icefall-multi-kd-finetune-amp-fp16
/
inference_audio_tagging
/log-decode-epoch-15-avg-7-use-averaged-model-2024-08-22-11-59-30
| 2024-08-22 11:59:30,532 INFO [inference_audio_tagging.py:316] Evaluation started | |
| 2024-08-22 11:59:30,532 INFO [inference_audio_tagging.py:318] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'warm_step': 2000, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': 'e400fa3b456faf8afe0ee5bfe572946b4921a3db', 'k2-git-date': 'Sat Jul 15 04:21:50 2023', 'lhotse-version': '1.16.0', 'torch-version': '2.0.1+cu117', 'torch-cuda-available': True, 'torch-cuda-version': '11.7', 'python-version': '3.9', 'icefall-git-branch': 'multi_KD_with_wenet', 'icefall-git-sha1': 'a932ad6d-clean', 'icefall-git-date': 'Wed Aug 21 18:06:09 2024', 'icefall-path': '/xy/mnt/yangxiaoyu/workspace/icefall_multi_KD', 'k2-path': '/root/anaconda3/lib/python3.9/site-packages/k2/__init__.py', 'lhotse-path': '/root/anaconda3/lib/python3.9/site-packages/lhotse/__init__.py', 'hostname': 'NGK_xiaoyu'}, 'epoch': 15, 'iter': 0, 'avg': 7, 'use_averaged_model': True, 'exp_dir': PosixPath('multi_KD/exp_finetune_asr_full_libri1_6-fold_do_AT1_KD_as_unbalanced_scale2.0_do_SV1_only_vox2_scale10.0_freeze_12000steps_encoder_lr_scale0.2_freeze_3layers_ecapa_lr_scale0.2_init_3_tasks_pretrain_avg_musan0_sync_task_md1500'), 'trained_with_distillation': True, 'trained_with_multitask': False, 'freeze_encoder': False, 'num_events': 527, 'eval_subset': 'eval', 'vocab_size': 500, 'blank_id': 0, 'context_size': 2, 'do_audio_tagging': True, 'use_encoder_projection': True, 'encoder_projection_dim': 2560, 'freezing_encoder_layer_index': '-1', 'freeze_encoder_steps': -1, 'save_logits': False, 'num_encoder_layers': '2,2,3,4,3,2', 'downsampling_factor': '1,2,4,8,4,2', 'feedforward_dim': '512,768,1024,1536,1024,768', 'num_heads': '4,4,4,8,4,4', 'encoder_dim': '192,256,384,512,384,256', 'query_head_dim': '32', 'value_head_dim': '12', 'pos_head_dim': '4', 'pos_dim': 48, 'encoder_unmasked_dim': '192,192,256,256,256,192', 'cnn_module_kernel': '31,31,15,15,15,31', 'decoder_dim': 512, 'joiner_dim': 512, 'causal': False, 'chunk_size': '32', 'left_context_frames': '256', 'use_transducer': True, 'use_ctc': False, 'speaker_input_idx': 2, 'whisper_dim': 1280, 'use_task_id': False, 'num_codebooks': 32, 'mvq_kd_layer_idx': -1, 'use_subsampled_output': True, 'delta_t': 0, 'full_libri': True, 'mini_libri': False, 'use_libriheavy': False, 'libriheavy_subset': 'small', 'use_librispeech': False, 'use_gigaspeech': False, 'gigaspeech_subset': None, 'use_wenetspeech': False, 'use_audioset': False, 'audioset_subset': 'balanced', 'use_voxceleb': False, 'voxceleb_subset': 'vox1', 'use_fma': False, 'fma_subset': 'large', 'manifest_dir': PosixPath('data/fbank_LS_Vox_AS_fma'), 'max_duration': 300, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'enable_audioset': False, 'use_musan_separately': False, 'input_strategy': 'PrecomputedFeatures', 'drop_features': False, 'return_audio': False, 'use_beats': True, 'use_ecapa': False, 'use_whisper': True, 'whisper_mvq': False, 'beats_ckpt': 'data/models/BEATs/BEATs_iter3_plus_AS2M_finetuned_on_AS2M_cpt2.pt', 'whisper_version': 'small.en', 'use_mert': False, 'lm_vocab_size': 500, 'lm_epoch': 7, 'lm_avg': 1, 'lm_exp_dir': None, 'rnn_lm_embedding_dim': 2048, 'rnn_lm_hidden_dim': 2048, 'rnn_lm_num_layers': 3, 'rnn_lm_tie_weights': True, 'transformer_lm_exp_dir': None, 'transformer_lm_dim_feedforward': 2048, 'transformer_lm_encoder_dim': 768, 'transformer_lm_embedding_dim': 768, 'transformer_lm_nhead': 8, 'transformer_lm_num_layers': 16, 'transformer_lm_tie_weights': True, 'res_dir': PosixPath('multi_KD/exp_finetune_asr_full_libri1_6-fold_do_AT1_KD_as_unbalanced_scale2.0_do_SV1_only_vox2_scale10.0_freeze_12000steps_encoder_lr_scale0.2_freeze_3layers_ecapa_lr_scale0.2_init_3_tasks_pretrain_avg_musan0_sync_task_md1500/inference_audio_tagging'), 'suffix': 'epoch-15-avg-7-use-averaged-model'} | |
| 2024-08-22 11:59:30,532 INFO [inference_audio_tagging.py:324] About to create model | |
| 2024-08-22 11:59:30,889 INFO [inference_audio_tagging.py:403] Calculating the averaged model over epoch range from 8 (excluded) to 15 | |
| 2024-08-22 11:59:44,748 INFO [inference_audio_tagging.py:421] Number of model parameters: 65577734 | |
| 2024-08-22 11:59:44,748 INFO [kd_datamodule.py:900] About to get the audioset eval cuts. | |
| 2024-08-22 11:59:44,798 INFO [kd_datamodule.py:584] About to create dev dataset | |
| 2024-08-22 11:59:45,218 INFO [kd_datamodule.py:605] About to create dev dataloader | |
| 2024-08-22 11:59:52,946 INFO [inference_audio_tagging.py:286] Processed 60 cuts already. | |
| 2024-08-22 11:59:58,092 INFO [inference_audio_tagging.py:286] Processed 660 cuts already. | |
| 2024-08-22 12:00:01,577 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.5446, 3.9109, 4.1468, 3.8065], device='cuda:0') | |
| 2024-08-22 12:00:03,683 INFO [inference_audio_tagging.py:286] Processed 1260 cuts already. | |
| 2024-08-22 12:00:08,890 INFO [inference_audio_tagging.py:286] Processed 1860 cuts already. | |
| 2024-08-22 12:00:11,662 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([6.0674, 5.8825, 6.0381, 6.0684], device='cuda:0') | |
| 2024-08-22 12:00:13,629 INFO [inference_audio_tagging.py:286] Processed 2460 cuts already. | |
| 2024-08-22 12:00:18,353 INFO [inference_audio_tagging.py:286] Processed 3060 cuts already. | |
| 2024-08-22 12:00:22,936 INFO [inference_audio_tagging.py:286] Processed 3660 cuts already. | |
| 2024-08-22 12:00:24,367 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.9725, 3.1965, 3.8187, 3.5284, 3.2349, 3.6825, 3.8033, 3.6414], | |
| device='cuda:0') | |
| 2024-08-22 12:00:25,913 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([5.1313, 4.8871, 5.0000, 4.3796], device='cuda:0') | |
| 2024-08-22 12:00:27,342 INFO [inference_audio_tagging.py:286] Processed 4260 cuts already. | |
| 2024-08-22 12:00:31,868 INFO [inference_audio_tagging.py:286] Processed 4860 cuts already. | |
| 2024-08-22 12:00:33,943 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.8808, 3.1099, 3.0571, 3.2618, 3.1449, 3.1382, 3.1140, 3.0702], | |
| device='cuda:0') | |
| 2024-08-22 12:00:36,406 INFO [inference_audio_tagging.py:286] Processed 5460 cuts already. | |
| 2024-08-22 12:00:37,434 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([6.0808, 5.8437, 6.0095, 6.0612], device='cuda:0') | |
| 2024-08-22 12:00:40,650 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.6085, 4.0373, 3.5733, 3.9854], device='cuda:0') | |
| 2024-08-22 12:00:41,067 INFO [inference_audio_tagging.py:286] Processed 6060 cuts already. | |
| 2024-08-22 12:00:42,213 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.9572, 4.2396, 4.4062, 3.9576], device='cuda:0') | |
| 2024-08-22 12:00:45,900 INFO [inference_audio_tagging.py:286] Processed 6660 cuts already. | |
| 2024-08-22 12:00:50,510 INFO [inference_audio_tagging.py:286] Processed 7260 cuts already. | |
| 2024-08-22 12:00:51,631 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.9668, 4.1681, 4.3864, 3.6854], device='cuda:0') | |
| 2024-08-22 12:00:55,133 INFO [inference_audio_tagging.py:286] Processed 7860 cuts already. | |
| 2024-08-22 12:00:55,545 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.5968, 4.0222, 3.6336, 4.0929], device='cuda:0') | |
| 2024-08-22 12:00:56,495 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.9679, 3.0367, 3.3920, 3.5204, 3.1459, 3.3276, 3.5428, 3.1649], | |
| device='cuda:0') | |
| 2024-08-22 12:00:59,613 INFO [inference_audio_tagging.py:286] Processed 8460 cuts already. | |
| 2024-08-22 12:01:02,391 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.9712, 3.1109, 3.4402, 3.6399, 3.2491, 3.2421, 3.5565, 3.2025], | |
| device='cuda:0') | |
| 2024-08-22 12:01:04,243 INFO [inference_audio_tagging.py:286] Processed 9060 cuts already. | |
| 2024-08-22 12:01:08,600 INFO [inference_audio_tagging.py:286] Processed 9660 cuts already. | |
| 2024-08-22 12:01:13,094 INFO [inference_audio_tagging.py:286] Processed 10260 cuts already. | |
| 2024-08-22 12:01:17,604 INFO [inference_audio_tagging.py:286] Processed 10860 cuts already. | |
| 2024-08-22 12:01:22,037 INFO [inference_audio_tagging.py:286] Processed 11460 cuts already. | |
| 2024-08-22 12:01:25,954 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.9487, 3.5733, 3.9973, 3.7565], device='cuda:0') | |
| 2024-08-22 12:01:26,682 INFO [inference_audio_tagging.py:286] Processed 12060 cuts already. | |
| 2024-08-22 12:01:31,433 INFO [inference_audio_tagging.py:286] Processed 12660 cuts already. | |
| 2024-08-22 12:01:36,033 INFO [inference_audio_tagging.py:286] Processed 13260 cuts already. | |
| 2024-08-22 12:01:40,870 INFO [inference_audio_tagging.py:286] Processed 13860 cuts already. | |
| 2024-08-22 12:01:45,605 INFO [inference_audio_tagging.py:286] Processed 14460 cuts already. | |
| 2024-08-22 12:01:50,309 INFO [inference_audio_tagging.py:286] Processed 15060 cuts already. | |
| 2024-08-22 12:01:50,808 INFO [inference_audio_tagging.py:287] Finish collecting audio logits | |
| 2024-08-22 12:01:52,158 INFO [inference_audio_tagging.py:454] mAP for audioset eval is: 0.458908065702477 | |
| 2024-08-22 12:01:52,158 INFO [inference_audio_tagging.py:456] Done | |