301
+ "no_log_learning_rate_to_tensorboard": false,
302
+ "log_validation_ppl_to_tensorboard": true,
303
+ "log_memory_to_tensorboard": true,
304
+ "logging_level": null,
305
+ "wandb_project": "gigaspeech_de",
306
+ "wandb_exp_name": "omni-gigaspeech-de-s_v4_ner_baseline_aligned_rate1.0_k20_final",
307
+ "wandb_save_dir": null,
308
+ "eval_iters": -1,
309
+ "eval_interval": 200,
310
+ "seq_length": 2048,
311
+ "num_workers": 8,
312
+ "extra_megatron_kwargs": {},
313
+ "add_version": true,
314
+ "rank": 0,
315
+ "global_world_size": 2,
316
+ "local_world_size": 2,
317
+ "model_suffix": "Qwen3-Omni-30B-A3B-Instruct",
318
+ "model_info": "ModelInfo(model_type='qwen3_omni', model_dir='/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct', torch_dtype=torch.bfloat16, max_model_len=65536, quant_method=None, quant_bits=None, rope_scaling={'interleaved': True, 'mrope_section': [24, 20, 20], 'rope_type': 'default', 'type': 'default'}, is_moe_model=True, config=None, task_type='causal_lm', num_labels=None)",
319
+ "model_meta": "ModelMeta(model_type='qwen3_omni', model_groups=[ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Instruct', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Thinking', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Thinking', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Captioner', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Captioner', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='qwen3_omni', get_function=<function get_model_tokenizer_qwen3_omni at 0x150f8b44ea20>, model_arch=MultiModelKeys(arch_name='qwen3_omni', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['thinker.model'], aligner=['thinker.audio_tower.proj1', 'thinker.audio_tower.proj2', 'thinker.visual.merger', 'thinker.visual.merger_list'], vision_tower=['thinker.audio_tower', 'thinker.visual'], generator=['talker', 'token2wav']), architectures=['Qwen3OmniMoeForConditionalGeneration'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, task_type=None, ignore_patterns=None, requires=['transformers>=4.57.dev0', 'soundfile', 'decord', 'qwen_omni_utils'], tags=['vision', 'video', 'audio'])",
320
+ "model_dir": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct",
321
+ "hub": "<class 'swift.hub.hub.MSHub'>",
322
+ "megatron_model_meta": "MMGPTMegatronModelMeta(megatron_model_type='qwen3_omni', model_types=['qwen3_omni'], convert_mcore2hf=<function convert_mcore2hf_qwen3_omni at 0x150ef4ae1300>, convert_hf2mcore=<function convert_hf2mcore_qwen3_omni at 0x150ef4ae1260>, model_cls=<class 'swift.megatron.model.mm_gpt.qwen3_vl.Qwen3VLGPTModel'>, convert_hf_config=<function convert_gpt_hf_config at 0x150ef4a899e0>, get_transformer_layer_spec=None, model_provider=<function model_provider at 0x150ef4a887c0>, visual_cls=<class 'swift.megatron.model.mm_gpt.qwen3_vl.Qwen3Omni_Vit'>, extra_args_provider=None)",
323
+ "extra_args": {
324
+ "model": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct/",
325
+ "model_type": "qwen3_omni",
326
+ "model_revision": null,
327
+ "task_type": "causal_lm",
328
+ "torch_dtype": "bfloat16",
329
+ "attn_impl": null,
330
+ "new_special_tokens": [],
331
+ "num_labels": null,
332
+ "problem_type": null,
333
+ "rope_scaling": {
334
+ "interleaved": true,
335
+ "mrope_interleaved": true,
336
+ "mrope_section": [
337
+ 24,
338
+ 20,
339
+ 20
340
+ ],
341
+ "rope_type": "default",
342
+ "type": "default"
343
+ },
344
+ "device_map": null,
345
+ "max_memory": {},
346
+ "max_model_len": null,
347
+ "local_repo_path": null,
348
+ "init_strategy": null,
349
+ "template": "qwen3_omni",
350
+ "system": null,
351
+ "max_length": 2048,
352
+ "truncation_strategy": "delete",
353
+ "max_pixels": null,
354
+ "agent_template": null,
355
+ "norm_bbox": null,
356
+ "use_chat_template": true,
357
+ "padding_free": true,
358
+ "padding_side": "right",
359
+ "sequence_parallel_size": 1,
360
+ "response_prefix": null,
361
+ "template_backend": "swift",
362
+ "dataset": [
363
+ "/data/group_data/li_lab/siqiouya/datasets/gigaspeech/manifests/train_s_de_v4_ner_baseline_aligned_rate1.0_k20_final.jsonl"
364
+ ],
365
+ "val_dataset": [],
366
+ "split_dataset_ratio": 0.01,
367
+ "data_seed": 42,
368
+ "dataset_num_proc": 8,
369
+ "load_from_cache_file": true,
370
+ "dataset_shuffle": true,
371
+ "val_dataset_shuffle": false,
372
+ "streaming": false,
373
+ "interleave_prob": null,
374
+ "stopping_strategy": "first_exhausted",
375
+ "shuffle_buffer_size": 1000,
376
+ "download_mode": "reuse_dataset_if_exists",
377
+ "columns": {},
378
+ "strict": false,
379
+ "remove_unused_columns": true,
380
+ "model_name": null,
381
+ "model_author": null,
382
+ "custom_dataset_info": [],
383
+ "quant_method": null,
384
+ "quant_bits": null,
385
+ "hqq_axis": null,
386
+ "bnb_4bit_compute_dtype": "bfloat16",
387
+ "bnb_4bit_quant_type": "nf4",
388
+ "bnb_4bit_use_double_quant": true,
389
+ "bnb_4bit_quant_storage": null,
390
+ "max_new_tokens": null,
391
+ "temperature": null,
392
+ "top_k": null,
393
+ "top_p": null,
394
+ "repetition_penalty": null,
395
+ "num_beams": 1,
396
+ "stream": false,
397
+ "stop_words": [],
398
+ "logprobs": false,
399
+ "top_logprobs": null,
400
+ "ckpt_dir": "/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct-mcore",
401
+ "lora_modules": [],
402
+ "tuner_backend": "peft",
403
+ "train_type": "lora",
404
+ "adapters": [],
405
+ "external_plugins": [],
406
+ "model_kwargs": {},
407
+ "load_args": false,
408
+ "load_data_args": false,
409
+ "packing": true,
410
+ "packing_length": 2048,
411
+ "lazy_tokenize": false,
412
+ "cached_dataset": [],
413
+ "custom_register_path": [],
414
+ "use_hf": false,
415
+ "hub_token": null,
416
+ "ddp_timeout": 18000000,
417
+ "ddp_backend": null,
418
+ "ignore_args_error": false,
419
+ "use_swift_lora": false,
420
+ "freeze_llm": false,
421
+ "freeze_vit": true,
422
+ "freeze_aligner": true,
423
+ "freeze_parameters": [],
424
+ "freeze_parameters_regex": null,
425
+ "freeze_parameters_ratio": 0.0,
426
+ "trainable_parameters": [],
427
+ "trainable_parameters_regex": null,
428
+ "adapter_load": null,
429
+ "target_modules": [
430
+ "all-linear"
431
+ ],
432
+ "target_regex": null,
433
+ "modules_to_save": [],
434
+ "lora_rank": 8,
435
+ "lora_alpha": 32,
436
+ "lora_dropout": 0.05,
437
+ "lora_bias": "none",
438
+ "lora_dtype": null,
439
+ "use_rslora": false,
440
+ "rlhf_type": null,
441
+ "ref_load": null,
442
+ "ref_adapter_load": null,
443
+ "beta": 0.1,
444
+ "rpo_alpha": null,
445
+ "reference_free": false,
446
+ "label_smoothing": 0.0,
447
+ "f_divergence_type": "reverse_kl",
448
+ "loss_type": null,
449
+ "desirable_weight": 1.0,
450
+ "undesirable_weight": 1.0,
451
+ "calculate_KL": null,
452
+ "center_rewards_coefficient": null,
453
+ "padded_vocab_size": 152064,
454
+ "initialize_embedding": false,
455
+ "mlp_padding_free": false,
456
+ "dataloader_persistent_workers": true,
457
+ "dataloader_prefetch_factor": 10,
458
+ "architectures": "Qwen3OmniMoeForConditionalGeneration",
459
+ "llm_architectures": null,
460
+ "max_epochs": 1,
461
+ "enable_dft_loss": false,
462
+ "enable_channel_loss": false,
463
+ "original_max_position_embeddings": null,
464
+ "partial_rotary_factor": null,
465
+ "use_shared_expert_gate": false,
466
+ "vit_gradient_checkpointing": false,
467
+ "gradient_checkpointing_kwargs": null,
468
+ "linear_num_value_heads": null,
469
+ "linear_num_key_heads": null,
470
+ "linear_key_head_dim": null,
471
+ "linear_value_head_dim": null,
472
+ "linear_conv_kernel_dim": null,
473
+ "layer_types": null,
474
+ "mrope_interleaved": true,
475
+ "add_version": true,
476
+ "model_info": "ModelInfo(model_type='qwen3_omni', model_dir='/data/user_data/siqiouya/ckpts/pretrained/llm/Qwen3-Omni-30B-A3B-Instruct', torch_dtype=torch.bfloat16, max_model_len=65536, quant_method=None, quant_bits=None, rope_scaling={'interleaved': True, 'mrope_section': [24, 20, 20], 'rope_type': 'default', 'type': 'default'}, is_moe_model=True, config=None, task_type='causal_lm', num_labels=None)",
477
+ "model_meta": "ModelMeta(model_type='qwen3_omni', model_groups=[ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Instruct', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Thinking', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Thinking', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3-Omni-30B-A3B-Captioner', hf_model_id='Qwen/Qwen3-Omni-30B-A3B-Captioner', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='qwen3_omni', get_function=<function get_model_tokenizer_qwen3_omni at 0x150f8b44ea20>, model_arch=MultiModelKeys(arch_name='qwen3_omni', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['thinker.model'], aligner=['thinker.audio_tower.proj1', 'thinker.audio_tower.proj2', 'thinker.visual.merger', 'thinker.visual.merger_list'], vision_tower=['thinker.audio_tower', 'thinker.visual'], generator=['talker', 'token2wav']), architectures=['Qwen3OmniMoeForConditionalGeneration'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, task_type=None, ignore_patterns=None, requires=['transformers>=4.57.dev0', 'soundfile', 'decord', 'qwen_omni_utils'], tags=['vision', 'video', 'audio'])",
478
+ "megatron_model_meta": "MMGPTMegatronModelMeta(megatron_model_type='qwen3_omni', model_types=['qwen3_omni'], convert_mcore2hf=<function convert_mcore2hf_qwen3_omni at 0x150ef4ae1300>, convert_hf2mcore=<function convert_hf2mcore_qwen3_omni at 0x150ef4ae1260>, model_cls=<class 'swift.megatron.model.mm_gpt.qwen3_vl.Qwen3VLGPTModel'>, convert_hf_config=<function convert_gpt_hf_config at 0x150ef4a899e0>, get_transformer_layer_spec=None, model_provider=<function model_provider at 0x150ef4a887c0>, visual_cls=<class 'swift.megatron.model.mm_gpt.qwen3_vl.Qwen3Omni_Vit'>, extra_args_provider=None)"
479
+ }
480
+ }
chat_template.jinja ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
9
+ {{- "<|vision_start|><|image_pad|><|vision_end|>" }}
10
+ {%- elif content.type == 'audio' or 'audio' in content or 'audio_url' in content %}
11
+ {{- "<|audio_start|><|audio_pad|><|audio_end|>" }}
12
+ {%- elif content.type == 'video' or 'video' in content %}
13
+ {{- "<|vision_start|><|video_pad|><|vision_end|>" }}
14
+ {%- elif content.type == 'text' %}
15
+ {{- content.text }}
16
+ {%- endif %}
17
+ {%- endfor %}
18
+ {%- endif %}
19
+ {%- endif %}
20
+ {{- '\n\n' }}
21
+ {{- "# 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>" }}
22
+ {%- for tool in tools %}
23
+ {{- "\n" }}
24
+ {{- tool | tojson }}
25
+ {%- endfor %}
26
+ {{- "\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" }}
27
+ {%- else %}
28
+ {%- if messages[0].role == 'system' %}
29
+ {%- if messages[0].content is string %}
30
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
31
+ {%- else %}
32
+ {%- for content in messages[0].content %}
33
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
34
+ {{- '<|im_start|>system\n' +"<|vision_start|><|image_pad|><|vision_end|>"+ '<|im_end|>\n' }}
35
+ {%- elif content.type == 'audio' or 'audio' in content or 'audio_url' in content %}
36
+ {{- '<|im_start|>system\n' +"<|audio_start|><|audio_pad|><|audio_end|>"+ '<|im_end|>\n' }}
37
+ {%- elif content.type == 'video' or 'video' in content %}
38
+ {{- '<|im_start|>system\n' +"<|vision_start|><|video_pad|><|vision_end|>"+ '<|im_end|>\n' }}
39
+ {%- elif content.type == 'text' %}
40
+ {{- '<|im_start|>system\n' +content.text+ '<|im_end|>\n' }}
41
+ {%- endif %}
42
+ {%- endfor %}
43
+ {%- endif %}
44
+ {%- endif %}
45
+ {%- endif %}
46
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
47
+ {%- for message in messages[::-1] %}
48
+ {%- set index = (messages|length - 1) - loop.index0 %}
49
+ {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
50
+ {%- set ns.multi_step_tool = false %}
51
+ {%- set ns.last_query_index = index %}
52
+ {%- endif %}
53
+ {%- endfor %}
54
+ {%- for message in messages %}
55
+ {%- if message.content is string %}
56
+ {%- set content = message.content %}
57
+ {%- else %}
58
+ {%- set content = namespace(text="") %}
59
+ {%- for mcontent in message.content %}
60
+ {%- if mcontent.type == 'image' or 'image' in mcontent or 'image_url' in mcontent %}
61
+ {%- set content.text = content.text~"<|vision_start|><|image_pad|><|vision_end|>" %}
62
+ {%- elif mcontent.type == 'audio' or 'audio' in mcontent or 'audio_url' in mcontent %}
63
+ {%- set content.text = content.text~"<|audio_start|><|audio_pad|><|audio_end|>" %}
64
+ {%- elif mcontent.type == 'video' or 'video' in mcontent %}
65
+ {%- set content.text = content.text~"<|vision_start|><|video_pad|><|vision_end|>" %}
66
+ {%- elif mcontent.type == 'text' %}
67
+ {%- set content.text = content.text~mcontent.text %}
68
+ {%- endif %}
69
+ {%- endfor %}
70
+ {%- set content = content.text %}
71
+ {%- endif %}
72
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
73
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
74
+ {%- elif message.role == "assistant" %}
75
+ {%- set reasoning_content = "" %}
76
+ {%- if message.reasoning_content is string %}
77
+ {%- set reasoning_content = message.reasoning_content %}
78
+ {%- else %}
79
+ {%- if '</think>' in content %}
80
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
81
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
82
+ {%- endif %}
83
+ {%- endif %}
84
+ {%- if loop.index0 > ns.last_query_index %}
85
+ {%- if loop.last or (not loop.last and reasoning_content) %}
86
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip("\n") + '\n</think>\n\n' + content.lstrip('\n') }}
87
+ {%- else %}
88
+ {{- '<|im_start|>' + message.role + '\n' + content }}
89
+ {%- endif %}
90
+ {%- else %}
91
+ {{- '<|im_start|>' + message.role + '\n' + content }}
92
+ {%- endif %}
93
+ {%- if message.tool_calls %}
94
+ {%- for tool_call in message.tool_calls %}
95
+ {%- if (loop.first and content) or (not loop.first) %}{{- '\n' }}{%- endif %}
96
+ {%- if tool_call.function %}
97
+ {%- set tool_call = tool_call.function %}
98
+ {%- endif %}
99
+ {{- '<tool_call>\n{"name": "' }}
100
+ {{- tool_call.name }}
101
+ {{- '", "arguments": ' }}
102
+ {%- if tool_call.arguments is string %}
103
+ {{- tool_call.arguments }}
104
+ {%- else %}
105
+ {{- tool_call.arguments | tojson }}
106
+ {%- endif %}
107
+ {{- '}\n</tool_call>' }}
108
+ {%- endfor %}
109
+ {%- endif %}
110
+ {{- '<|im_end|>\n' }}
111
+ {%- elif message.role == "tool" %}
112
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}{{- '<|im_start|>user' }}{%- endif %}
113
+ {{- '\n<tool_response>\n' }}
114
+ {{- content }}
115
+ {{- '\n</tool_response>' }}
116
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}{{- '<|im_end|>\n' }}{%- endif %}
117
+ {%- endif %}
118
+ {%- endfor %}
119
+ {%- if add_generation_prompt %}
120
+ {{- '<|im_start|>assistant\n' }}
121
+ {%- if enable_thinking is defined and enable_thinking is false %}{{- '<think>\n\n</think>\n\n' }}{%- endif %}
122
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,560 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3OmniMoeForConditionalGeneration"
4
+ ],
5
+ "assistant_token_id": 77091,
6
+ "code2wav_config": {
7
+ "attention_bias": false,
8
+ "attention_dropout": 0.0,
9
+ "codebook_dim": 512,
10
+ "codebook_size": 2048,
11
+ "decoder_dim": 1536,
12
+ "dtype": "bfloat16",
13
+ "hidden_act": "silu",
14
+ "hidden_size": 1024,
15
+ "intermediate_size": 3072,
16
+ "layer_scale_initial_scale": 0.01,
17
+ "max_position_embeddings": 8000,
18
+ "model_type": "",
19
+ "num_attention_heads": 16,
20
+ "num_hidden_layers": 8,
21
+ "num_key_value_heads": 16,
22
+ "num_quantizers": 16,
23
+ "num_semantic_quantizers": 1,
24
+ "pad_token_id": 151643,
25
+ "rms_norm_eps": 1e-05,
26
+ "rope_theta": 10000,
27
+ "semantic_codebook_size": 4096,
28
+ "sliding_window": 72,
29
+ "upsample_rates": [
30
+ 8,
31
+ 5,
32
+ 4,
33
+ 3
34
+ ],
35
+ "upsampling_ratios": [
36
+ 2,
37
+ 2
38
+ ],
39
+ "vector_quantization_hidden_dimension": 512
40
+ },
41
+ "dtype": "bfloat16",
42
+ "enable_audio_output": true,
43
+ "hidden_size": 1024,
44
+ "im_end_token_id": 151645,
45
+ "im_start_token_id": 151644,
46
+ "keys_to_ignore_at_inference": [
47
+ "past_key_values",
48
+ "hidden_states",
49
+ "attention_mask"
50
+ ],
51
+ "model_type": "qwen3_omni_moe",
52
+ "pad_token_id": 151643,
53
+ "system_token_id": 8948,
54
+ "talker_config": {
55
+ "accept_hidden_layer": 24,
56
+ "audio_end_token_id": 151670,
57
+ "audio_start_token_id": 151669,
58
+ "audio_token_id": 151675,
59
+ "code_predictor_config": {
60
+ "_name_or_path": "",
61
+ "add_cross_attention": false,
62
+ "architectures": null,
63
+ "attention_bias": false,
64
+ "attention_dropout": 0,
65
+ "bad_words_ids": null,
66
+ "begin_suppress_tokens": null,
67
+ "bos_token_id": null,
68
+ "chunk_size_feed_forward": 0,
69
+ "cross_attention_hidden_size": null,
70
+ "decoder_start_token_id": null,
71
+ "diversity_penalty": 0.0,
72
+ "do_sample": false,
73
+ "dtype": null,
74
+ "early_stopping": false,
75
+ "encoder_no_repeat_ngram_size": 0,
76
+ "eos_token_id": null,
77
+ "exponential_decay_length_penalty": null,
78
+ "finetuning_task": null,
79
+ "forced_bos_token_id": null,
80
+ "forced_eos_token_id": null,
81
+ "head_dim": 128,
82
+ "hidden_act": "silu",
83
+ "hidden_size": 1024,
84
+ "id2label": {
85
+ "0": "LABEL_0",
86
+ "1": "LABEL_1"
87
+ },
88
+ "initializer_range": 0.02,
89
+ "intermediate_size": 3072,
90
+ "is_decoder": false,
91
+ "is_encoder_decoder": false,
92
+ "label2id": {
93
+ "LABEL_0": 0,
94
+ "LABEL_1": 1
95
+ },
96
+ "layer_types": [
97
+ "full_attention",
98
+ "full_attention",
99
+ "full_attention",
100
+ "full_attention",
101
+ "full_attention"
102
+ ],
103
+ "length_penalty": 1.0,
104
+ "max_length": 20,
105
+ "max_position_embeddings": 32768,
106
+ "max_window_layers": 28,
107
+ "min_length": 0,
108
+ "model_type": "qwen3_omni_moe_talker_code_predictor",
109
+ "no_repeat_ngram_size": 0,
110
+ "num_attention_heads": 16,
111
+ "num_beam_groups": 1,
112
+ "num_beams": 1,
113
+ "num_code_groups": 16,
114
+ "num_hidden_layers": 5,
115
+ "num_key_value_heads": 8,
116
+ "num_return_sequences": 1,
117
+ "output_attentions": false,
118
+ "output_hidden_states": false,
119
+ "output_scores": false,
120
+ "pad_token_id": 151643,
121
+ "prefix": null,
122
+ "problem_type": null,
123
+ "pruned_heads": {},
124
+ "remove_invalid_values": false,
125
+ "repetition_penalty": 1.0,
126
+ "return_dict": true,
127
+ "return_dict_in_generate": false,
128
+ "rms_norm_eps": 1e-06,
129
+ "rope_scaling": null,
130
+ "rope_theta": 1000000,
131
+ "sep_token_id": null,
132
+ "sliding_window": null,
133
+ "suppress_tokens": null,
134
+ "task_specific_params": null,
135
+ "temperature": 1.0,
136
+ "tf_legacy_loss": false,
137
+ "tie_encoder_decoder": false,
138
+ "tie_word_embeddings": false,
139
+ "tokenizer_class": null,
140
+ "top_k": 50,
141
+ "top_p": 1.0,
142
+ "torchscript": false,
143
+ "typical_p": 1.0,
144
+ "use_bfloat16": false,
145
+ "use_cache": true,
146
+ "use_sliding_window": false,
147
+ "vocab_size": 2048
148
+ },
149
+ "codec_bos_id": 2149,
150
+ "codec_eos_token_id": 2150,
151
+ "codec_nothink_id": 2155,
152
+ "codec_pad_id": 2148,
153
+ "codec_think_bos_id": 2156,
154
+ "codec_think_eos_id": 2157,
155
+ "dtype": "bfloat16",
156
+ "image_token_id": 151655,
157
+ "model_type": "",
158
+ "num_code_groups": 16,
159
+ "output_router_logits": false,
160
+ "position_id_per_seconds": 13,
161
+ "seconds_per_chunk": 2,
162
+ "spatial_merge_size": 2,
163
+ "speaker_id": {
164
+ "aiden": 2303,
165
+ "chelsie": 2301,
166
+ "ethan": 2302
167
+ },
168
+ "text_config": {
169
+ "_name_or_path": "",
170
+ "add_cross_attention": false,
171
+ "architectures": null,
172
+ "attention_bias": false,
173
+ "attention_dropout": 0,
174
+ "bad_words_ids": null,
175
+ "begin_suppress_tokens": null,
176
+ "bos_token_id": null,
177
+ "chunk_size_feed_forward": 0,
178
+ "cross_attention_hidden_size": null,
179
+ "decoder_sparse_step": 1,
180
+ "decoder_start_token_id": null,
181
+ "diversity_penalty": 0.0,
182
+ "do_sample": false,
183
+ "dtype": "bfloat16",
184
+ "early_stopping": false,
185
+ "encoder_no_repeat_ngram_size": 0,
186
+ "eos_token_id": null,
187
+ "exponential_decay_length_penalty": null,
188
+ "finetuning_task": null,
189
+ "forced_bos_token_id": null,
190
+ "forced_eos_token_id": null,
191
+ "head_dim": 128,
192
+ "hidden_act": "silu",
193
+ "hidden_size": 1024,
194
+ "id2label": {
195
+ "0": "LABEL_0",
196
+ "1": "LABEL_1"
197
+ },
198
+ "initializer_range": 0.02,
199
+ "intermediate_size": 2048,
200
+ "is_decoder": false,
201
+ "is_encoder_decoder": false,
202
+ "label2id": {
203
+ "LABEL_0": 0,
204
+ "LABEL_1": 1
205
+ },
206
+ "length_penalty": 1.0,
207
+ "max_length": 20,
208
+ "max_position_embeddings": 65536,
209
+ "min_length": 0,
210
+ "mlp_only_layers": [],
211
+ "model_type": "qwen3_omni_moe_talker_text",
212
+ "moe_intermediate_size": 384,
213
+ "no_repeat_ngram_size": 0,
214
+ "norm_topk_prob": true,
215
+ "num_attention_heads": 16,
216
+ "num_beam_groups": 1,
217
+ "num_beams": 1,
218
+ "num_experts": 128,
219
+ "num_experts_per_tok": 6,
220
+ "num_hidden_layers": 20,
221
+ "num_key_value_heads": 2,
222
+ "num_return_sequences": 1,
223
+ "output_attentions": false,
224
+ "output_hidden_states": false,
225
+ "output_router_logits": false,
226
+ "output_scores": false,
227
+ "pad_token_id": 151643,
228
+ "prefix": null,
229
+ "problem_type": null,
230
+ "pruned_heads": {},
231
+ "remove_invalid_values": false,
232
+ "repetition_penalty": 1.0,
233
+ "return_dict": true,
234
+ "return_dict_in_generate": false,
235
+ "rms_norm_eps": 1e-06,
236
+ "rope_scaling": {
237
+ "interleaved": true,
238
+ "mrope_interleaved": true,
239
+ "mrope_section": [
240
+ 24,
241
+ 20,
242
+ 20
243
+ ],
244
+ "rope_type": "default",
245
+ "type": "default"
246
+ },
247
+ "rope_theta": 1000000,
248
+ "router_aux_loss_coef": 0.001,
249
+ "sep_token_id": null,
250
+ "shared_expert_intermediate_size": 768,
251
+ "sliding_window": null,
252
+ "suppress_tokens": null,
253
+ "task_specific_params": null,
254
+ "temperature": 1.0,
255
+ "tf_legacy_loss": false,
256
+ "tie_encoder_decoder": false,
257
+ "tie_word_embeddings": false,
258
+ "tokenizer_class": null,
259
+ "top_k": 50,
260
+ "top_p": 1.0,
261
+ "torchscript": false,
262
+ "typical_p": 1.0,
263
+ "use_bfloat16": false,
264
+ "use_cache": true,
265
+ "use_sliding_window": false,
266
+ "vocab_size": 3072
267
+ },
268
+ "thinker_hidden_size": 2048,
269
+ "video_token_id": 151656,
270
+ "vision_start_token_id": 151652
271
+ },
272
+ "thinker_config": {
273
+ "audio_config": {
274
+ "_name_or_path": "",
275
+ "activation_dropout": 0,
276
+ "activation_function": "gelu",
277
+ "add_cross_attention": false,
278
+ "architectures": null,
279
+ "attention_dropout": 0,
280
+ "bad_words_ids": null,
281
+ "begin_suppress_tokens": null,
282
+ "bos_token_id": null,
283
+ "chunk_size_feed_forward": 0,
284
+ "conv_chunksize": 500,
285
+ "cross_attention_hidden_size": null,
286
+ "d_model": 1280,
287
+ "decoder_start_token_id": null,
288
+ "diversity_penalty": 0.0,
289
+ "do_sample": false,
290
+ "downsample_hidden_size": 480,
291
+ "dropout": 0,
292
+ "dtype": "bfloat16",
293
+ "early_stopping": false,
294
+ "encoder_attention_heads": 20,
295
+ "encoder_ffn_dim": 5120,
296
+ "encoder_layers": 32,
297
+ "encoder_no_repeat_ngram_size": 0,
298
+ "eos_token_id": null,
299
+ "exponential_decay_length_penalty": null,
300
+ "finetuning_task": null,
301
+ "forced_bos_token_id": null,
302
+ "forced_eos_token_id": null,
303
+ "id2label": {
304
+ "0": "LABEL_0",
305
+ "1": "LABEL_1"
306
+ },
307
+ "initializer_range": 0.02,
308
+ "is_decoder": false,
309
+ "is_encoder_decoder": false,
310
+ "label2id": {
311
+ "LABEL_0": 0,
312
+ "LABEL_1": 1
313
+ },
314
+ "length_penalty": 1.0,
315
+ "max_length": 20,
316
+ "max_source_positions": 1500,
317
+ "min_length": 0,
318
+ "model_type": "qwen3_omni_moe_audio_encoder",
319
+ "n_window": 50,
320
+ "n_window_infer": 800,
321
+ "no_repeat_ngram_size": 0,
322
+ "num_beam_groups": 1,
323
+ "num_beams": 1,
324
+ "num_hidden_layers": 32,
325
+ "num_mel_bins": 128,
326
+ "num_return_sequences": 1,
327
+ "output_attentions": false,
328
+ "output_dim": 2048,
329
+ "output_hidden_states": false,
330
+ "output_scores": false,
331
+ "pad_token_id": 151643,
332
+ "prefix": null,
333
+ "problem_type": null,
334
+ "pruned_heads": {},
335
+ "remove_invalid_values": false,
336
+ "repetition_penalty": 1.0,
337
+ "return_dict": true,
338
+ "return_dict_in_generate": false,
339
+ "scale_embedding": false,
340
+ "sep_token_id": null,
341
+ "suppress_tokens": null,
342
+ "task_specific_params": null,
343
+ "temperature": 1.0,
344
+ "tf_legacy_loss": false,
345
+ "tie_encoder_decoder": false,
346
+ "tie_word_embeddings": true,
347
+ "tokenizer_class": null,
348
+ "top_k": 50,
349
+ "top_p": 1.0,
350
+ "torchscript": false,
351
+ "typical_p": 1.0,
352
+ "use_bfloat16": false
353
+ },
354
+ "audio_end_token_id": 151670,
355
+ "audio_start_token_id": 151669,
356
+ "audio_token_id": 151675,
357
+ "dtype": "bfloat16",
358
+ "image_token_id": 151655,
359
+ "initializer_range": 0.02,
360
+ "model_type": "qwen3_omni_moe_thinker",
361
+ "pad_token_id": 151643,
362
+ "position_id_per_seconds": 13,
363
+ "seconds_per_chunk": 2,
364
+ "text_config": {
365
+ "_name_or_path": "",
366
+ "add_cross_attention": false,
367
+ "architectures": null,
368
+ "attention_bias": false,
369
+ "attention_dropout": 0.0,
370
+ "bad_words_ids": null,
371
+ "begin_suppress_tokens": null,
372
+ "bos_token_id": null,
373
+ "chunk_size_feed_forward": 0,
374
+ "cross_attention_hidden_size": null,
375
+ "decoder_sparse_step": 1,
376
+ "decoder_start_token_id": null,
377
+ "diversity_penalty": 0.0,
378
+ "do_sample": false,
379
+ "dtype": "bfloat16",
380
+ "early_stopping": false,
381
+ "encoder_no_repeat_ngram_size": 0,
382
+ "eos_token_id": null,
383
+ "exponential_decay_length_penalty": null,
384
+ "finetuning_task": null,
385
+ "forced_bos_token_id": null,
386
+ "forced_eos_token_id": null,
387
+ "head_dim": 128,
388
+ "hidden_act": "silu",
389
+ "hidden_size": 2048,
390
+ "id2label": {
391
+ "0": "LABEL_0",
392
+ "1": "LABEL_1"
393
+ },
394
+ "initializer_range": 0.02,
395
+ "intermediate_size": 768,
396
+ "is_decoder": false,
397
+ "is_encoder_decoder": false,
398
+ "label2id": {
399
+ "LABEL_0": 0,
400
+ "LABEL_1": 1
401
+ },
402
+ "length_penalty": 1.0,
403
+ "max_length": 20,
404
+ "max_position_embeddings": 65536,
405
+ "min_length": 0,
406
+ "mlp_only_layers": [],
407
+ "model_type": "qwen3_omni_moe_text",
408
+ "moe_intermediate_size": 768,
409
+ "no_repeat_ngram_size": 0,
410
+ "norm_topk_prob": true,
411
+ "num_attention_heads": 32,
412
+ "num_beam_groups": 1,
413
+ "num_beams": 1,
414
+ "num_experts": 128,
415
+ "num_experts_per_tok": 8,
416
+ "num_hidden_layers": 48,
417
+ "num_key_value_heads": 4,
418
+ "num_return_sequences": 1,
419
+ "output_attentions": false,
420
+ "output_hidden_states": false,
421
+ "output_router_logits": false,
422
+ "output_scores": false,
423
+ "pad_token_id": 151643,
424
+ "prefix": null,
425
+ "problem_type": null,
426
+ "pruned_heads": {},
427
+ "remove_invalid_values": false,
428
+ "repetition_penalty": 1.0,
429
+ "return_dict": true,
430
+ "return_dict_in_generate": false,
431
+ "rms_norm_eps": 1e-06,
432
+ "rope_scaling": {
433
+ "interleaved": true,
434
+ "mrope_interleaved": true,
435
+ "mrope_section": [
436
+ 24,
437
+ 20,
438
+ 20
439
+ ],
440
+ "rope_type": "default",
441
+ "type": "default"
442
+ },
443
+ "rope_theta": 1000000,
444
+ "router_aux_loss_coef": 0.001,
445
+ "sep_token_id": null,
446
+ "shared_expert_intermediate_size": 0,
447
+ "sliding_window": null,
448
+ "suppress_tokens": null,
449
+ "task_specific_params": null,
450
+ "temperature": 1.0,
451
+ "tf_legacy_loss": false,
452
+ "tie_encoder_decoder": false,
453
+ "tie_word_embeddings": false,
454
+ "tokenizer_class": null,
455
+ "top_k": 50,
456
+ "top_p": 1.0,
457
+ "torchscript": false,
458
+ "typical_p": 1.0,
459
+ "use_bfloat16": false,
460
+ "use_cache": true,
461
+ "use_qk_norm": true,
462
+ "use_sliding_window": false,
463
+ "vocab_size": 152064
464
+ },
465
+ "user_token_id": 872,
466
+ "video_token_id": 151656,
467
+ "vision_config": {
468
+ "_name_or_path": "",
469
+ "add_cross_attention": false,
470
+ "apply_vit_abs_pos_embed": true,
471
+ "architectures": null,
472
+ "bad_words_ids": null,
473
+ "begin_suppress_tokens": null,
474
+ "bos_token_id": null,
475
+ "chunk_size_feed_forward": 0,
476
+ "cross_attention_hidden_size": null,
477
+ "decoder_start_token_id": null,
478
+ "deepstack_visual_indexes": [
479
+ 8,
480
+ 16,
481
+ 24
482
+ ],
483
+ "depth": 27,
484
+ "diversity_penalty": 0.0,
485
+ "do_sample": false,
486
+ "dtype": "bfloat16",
487
+ "early_stopping": false,
488
+ "encoder_no_repeat_ngram_size": 0,
489
+ "eos_token_id": null,
490
+ "exponential_decay_length_penalty": null,
491
+ "finetuning_task": null,
492
+ "forced_bos_token_id": null,
493
+ "forced_eos_token_id": null,
494
+ "hidden_act": "gelu_pytorch_tanh",
495
+ "hidden_size": 1152,
496
+ "id2label": {
497
+ "0": "LABEL_0",
498
+ "1": "LABEL_1"
499
+ },
500
+ "image_size": 768,
501
+ "in_channels": 3,
502
+ "in_chans": 3,
503
+ "initializer_range": 0.02,
504
+ "intermediate_size": 4304,
505
+ "is_decoder": false,
506
+ "is_encoder_decoder": false,
507
+ "label2id": {
508
+ "LABEL_0": 0,
509
+ "LABEL_1": 1
510
+ },
511
+ "length_penalty": 1.0,
512
+ "max_length": 20,
513
+ "min_length": 0,
514
+ "model_type": "qwen3_omni_moe_vision_encoder",
515
+ "no_repeat_ngram_size": 0,
516
+ "num_beam_groups": 1,
517
+ "num_beams": 1,
518
+ "num_heads": 16,
519
+ "num_position_embeddings": 2304,
520
+ "num_return_sequences": 1,
521
+ "out_hidden_size": 2048,
522
+ "output_attentions": false,
523
+ "output_hidden_states": false,
524
+ "output_scores": false,
525
+ "pad_token_id": 151643,
526
+ "patch_size": 16,
527
+ "prefix": null,
528
+ "problem_type": null,
529
+ "pruned_heads": {},
530
+ "remove_invalid_values": false,
531
+ "repetition_penalty": 1.0,
532
+ "return_dict": true,
533
+ "return_dict_in_generate": false,
534
+ "sep_token_id": null,
535
+ "spatial_merge_size": 2,
536
+ "spatial_patch_size": 16,
537
+ "suppress_tokens": null,
538
+ "task_specific_params": null,
539
+ "temperature": 1.0,
540
+ "temporal_patch_size": 2,
541
+ "tf_legacy_loss": false,
542
+ "tie_encoder_decoder": false,
543
+ "tie_word_embeddings": true,
544
+ "tokenizer_class": null,
545
+ "tokens_per_second": 2,
546
+ "top_k": 50,
547
+ "top_p": 1.0,
548
+ "torchscript": false,
549
+ "typical_p": 1.0,
550
+ "use_bfloat16": false
551
+ },
552
+ "vision_end_token_id": 151653,
553
+ "vision_start_token_id": 151652
554
+ },
555
+ "transformers_version": "4.57.1",
556
+ "tts_bos_token_id": 151672,
557
+ "tts_eos_token_id": 151673,
558
+ "tts_pad_token_id": 151671,
559
+ "user_token_id": 872
560
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "eos_token_id": [
3
+ 151643,
4
+ 151645
5
+ ],
6
+ "talker_max_new_tokens": 4096,
7
+ "talker_repetition_penalty": 1.05,
8
+ "talker_temperature": 0.9,
9
+ "talker_top_k": 50,
10
+ "talker_top_p": 1.0,
11
+ "transformers_version": "4.57.1"
12
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f44e70f971d30228d7f55ed4166603380ef6ede4fb259bd9520a01312a26fd6
3
+ size 4997899632
model-00002-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:987f9fdfabb3f90b00b4751a16aa50f50cff8e5d35637a487f68f06aa43546e9
3
+ size 4997754216
model-00003-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb026a05ef13968265647c1051b79d1352f52c781f2f7c43887d5d1ab35e7ba2
3
+ size 4997754216
model-00004-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2375cdb34b2c036819a6f50608167e7b09ded9547d841afb34e9698e84a2e2f9
3
+ size 4997755648
model-00005-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8225a98fa5afa91b322081d67b8d18a35bda2357cf6cac929f766f4d4be64ac9
3
+ size 4997755792
model-00006-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ecb4af6f0edada5aad83d92fe1174a70c1f2e3ae26351579c637d0003962f63
3
+ size 4997755792
model-00007-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c7b27c42418def7118949b4b4296d3c734fb00c8dc11350d9fcfa4122ae4f8c
3
+ size 4997755792
model-00008-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85b62d261ea78ee843ede5bee156c6fd3b50e62e691e76cf48b711af02148271
3
+ size 4997755792
model-00009-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c942ec86a356cc597ba20f4627c1a8178b15ae6839d929bbb8933cfaa7b55765
3
+ size 4997755792
model-00010-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54c053b3b9b99612ee8ced0929be1a49eda2733576d86a83eff4b3e5786dab6b
3
+ size 4997755792
model-00011-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24c3f98b65acb5fe29671537ad47136c1c8c7e2dd2414dee06cf07d33833b784
3
+ size 4997755792
model-00012-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5181cef513eb59e15a4916f1c99fd94f72928d5b29d97e8b906a9470e24ca5e9
3
+ size 4997755792
model-00013-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e002c98479bdbefed0d4e1d66cdd1897969c842753555573ae43eceb89519095
3
+ size 4999771808
model-00014-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73f09f187ff7b469a8937b39b8ac772e539e0d61311526c345558c5e6a6eb09a
3
+ size 4996618552
model-00015-of-00015.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68880d566a50643807a4e73f5227c4f468bb31532925ed5fae78a51bdd39492a
3
+ size 553698794
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dither": 0.0,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 128,
5
+ "hop_length": 160,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "Qwen2VLImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "max_pixels": 12845056,
18
+ "merge_size": 2,
19
+ "min_pixels": 3136,
20
+ "n_fft": 400,
21
+ "n_samples": 4800000,
22
+ "nb_max_frames": 30000,
23
+ "padding_side": "right",
24
+ "padding_value": 0.0,
25
+ "patch_size": 16,
26
+ "processor_class": "Qwen3OmniMoeProcessor",
27
+ "return_attention_mask": true,
28
+ "sampling_rate": 16000,
29
+ "temporal_patch_size": 2
30
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>",
16
+ "<|audio_start|>",
17
+ "<|audio_end|>",
18
+ "<tts_pad>",
19
+ "<tts_text_bos>",
20
+ "<tts_text_bos_single>",
21
+ "<|audio_pad|>"
22
+ ],
23
+ "audio_bos_token": "<|audio_start|>",
24
+ "audio_eos_token": "<|audio_end|>",
25
+ "audio_token": "<|audio_pad|>",
26
+ "eos_token": {
27
+ "content": "<|im_end|>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ },
33
+ "image_token": "<|image_pad|>",
34
+ "pad_token": {
35
+ "content": "<|endoftext|>",
36
+ "lstrip": false,
37
+ "normalized": false,
38
+ "rstrip": false,
39
+ "single_word": false
40
+ },
41
+ "video_token": "<|video_pad|>",
42
+ "vision_bos_token": "<|vision_start|>",
43
+ "vision_eos_token": "<|vision_end|>"
44
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09267689b8362020b9763b65dd5be7e086b31e28d72e02837a9e781de9a91bc7
3
+ size 11423986
tokenizer_config.json ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ },
213
+ "151669": {
214
+ "content": "<|audio_start|>",
215
+ "lstrip": false,
216
+ "normalized": false,
217
+ "rstrip": false,
218
+ "single_word": false,
219
+ "special": true
220
+ },
221
+ "151670": {
222
+ "content": "<|audio_end|>",
223
+ "lstrip": false,
224
+ "normalized": false,
225
+ "rstrip": false,
226
+ "single_word": false,
227
+ "special": true
228
+ },
229
+ "151671": {
230
+ "content": "<tts_pad>",
231
+ "lstrip": false,
232
+ "normalized": false,
233
+ "rstrip": false,
234
+ "single_word": false,
235
+ "special": true
236
+ },
237
+ "151672": {
238
+ "content": "<tts_text_bos>",
239
+ "lstrip": false,
240
+ "normalized": false,
241
+ "rstrip": false,
242
+ "single_word": false,
243
+ "special": true
244
+ },
245
+ "151673": {
246
+ "content": "<tts_text_eod>",
247
+ "lstrip": false,
248
+ "normalized": false,
249
+ "rstrip": false,
250
+ "single_word": false,
251
+ "special": true
252
+ },
253
+ "151674": {
254
+ "content": "<tts_text_bos_single>",
255
+ "lstrip": false,
256
+ "normalized": false,
257
+ "rstrip": false,
258
+ "single_word": false,
259
+ "special": true
260
+ },
261
+ "151675": {
262
+ "content": "<|audio_pad|>",
263
+ "lstrip": false,
264
+ "normalized": false,
265
+ "rstrip": false,
266
+ "single_word": false,
267
+ "special": true
268
+ }
269
+ },
270
+ "additional_special_tokens": [
271
+ "<|im_start|>",
272
+ "<|im_end|>",
273
+ "<|object_ref_start|>",
274
+ "<|object_ref_end|>",
275
+ "<|box_start|>",
276
+ "<|box_end|>",
277
+ "<|quad_start|>",
278
+ "<|quad_end|>",
279
+ "<|vision_start|>",
280
+ "<|vision_end|>",
281
+ "<|vision_pad|>",
282
+ "<|image_pad|>",
283
+ "<|video_pad|>",
284
+ "<|audio_start|>",
285
+ "<|audio_end|>",
286
+ "<tts_pad>",
287
+ "<tts_text_bos>",
288
+ "<tts_text_bos_single>",
289
+ "<|audio_pad|>"
290
+ ],
291
+ "audio_bos_token": "<|audio_start|>",
292
+ "audio_eos_token": "<|audio_end|>",
293
+ "audio_token": "<|audio_pad|>",
294
+ "bos_token": null,
295
+ "clean_up_tokenization_spaces": false,
296
+ "eos_token": "<|im_end|>",
297
+ "errors": "replace",
298
+ "extra_special_tokens": {
299
+ "audio_bos_token": "<|audio_start|>",
300
+ "audio_eos_token": "<|audio_end|>",
301
+ "audio_token": "<|audio_pad|>",
302
+ "image_token": "<|image_pad|>",
303
+ "video_token": "<|video_pad|>",
304
+ "vision_bos_token": "<|vision_start|>",
305
+ "vision_eos_token": "<|vision_end|>"
306
+ },
307
+ "image_token": "<|image_pad|>",
308
+ "model_max_length": 131072,
309
+ "pad_token": "<|endoftext|>",
310
+ "processor_class": "Qwen3OmniMoeProcessor",
311
+ "split_special_tokens": false,
312
+ "tokenizer_class": "Qwen2Tokenizer",
313
+ "unk_token": null,
314
+ "video_token": "<|video_pad|>",
315
+ "vision_bos_token": "<|vision_start|>",
316
+ "vision_eos_token": "<|vision_end|>"
317
+ }
video_preprocessor_config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": true,
5
+ "device": null,
6
+ "dither": 0.0,
7
+ "do_center_crop": null,
8
+ "do_convert_rgb": true,
9
+ "do_normalize": true,
10
+ "do_rescale": true,
11
+ "do_resize": true,
12
+ "do_sample_frames": false,
13
+ "feature_extractor_type": "WhisperFeatureExtractor",
14
+ "feature_size": 128,
15
+ "fps": null,
16
+ "hop_length": 160,
17
+ "image_mean": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "image_std": [
23
+ 0.5,
24
+ 0.5,
25
+ 0.5
26
+ ],
27
+ "input_data_format": null,
28
+ "max_frames": 768,
29
+ "max_pixels": 12845056,
30
+ "merge_size": 2,
31
+ "min_frames": 4,
32
+ "min_pixels": 3136,
33
+ "n_fft": 400,
34
+ "n_samples": 4800000,
35
+ "nb_max_frames": 30000,
36
+ "num_frames": null,
37
+ "pad_size": null,
38
+ "padding_side": "right",
39
+ "padding_value": 0.0,
40
+ "patch_size": 16,
41
+ "processor_class": "Qwen3OmniMoeProcessor",
42
+ "resample": 3,
43
+ "rescale_factor": 0.00392156862745098,
44
+ "return_attention_mask": true,
45
+ "return_metadata": false,
46
+ "sampling_rate": 16000,
47
+ "size": {
48
+ "longest_edge": 12845056,
49
+ "shortest_edge": 3136
50
+ },
51
+ "temporal_patch_size": 2,
52
+ "video_metadata": null,
53
+ "video_processor_type": "Qwen2VLVideoProcessor"
54
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff